Economic Analysis for the
Final Ground Water Rule
-------
Office of Water (4606-M) EPA 815-R-06-014 October 2006 www.epa.gov/safewater
-------
Contents
Appendices vii
Exhibits vi
Acronyms xiii
Health Risk Reduction and Cost Analysis xvi
Executive Summary
ES.l Need forthe Rule ES-1
ES.2 Consideration of Regulatory Alternatives ES-2
ES.3 Summary of the Final GWR Requirements ES-2
ES.4 Systems Subjectto the GWR ES-6
ES.5 National Benefits and Costs of the GWR ES-8
ES.5.1 Derivation of Benefits ES-9
ES.5.2 Derivation of Costs ES-14
ES.6 Projected Impacts on Household Costs ES-15
ES.7 Comparison of Benefits and Costs, and of Regulatory Alternatives of the GWR .. ES-16
ES.8 Conclusions ES-19
1. Introduction
1.1 Summary of the Ground Water Rule 1-1
1.2 Document Organization 1-2
1.3 Calculations and Citations 1-2
2. Statement of Need for the Rule
2.1 Introduction 2-1
2.1.1 Description of the Issue 2-1
2.2 Public Health Concerns to Be Addressed 2-2
2.2.1 Contaminants and Their Health Effects 2-2
2.2.2 Sources of Contaminants 2-5
2.2.2.1 Ground Water Contamination through the Subsurface 2-6
2.2.2.2 Ground Water Contamination through the Wellhead 2-6
2.2.2.3 Contamination of Drinking Water in Distribution Systems 2-7
2.3 Statutory Authority for Promulgating the Rule 2-8
2.4 Regulatory History 2-8
2.4.1 1979 Total Trihalomethane Rule 2-9
2.4.2 1989 Total Coliform Rule 2-9
2.4.3 1996 Information Collection Rule 2-9
2.4.4 1998 Stage 1 Disinfectants and Disinfection Byproducts Rule 2-10
2.4.5 2001 Arsenic Rule 2-10
2.4.6 2006 Stage 2 Disinfectants and Disinfection Byproducts Rule 2-10
2.4.7 Underground Injection Control Program 2-11
2.5 Economic Rationale 2-12
Economic Analysis for the i October 2006
Final Ground Water Rule
-------
3. Consideration of Regulatory Alternatives
3.1 Introduction 3-1
3.2 Process for Development of Regulatory Alternatives 3-1
3.3 Regulatory Alternatives Considered 3-2
3.4 Final Rule Requirements 3-3
3.4.1 Mandatory Rule Components 3-6
3.4.2 Optional Provision 3-11
3.5 Other Changes Since Proposal 3-11
4. Baseline Conditions
4.1 Introduction 4-1
4.2 Industry Profile 4-1
4.2.1 Data Sources 4-2
4.2.2 Water System Characterization 4-2
4.2.3 Baseline Number of Systems, Entry Points, and Population 4-4
4.2.4 Water Treatment Plant Design and Average Daily Flows 4-15
4.2.5 Treatment Practices Baseline 4-17
4.2.6 Number of Households Served 4-17
4.2.7 Triggered Monitoring Baseline 4-18
4.3 Water Quality Baseline 4-23
4.3.1 Background 4-24
4.3.1.1 Representative Pathogens 4-24
4.3.2 Enterovirus and E. coli Occurrence in PWS Well Source Ground Water . . . 4-25
4.3.2.1 Lieberman et al. 2002 Study 4-27
4.3.2.2 Abbaszadegan et al. 2003 4-29
4.3.2.3 PennsylvaniaNoncommunity Wells (Lindsey etal., 2002) 4-31
4.3.2.4 Southeast Michigan (Francy et al., 2004) 4-32
4.3.2.5 New Jersey (Atherholt et al., 2003) 4-33
4.3.2.6 Missouri Ozark Plateau #1 (Davis and Witt, 2000) 4-34
4.3.2.7 Missouri Ozark Plateau #2 (Femmer, 2000) 4-35
4.3.2.8 Wisconsin Migrant Worker Camp (USEPA et al., 1998) 4-35
4.3.2.9 New England (Doherty et al., 1998) 4-36
4.3.2.10 Three-State Study: (Wisconsin-Battigelli, 1999) 4-37
4.3.2.11 Three-State Study: (Maryland-Banks etal, 2001) 4-37
4.3.2.12 Three-State Study: (Maryland-Banks and Battigelli, 2002) 4-38
4.3.2.13 Three-State Study: (Minnesota-Banks and Battigelli, 2002) 4-38
4.3.2.14 EPA Vulnerability Study (USEPA, 1998c) 4-39
4.3.2.15 Montana Study 4-39
4.3.2.16 Summary of New Data 4-40
4.3.3 Well Vulnerability 4-46
4.3.3.1 Background 4-46
4.3.3.2 Estimating percent wells invulnerability categories 4-47
4.3.4 Occurrence Analyses 4-49
4.3.4.1 Viral and Fecal Indicator Hit Rates 4-50
4.3.4.2 Pathogen Concentration Analysis 4-64
4.4 Outbreak Baseline and Causes of Contamination 4-68
4.5 Summary of Uncertainties in Development of GWR Baselines 4-70
Economic Analysis for the ii October 2006
Final Ground Water Rule
-------
5. Benefits Analysis
5.1 Introduction 5-1
5.1.1 Quantified Benefits 5-2
5.1.2 Nonqualified Benefits 5-4
5.2 Quantified Health Benefits from Reduction in Exposure to Viruses 5-4
5.2.1 Overview of Risk Assessment Methodology 5-5
5.2.2 Hazard Identification 5-7
5.2.2.1 Health Effects of Viral Infections 5-7
5.2.2.2 Sensitive Subgroups 5-8
5.2.3 Exposure Assessment 5-11
5.2.3.1 Source Water Viral Occurrence and Concentration 5-12
5.2.3.2 Finished Water Concentrations in Disinfecting
Ground Water Systems 5-12
5.2.3.3 Size of Exposed Population 5-13
5.2.3.4 Drinking Water Consumption Factors 5-13
5.2.4 Probability of Infection, Illness, and Mortality 5-17
5.2.4.1 Infectivity from Dose response Modeling of Human
Challenge Study Data 5-17
5.2.4.2 Morbidity and Mortality Data Sources and Uncertainty 5-21
5.2.4.3 Quantified GWR Benefits (Predictions of Illnesses and Deaths)-
Model Input Values 5-26
5.2.5 Risk Characterization 5-34
5.2.5.1 Risk Assessment Methodology for Baseline (Pre-GWR)
Risk Calculations 5-34
5.2.5.2 Results of the Baseline Risk Calculations 5-42
5.2.5.3 Baseline Illnesses and Deaths in Sensitive Subgroups 5-43
5.2.5.4 Baseline Risk to a Highly Exposed Individual 5-44
5.2.5.5 Sensitivity of Baseline Estimates to Quantified Uncertainty Inputs . 5-45
5.2.5.6 Methodology for Estimating Risk Reductions 5-48
5.2.5.7 Results for Risk Reduction for the Final GWR 5-54
5.2.5.8 Results for Reduction in Individual Risks for the Final GWR 5-56
5.2.5.9 Potential Increases in Health Risks 5-60
5.3 Monetized Benefits from Reduction in Exposure to Waterborne Pathogens 5-61
5.3.1 Value of Reduction in Type A and Type B Virus Cases 5-61
5.3.1.1 Value of Viral Illnesses Avoided 5-61
5.3.1.2 Value of Mortality Avoided 5-76
5.3.1.3 Measuring Benefits Over the GWR Implementation Schedule 5-76
5.3.1.4 Adjustments for Income Elasticity 5-76
5.3.1.5 Present Value of Future Benefits 5-78
5.3.2 Summary of Quantified Benefits of GWR 5-78
5.3.3 Quantified Benefits to Sensitive Subpopulations 5-81
5.4 Nonqualified Benefits of GWR Provisions 5-82
5.4.1 Decreased Incidence of Illness Caused by Other Type A Viruses 5-82
5.4.1.1 Norovirus 5-82
5.4.1.2 Other Type A viruses 5-86
5.4.2 Decreased Incidence of Other Illness Caused by Type B Viruses 5-88
5.4.3 Decreased Incidence of Bacterial Illness and Death 5-92
5.4.3.1 Bacterial Pathogen Occurrence 5-92
Economic Analysis for the Hi October 2006
Final Ground Water Rule
-------
5.4.3.2 Estimate of Potentially Avoided Bacterial Caused Deaths
by GWR 5-95
5.4.3.3 Estimate of a Hospitalization Rate for Waterborne Bacterial
Illness 5-98
5.4.4 Other Chronic and Acute Illness Potentially Avoided 5-100
5.4.5 Reduction in Outbreak Risk and Response Costs 5-102
5.4.6 Reduced Disinfection Treatment Failure Rates and Associated
Waterborne Disease 5-103
5.4.7 Distribution System Contamination 5-104
5.4.8 Benefits From the Reduction of Co-Occurring and Emerging
Contaminants 5-104
5.4.9 Reduced Uncertainty/Costs to Households to Avert Infection 5-105
5.4.10 Summary of Nonqualified Benefits 5-105
5.5 Alternative Analyses 5-107
5.5.1 Alternative Viral Concentration Baseline 5-107
5.5.2 Alternative Type A Dose response 5-107
5.5.3 Alternative Type B Dose Response 5-108
5.5.4 Alternative Occurrence (Peer Review) Data 5-108
5.6 Summary of Uncertainty 5-109
5.7 Regulatory Alternatives 5-112
6. Cost Analysis
6.1 Introduction 6-1
6.2 General Costing Assumptions and Methodology 6-2
6.2.1 Labor Rates 6-2
6.2.2 Laboratory Fees 6-4
6.2.3 Technology Unit Costs and Compliance Forecasts 6-5
6.2.4 Cost Model 6-6
6.2.5 Modeled Variability and Uncertainty in National Costs 6-6
6.3 Projecting and Discounting National Costs 6-6
6.4 Derivation of Costs for Systems and States 6-8
6.4.1 Rule Implementation and Annual Administration 6-11
6.4.2 Sanitary Surveys 6-16
6.4.3 Triggered Source Water Monitoring 6-25
6.4.4 Corrective Actions 6-32
6.4.4.1 Sanitary Survey Corrective Actions 6-32
6.4.4.2 Source Water Contamination Corrective Actions 6-36
6.4.5 Compliance Monitoring 6-44
6.4.6 Total Capital and One-Time Costs 6-53
6.4.7 Uncertainty in Unit Costs 6-53
6.4.8 Alternative Cost Analysis 6-54
6.5 Household Costs 6-54
6.6 Nonqualified Costs 6-56
6.7 Uncertainty Analysis 6-57
6.8 Total Annualized Cost for Final GWR Regulatory Alternative 6-61
6.9 Comparison of Regulatory Alternatives 6-62
Economic Analysis for the iv October 2006
Final Ground Water Rule
-------
7. Economic Impact Analysis
7.1 Introduction 7-1
7.2 Regulatory Flexibility Act and Small Business Regulatory Enforcement Fairness Act 7-1
7.3 Small Drinking Water System Variances 7-4
7.4 Feasible Treatment Technologies for All Systems 7-4
7.5 Effect of Compliance with the GWR on the Technical, Financial, and
Managerial Capacity of Public Water Systems 7-4
7.5.1 Requirements of the Final GWR 7-5
7.5.2 Systems Subject to the GWR 7-6
7.5.3 Impact of the GWR on System Capacity 7-6
7.5.4 Derivation of GWR Scores 7-9
7.5.4.1 Small Water Systems (Those Serving 10,000 or Fewer People) 7-9
7.5.4.2 Large Water Systems (Those Serving at Least 10,000 People) 7-10
7.6 Paperwork Reduction Act 7-10
7.7 Unfunded Mandates Reform Act 7-12
7.7.1 Social Benefits and Costs 7-13
7.7.2 Disproportionate Budgetary Effects 7-15
7.7.3 Macroeconomic Effects 7-17
7.7.4 Consultation with Small Governments 7-18
7.7.5 Consultation with State, Local, and Tribal Governments 7-18
7.7.6 Regulatory Alternatives Considered 7-19
7.7.7 Impacts on Small Governments 7-19
7.8 Indian Tribal Governments 7-19
7.9 Impacts on Sensitive Subpopulations 7-22
7.9.1 Protection of Children from Environmental Health Risks and Safety Risks . 7-23
7.10 Environmental Justice 7-23
7.11 Federalism 7-24
7.12 Actions Concerning Regulations That Significantly Affect Energy Supply,
Distribution, or Use 7-25
8. Comparison of Benefits and Costs of the LT2ESWTR
8.1 National Quantified Benefits and Costs of the GWR 8-1
8.1.1 National Quantified Benefits Summary 8-1
8.1.2 National Cost Summary 8-6
8.1.3 Comparison of National Quantified Benefits and Costs 8-6
8.2 Effect of Uncertainties and Nonquantified Benefit/Cost Estimates on the
Estimation of Net National Benefits 8-8
8.3 Breakeven Analysis 8-9
8.4 Comparison of Regulatory Alternatives 8-10
8.4.1 Comparison of Benefits and Costs 8-11
8.4.2 Cost-Effectiveness Measures 8-14
8.5 Summary of Conclusions 8-19
9. References
Economic Analysis for the v October 2006
Final Ground Water Rule
-------
Appendices
Appendix A: Calculations Supporting the Cost of Illness (COI) Analysis
Appendix B: Detail of Benefits Valuation Inputs
Appendix C: Benefits Detail
Appendix D: Costs Details
Appendix E: Potential Implications of Population Dynamics and Secondary Transmission of Infection
on the Benefits of the Groundwater Rule
Appendix F: Infectivity Dose Response Relationships: Description of Analysis Conducted to Select
Model Forms and Estimate Model Parameters
Appendix G: Summary Flowcharts for Baseline Risk and Benefits Model
Appendix H: Cost Effectiveness Analysis Using a Quality-Adjusted Life Years Approach
Appendix I: Analysis of Total Coliform Hit Rates in Drinking Water Systems With Ground Water
Sources
Appendix J: Changes in GWR Economic Analysis from Proposal to Final
Appendix K: Cost Details for Alternatives 1, 3, and 4
Appendix L: Summary Flow Charts for GWR Cost Model
Economic Analysis for the vi October 2006
Final Ground Water Rule
-------
Exhibits
Executive Summary
Exhibit ES.l Flowchart of Compliance with GWR Requirements for Systems
Exhibit ES.2 Implementation Timeline for the GWR
Exhibit ES.3 Summary of Rule Implications
Exhibit ES.4 Summary of Annual Avoided Viral Illnesses and Deaths by System Type
Exhibit ES.5 Summary of Annualized Present Value Quantified Benefits ($Millions, 2003$)
Exhibit ES.6 Summary of Benefits of the GWR
Exhibit ES.7 Total Initial Capital and One-Time Costs (SMillions, 2003$)
Exhibit ES.8 Total Annualized Present Value Costs of the GWR (SMillions, 2003$)
Exhibit ES.9 Summary of Annual Per-Household Costs for the GWR (2003$Year)
Exhibit ES. 10 Estimated Annualized National Benefits and Costs for the GWR
($Millions, 2003$)
. ES-4
. ES-5
. ES-7
ES-10
ES-11
ES-13
ES-15
ES-15
ES-16
ES-17
1. Introduction
2. Statement of Need for the Rule
Exhibit 2.1 Examples of Illnesses Caused by Known or Suspected Waterborne Fecal Viral
Pathogens 2-4
Exhibit 2.2 Examples of Illnesses Caused by Common Waterborne Bacterial Pathogens 2-4
Exhibit 2.3 Etiology of Waterborne Outbreaks in Ground Water Systems,
1991-2000 2-5
3. Consideration of Regulatory Alternatives
Exhibit 3.1 Flowchart of Mandatory PWS Ground Water Rule Requirements
4. Baseline Conditions
Exhibit 4.1 Ground Water Rule System Baseline 4-6
Exhibit 4.2 Ground Water Rule System Baseline: Disinfecting and Nondisinfecting
Systems 4-8
Exhibit 4.3 Ground Water Rule Entry Point Baseline 4-10
Exhibit 4.4 Ground Water Rule System Population Baseline 4-12
Exhibit 4.5 Ground Water Rule Entry Point Population Baseline 4-13
Exhibit 4.6 Design Flows and Average Daily Flows per Plant (MGD) 4-16
Exhibit 4.7 Disinfection Treatment Practices for Disinfecting Ground Water Systems 4-18
Exhibit 4.8 Ground Water Rule CWS Household Baseline 4-18
Exhibit 4.9 Total Coliform Positive Hit Rates 4-21
Exhibit 4.10 Estimated Number of Routine Total Coliform Samples
Taken Per System, Per Year, by Type and Size of System 4-22
Exhibit 4.11 Estimated Number of TC+ Samples Per System, Per Year,
by System Size and System Type 4-22
Exhibit 4.12 Results of the Abbaszadegan et al. 2003 Study 4-30
Exhibit 4.13a New Data Available since publication of the Proposed GWR 4-42
Economic Analysis for the
Final Ground Water Rule
October 2006
-------
Exhibit 4.13b Virus Assays and Positives for 1,253 Wells Assayed for Viruses 4-45
Exhibit 4.13c E.coli Assays and Positives for 687 Wells Assayed for E.coli 4-45
Exhibit 4.13d Number of Virus and E.coli Assays 4-46
Exhibit 4.14 Mean Percent of Systems with Acute or Monthly MCL Violations
by System Type and System Size 4-48
Exhibit 4.15 Number and Percent of Systems Disinfecting, By Type of System 4-48
Exhibit 4.16 Categories of Indicator and Viral Classification Among PWS Wells in U.S 4-51
Exhibit 4.17 Median of 10,000 Estimates of PI, P2, P3, and P4
(with error bars showing the 5th and 95th percentiles) 4-55
Exhibit 4.18 Median of 10,000 Estimates of Pwell for Virus and Indicator
(with error bars showing the 5th and 95th percentiles) 4-56
Exhibit 4.19 Scatter Plot of Pwell Pairs for Indicators and Viruses 4-57
Exhibit 4.20 Density Function Shapes of Psample for Viruses 4-58
Exhibit 4.21 Cumulative Distributions of Psample for Viruses 4-58
Exhibit 4.22 Median of 10,000 Estimates of Psample for Virus and Indicator
(with error bars showing the 5th and 95th percentiles) 4-59
Exhibit 4.23 Scatter Plot of Means of Psample Pairs for Indicators and Viruses 4-60
Exhibit 4.24 Mean of Psampie Versus Pwell for Viruses
(1,000 Pairs from Occurrence Model) 4-61
Exhibit 4.25 Mean of Psampie Versus Pwell for Indicators
(1,000 Pairs from Occurrence Model) 4-61
Exhibit 4.26 Cumulative Probability of an Indicator Positive as a
Function of Assay Number ~ All Wells (used for cost analysis) 4-63
Exhibit 4.27 Cumulative Probability of an Indicator Positive as a Function of
Assay Number ~ Virus Positive Wells (used for risk reduction analysis) 4-64
Exhibit 4.28 Summary of Virus Concentrations Observed
in the Lieberman et al. 2002 Study 4-66
Exhibit 4.29 Summary of Virus Concentrations Observed
in the Abbaszadegan et al. 2003 Study 4-67
Exhibit 4.30 Summary of Virus Concentrations Observed
in the Pennsylvania Noncommunity Study 4-68
Exhibit 4.31 Summary of Waterborne Disease Outbreaks Attributable to PWSs Served by
Wells using Ground Water: 1991-2000* 4-71
Exhibit 4.32 Summary of Uncertainties Affecting GWR Baseline Estimates 4-73
5. Benefits Analysis
Exhibit 5.1 Overview of Quantified and Nonqualified GWR Benefits 5-1
Exhibit 5.2 Overview of Viral Pathogen Risk Assessment and Benefits Valuation Procedure
for Quantified Benefits (Main Analysis 5-3
Exhibit 5.3 Health Risk Assessment Framework 5-6
Exhibit 5.4 Sensitive Populations in the United States 5-9
Exhibit 5.5 Distribution of Individual Daily Drinking Water Consumptionby Age Group
(L/person/day) 5-14
Exhibit 5.6 Derivation of Transient Noncommunity Water System Consumption Factor 5-16
Exhibit 5.7 Estimated Exposure Days by System Type 5-17
Exhibit 5.8 Rotavirus Dose-Response Data 5-19
Exhibit 5.9 Echovirus Dose-Response Data 5-19
Exhibits.10 Dose Response Assumptions of Viral Pathogens for the GWR Risk Assessment ... 5-26
Economic Analysis for the
Final Ground Water Rule
October 2006
-------
Exhibit 5.11 Summary of Virus Watch Morbidity Data 5-30
Exhibit 5.12 Summary Table of Risk Calculation Factors, Distribution Category (Variability,
Uncertainty, Constant) and Distribution Type) 5-37
Exhibit 5.13 Estimates of Annual Baseline Viral Illness and Death1 5-43
Exhibit 5.14 Baseline Illnesses and Deaths in Sensitive Subgroups 5-44
Exhibit 5.15a Summary of Pearson Correlation Coefficients for Uncertain Inputs, Virus Type A . 5-47
Exhibit 5.15b Summary of Pearson Correlation Coefficients for Uncertain Inputs, Virus Type B . 5-48
Exhibit 5.16 Example of Baseline Cases, Cases Remaining, and Cases Avoided
atthe Well Level 5-50
Exhibit 5.17 Annual Viral Illnesses and Deaths Avoided for the GWR by System Size and Type . 5-55
Exhibit 5.18 Summary of Annual Viral Illnesses and Deaths Avoided for the GWR 5-56
Exhibit 5.19 Viral Illnesses and Deaths Avoided in the 25th Year
after Implementation of the GWR 5-56
Exhibit 5.20a Comparison of Average Annual Individual Infectivity Risk Distributions for
Baseline and Rule Alternatives for Type A Viruses 5-58
Exhibit 5.20b Comparison of Average Annual Individual Infectivity Risk Distributions for
Baseline and Rule Alternatives for Type B Viruses 5-59
Exhibit 5.21a Estimates for Average Cost and Average Cost per Healthy Patient of Type A
Illness, by Age (Enhanced COI) 5-65
Exhibit 5.21b Estimates for Average Cost and Average Cost per
Immunocompromised Patient of Type A Illness, by Age (Enhanced COI) 5-66
Exhibit 5.21c: Estimates for Average Cost and Average Cost per Case of Type B Illness
Requiring No Medical Care, by Age (Enhanced COI) 5-67
Exhibit 5.2 Id Estimates for Average Cost and Average Cost per Case of Type B Illness
Requiring Outpatient Care, by Age (Enhanced COI) 5-68
Exhibit 5.21e Estimates for Average Cost and Average Cost per Case of Type B Illness
Requiring Inpatient Care, by Age (Enhanced COI) 5-69
Exhibit 5.22a Estimates for Average Cost and Average Cost per Healthy Patient of Type A
Illness, by Age (Traditional COI) 5-70
Exhibit 5.22b Estimates for Average Cost and Average Cost per
Immunocompromised Patient of Type A Illness, by Age (Traditional COI) 5-71
Exhibit 5.22c Estimates for Average Cost and Average Cost per Case of Type B Illness
Requiring No Medical Care, by Age (Traditional COI) 5-72
Exhibit 5.22d Estimates for Average Cost and Average Cost per Case of Type B Illness
Requiring Outpatient Care, by Age (Traditional COI) 5-73
Exhibit 5.22e Estimates for Average Cost and Average Cost per Case of Type B Illness
Requiring Inpatient Care, by Age (Traditional COI) 5-74
Exhibit 5.23a Annualized Quantified Benefits of Illnesses and Deaths Avoided, Final Rule,
Enhanced COI, All Systems by System Size and Type (SMillions, 2003) 5-79
Exhibit 5.23b Annualized Quantified Benefits of Illnesses and Deaths Avoided, Final Rule,
Traditional COI, All Systems, by System Size and Type (SMillions, 2003) 5-80
Exhibit 5.24 Annual Illnesses and Deaths Avoided at Full Implementation, and Quantified
Benefits of the GWR in Sensitive Populations 5-81
Exhibit 5.25 Estimated Bacterial Illnesses and Deaths Avoided 5-97
Exhibit 5.26 Waterborne Bacterial Illness Hospitalization Rates 5-99
Exhibit 5.27 Results of Alternative Analyses 5-109
Exhibit 5.28 Summary of Uncertainties Affecting GWR Estimates 5-110
Economic Analysis for the
Final Ground Water Rule
October 2006
-------
Exhibit 5.29 Remaining Number of Annual Viral Illnesses and Deaths for Each
Regulatory Alternative 5-113
Exhibit 5.30 Comparison of Number of Annual Viral Illnesses and Deaths Avoided for
Regulatory Alternatives 5-113
Exhibit 5.31 Annualized Value of Viral Illnesses and Deaths Avoided for
Regulatory Alternatives 5-113
6. Cost Analysis
Exhibit 6.1 Wage Rates by System Size 6-3
Exhibit 6.2 State Labor Rates 6-4
Exhibit 6.3 Source Water Monitoring Costs per Sample 6-5
Exhibit 6.4 Discount Rates for Private and Public Systems 6-7
Exhibit 6.5a GWR Baselines: Number of Systems, Entry Points, and Wells 6-9
Exhibit 6.5b Summary of Rule Implications 6-10
Exhibit 6.5c Annualized Costs for Meeting Each of the GWR Provisions to Systems and
States (SMillions, 2003$) 6-11
Exhibit 6.6 PWS Unit Burden and Cost Estimates for Implementation Activities 6-13
Exhibit 6.7a State Burden and Cost Estimates for Implementation Activities 6-15
Exhibit 6.7b State Burden and Cost Estimates for Annual Administration 6-15
Exhibit 6.8 PWS and State Cost Estimates for Implementation and Annual
Administration Activities (SMillions, 2003$) 6-16
Exhibit 6.9 Schematic of Sanitary Survey Process 6-18
Exhibit 6.10 Number of Full and Incremental Sanitary Surveys for Systems 6-20
Exhibit 6.1 la PWS Unit Burden and Cost Estimates for Performing Full and Incremental
Sanitary Surveys (Treatment) 6-21
Exhibit 6.1 Ib PWS Unit Burden and Cost Estimates for Performing Full and Incremental
Sanitary Surveys (No Treatment) 6-22
Exhibit 6.12a State Unit Burden and Cost Estimates for Performing Full and Incremental
Sanitary Surveys (Treatment) 6-23
Exhibit 6.12b State Unit Burden and Cost Estimates for Performing Full and Incremental
Sanitary Surveys (No Treatment) 6-24
Exhibit 6.13 PWS and State Cost Estimates for Sanitary Survey Performance
(SMillions, 2003$) 6-25
Exhibit 6.14 Schematic of Triggered Monitoring Process 6-26
Exhibit 6.15 Estimated Number of Triggered Samples Per Year Per Entry Point 6-27
Exhibit 6.16 PWS Unit Costs for Triggered Monitoring 6-30
Exhibit 6.17 PWS and State Cost Estimates for Performing Triggered Monitoring
($Millions, 2003$) 6-31
Exhibit 6.18 Estimated Distribution of Significant Deficiency Corrective Actions 6-33
Exhibit 6.19 Estimated Unit Costs of Significant Deficiency Corrective Actions 6-34
Exhibit 6.20a PWS and State Unit Costs for Corrective Action Plans 6-35
Exhibit 6.20b PWS and State Cost Estimates for Sanitary Survey Corrective Action
Activities ($Millions, 2003$) 6-36
Exhibit 6.2 la Summary Flow Chart Estimated Distribution of Source Water Contamination
Corrective Actions 6-39
Exhibit 6.2 Ib Estimated Distribution of Source Water Contamination Corrective Actions 6-40
Exhibit 6.22a Estimated Unit Costs of Nontreatment Corrective Actions
for Source Water Contamination 6-42
Economic Analysis for the
Final Ground Water Rule
October 2006
-------
Exhibit 6.22b Estimated Unit Costs of Treatment Corrective Actions for Source
Water Contamination 6-43
Exhibit 6.23 PWS and State Cost Estimates for Triggered Monitoring Corrective
Action Activities ($Millions, 2003$) 6-44
Exhibit 6.24 Schematic of Compliance Monitoring Process 6-45
Exhibit 6.25 Assumptions for Entry Points Subject to Compliance Monitoring 6-47
Exhibit 6.26 PWS Unit Costs for Compliance Monitoring for Initial State Notification and
Disinfection Failure Reports 6-49
Exhibit 6.27a PWS Compliance Monitoring Unit Costs for Systems Serving 3,300 or
Fewer People 6-50
Exhibit 6.27b PWS Compliance Monitoring Capital Unit and O&M Costs for Systems
Serving More than 3,300 People 6-51
Exhibit 6.28 State Unit Costs for Compliance Monitoring 6-52
Exhibit 6.29 PWS and State Cost Estimates for Compliance Monitoring Activities
(SMillions, 2003$) 6-52
Exhibit 6.30 Total Initial Capital and One-Time Costs (SMillions, 2003$) 6-53
Exhibit 6.31 Summary of Annual Per-Household Costs for the GWR (2003$/Year) 6-55
Exhibit 6.32 Cost Uncertainty Summary 6-60
Exhibit 6.33 Total Annualized Present Value Costs ($Millions, 2003$) 6-61
Exhibit 6.34 Total Annualized Costs to Systems by System Size and Type ($Millions, 2003$) . . . 6-62
Exhibit 6.35 Comparison of National Annual Costs by Regulatory Alternative
($Millions, 2003$) 6-63
7. Economic Impact Analysis
Exhibit 7. la Estimated Impact of the GWR on Small System's Technical, Managerial, and
Financial Capacity 7-7
Exhibit 7. Ib Estimated Impact of the GWR on Large System's Technical, Managerial, and
Financial Capacity 7-8
Exhibit 7.2 Average Annual Burden Hours and Costs for the GWR Information Collection
Request 3-Year Approval Period 7-11
Exhibit 7.3 Ground Water System, State, and Tribal Estimated 7-13
Exhibit 7.4 Mean Total Annualized Benefits and Costs of Regulatory Alternatives
($Millions, 2003$) 7-14
Exhibit 7.5 Annualized Compliance Costs by Type of Ground Water System 7-16
Exhibit 7.6 Mean Annualized Compliance Cost per Ground Water System by
System Size and Type 7-17
Exhibit 7.7 Annual Cost of Compliance for Tribal Systems by System Type and Size
(Annualized at 3 Percent) 7-21
Exhibit 7.8 Total Increased Annual National Energy Usage Attributable to the GWR 7-27
Exhibit 7.9 Sample Calculation for Determining Increase in Energy Usage: Chlorine
Dioxide (C1O2 Dose = 1.25 mg/L) 7-28
8. Comparison of Benefits and Costs of the LT2ESWTR
Exhibit 8.1 Summary of Annual Avoided Viral Illnesses and Deaths by System Type 8-2
Exhibit 8.2 Summary of Annualized Present Value Quantified Benefits ($Millions, 2003$) 8-4
Exhibit 8.3 Summary of Benefits of the GWR 8-5
Economic Analysis for the
Final Ground Water Rule
October 2006
-------
Exhibit 8.4 Summary of Quantified Costs, Final Rule (SMillions, 2003$) 8-6
Exhibit 8.5a Estimated Annualized National Benefits and Costs for the GWR
(SMillions, 2003$) 8-7
Exhibit 8.5b Estimated Net Benefits Including Annual Bacterial Illness and Death
Avoidance Estimate (SMillions 2003$) 8-8
Exhibit 8.6 Estimated Breakeven Points 8-10
Exhibit 8.7 Annualized Costs, by Regulatory Alternative ($Millions, 2003$) 8-11
Exhibit 8.8 Number of Annual Quantified Viral Illnesses and Deaths Avoided
by Regulatory Alternatives 8-11
Exhibit 8.9 Annualized Value of Quantified Viral Illnesses and Deaths Avoided,
by Regulatory Alternative ($Millions, 2003$) 8-12
Exhibit 8.10a Annualized Net Benefits by Regulatory Alternative ($Millions, 2003$) 8-13
Exhibit 8. lOb Annualized Mean Net Benefits for Final Rule Including Estimates for
Nonqualified Benefits ($Millions, 2003$) 8-13
Exhibit 8.1 la Mean Annualized Costs at Mean Benefit Level, Enhanced COI,
by Regulatory Alternative 8-15
Exhibit 8.1 Ib Mean Annualized Costs at Mean Benefit Level, Traditional COI,
by Regulatory Alternative 8-16
Exhibit 8.12 Cost Per Viral Illness or Death Avoided by Regulatory Alternative (2003$) 8-17
Exhibit 8.13a Incremental Net Quantified Benefits by Rule Alternative - Enhanced COI
(Annualized Present Value Mean, $Millions, 2003$) 8-18
Exhibit 8.13b Incremental Net Quantified Benefits by Rule Alternative - Traditional COI
(Annualized Present Value Mean, $Millions, 2003$) 8-19
9. References
Economic Analysis for the
Final Ground Water Rule
xi i
October 2006
-------
List of Acronyms
ADA American Diabetes Association
AIDS Acquired Immune Deficiency Syndrome
ASDWA Association of State Drinking Water Administrators
AWWA American Water Works Association
AWWARF American Water Works Association Research Foundation
AWWSCo American Water Works Service Company
BLS Bureau of Labor Statistics
BMP Best Management Practice
BGM Buffalo Green Monkey
CCR Consumer Confidence Report
CDC Centers for Disease Control and Prevention
CFR Combined Federal Register
Cl chlorine
C1O2 chlorine dioxide
COI Cost of Illness
CDBG Community Development Block Grant
CPI Consumer Price Index
CT product of the residual disinfectant concentration (C) & the disinfectant contact time (T)
CWS Community Water System
CWSS Community Water System Survey
DBP Disinfection Byproduct
DBPR Disinfectants and Disinfection Byproducts Rule
DOE Department of Energy
DWSRF Drinking Water State Revolving Fund
E Income Elasticity
EA Economic Analysis
EIA Economic Impact Analysis
EPA United States Environmental Protection Agency
FBRR Filter Backwash Recycling Rule
FDA Food and Drug Administration
FSIS Federalism summary impact statement
FR Federal Register
FTE Full time equivalent
GDP Gross Domestic Product
gpd gallons per day
gpm gallons per minute
GWR Ground Water Rule
GWSS Ground Water Supply Survey
GWUDI Ground Water Under Direct Influence of Surface Water
HAA Haloacetic Acid
HAAS Sum of 5 Haloacetic Acids
HAV hepatitis A virus
HCUP Hospital Cost and Utilization Project
HSA Hydrogeologic Sensitivity Analysis
I Income
ICR Information Collection Rule
IESWTR Interim Enhanced Surface Water Treatment Rule
Economic Analysis for the
Final Ground Water Rule
Xlll
October 2006
-------
IDDM Insulin-dependent diabetes mellitus
IDSE Initial Distribution System Evaluation
ICD International Class of Diseases
kgal kilogallons
kgpd kilogallons per day
kW kilowatt
kWh kilowatt hour
kWh/y kilowatt hour per year
L liter
LRAA locational running annual average
LT1ESWTR Long Term 1 Enhanced Surface Water Treatment Rule
MCL Maximum Contaminant Level
MCLGs Maximum Contaminant Level Goal
MF Microfiltration
mgd million gallons per day
mg/L milligrams per liter
Hg/L microgram per liter
MPN Most probable number
MPNIU Most probable number of infectious units
MRDL Maximum residual disinfectant level
MRDLG Maximum residual disinfectant level goal
NCHS National Center for Health Statistics
NHLBI National Health, Lung and Blood Institute
NCWS Noncommunity Water System
NDWAC National Drinking Water Advisory Committee
NFID National Foundation for Infectious Diseases
NRC National Research Council
NWRI National Water Research Institute
NF Nanofiltration
nm nanometers
NPDES National Pollutant Discharge Elimination System
NPDWR National Primary Drinking Water Regulations
NRC National Research Council
NTNCWS Nontransient Noncommunity Water System
O&M Operation and Maintenance
OGWDW Office of Ground Water and Drinking Water
OMB Office of Management and Budget
POE Point-of-Entry
PCR Polymerase Chain Reaction
POTW Publicly Owned Treatment Works
POU Point-of-Use
PPI Producer Price Index
ppm parts per million
PWS Public Water System
PWSS Public Water Systems Supervision
PV Present Value
RAA Running Annual Average
RFA Regulatory Flexibility Act
RNA Ribonucleic acid
Economic Analysis for the
Final Ground Water Rule
xiv
October 2006
-------
RO Reverse Osmosis
RTI Research Triangle Institute
RT-PCR Reverse Transcription-Polymerase Chain Reaction
RUS Rural Utility Service
SAB Science Advisory Board
SBA Small Business Administration
SBAR Small Business Advocacy Review
SBREFA Small Business Regulatory Enforcement Fairness Act
SDWA Safe Drinking Water Act
SDWIS Safe Drinking Water Information System
SER Small entity representative
SIC Standard Industrial Classification
SOC Standard Occupational Classification
SRSV Small round, structured viruses
SRMD Standards and Risk Management Division
SWAPP Source Water Assessment and Protection Program
SWTR Surface Water Treatment Rule
T&C Technology and Cost
TC Total Coliform
TCR Total Coliform Rule
TMF Technical, managerial, and financial
THM Trihalomethane
TNCWS Transient Noncommunity Water System
TTHM Total Trihalomethanes
UIC Underground Injection Control
UF Ultrafiltration
UMRA Unfunded Mandates Reform Act
USDA United States Department of Agriculture
UV Ultraviolet
VSL Value of a Statistical Life
WTP Willingness to Pay
Economic Analysis for the
Final Ground Water Rule
xv
October 2006
-------
Health Risk Reduction and Cost Analysis
Under the Safe Drinking Water Act (SDWA) Amendments of 1996, when proposing a national
primary drinking water regulation that includes an maximum contaminant level (MCL), the U.S.
Environmental Protection Agency (EPA or the Agency) must conduct a health risk reduction and cost
analysis (HRRCA). A HRRCA addresses seven requirements, all of which are addressed in this Economic
Analysis (EA) for the Ground Water Rule (GWR).
HRRCA Crosswalk Summary
HRRCA Requirement
Quantifiable and nonquantifiable health risk
reduction benefits
Quantifiable and nonquantifiable health risk
reduction benefits from co-occurring contaminants
Quantifiable and nonquantifiable costs
Incremental costs and benefits associated with
regulatory alternatives
Effects of the contaminants on the general
population and sensitive subpopulations
Increased health risk that may occur as a result of
compliance
Other relevant factors (quality and uncertainty of
information)
Addressed in Economic Analysis
Chapter 5 (All sections and exhibits)
Chapter 7 (Section 7.7.1 ; Exhibit 7.4)
Chapters (Sections 8.1; Exhibits 8.1-8.3, 8.5)
Chapter 5 (Section 5.4)
Chapters (All sections and exhibits)
Chapter 7 (Sections 7.2-7.8; Exhibits 7.1-7.7)
Chapter 8 (Sections 8.1 , 8.3; Exhibit 8.4, 8.6)
Chapters (Section 5.7; Exhibit 5.31)
Chapters (Sections 6.9; Exhibit 6.35)
Chapters (Section 8.4; Exhibits 8.7 - 8.13)
Chapters (Sections 5.2.2.2, 5.2.5.3, and 5.3.3)
Chapter 7 (Section 7.9)
Chapter 5 (Section 5.2.5.9)
Chapter 4 (Section 4.6; Exhibit 4. 32)
Chapter 5 (Section 5.6; Exhibit 5.28)
Chapter 6 (Sections 6.2 - 6.4, Section 6.7; Exhibit
6.32)
Chapter 8 (Section 8.2)
Economic Analysis for the
Final Ground Water Rule
xvi
October 2006
-------
Executive Summary
This Economic Analysis (EA) presents the evaluation of the benefits and costs of the Ground
Water Rule (GWR). The analysis is performed in compliance with Executive Order 12866, Regulatory
Planning and Review (58 FR 51735, September 1993), which requires the United States Environmental
Protection Agency (EPA or Agency) to estimate the economic impact of rules that have an annual effect
on the economy of over $100 million and make that analysis available to the public in conjunction with
publication of the final rule. Although EPA's analysis of the GWR has determined that its annual costs
are most likely below this threshold, EPA has chosen to publish a complete EA for this rule. Earlier, EPA
had prepared an EA (formerly known as the Regulatory Impact Analysis) to accompany the May 2000
proposed GWR.
EPA developed the GWR in collaboration with States and other interested stakeholders. The
primary goal of the GWR is to improve public health by identifying public ground water systems (GWSs)
that are susceptible to fecal contamination and to ensure that they take adequate measures to remove or
inactivate pathogens in drinking water they provide to the public.
ES.l Need for the Rule
An estimated 147,330 public water systems (PWSs) in the United States, serving over 114 million
people, use ground water as their primary water source. EPA is concerned about any potential adverse
health risks that may be associated with ground water sources and, in particular, the risks associated with
fecal contamination. Fecal contamination includes all of the bacteria and viruses—both pathogenic
(disease-causing) and non-pathogenic—found in feces. Under certain circumstances, these
microorganisms can make their way into ground water sources. Unlike for surface water sources, no
federal regulations currently require filtration or disinfection of ground water sources to remove microbial
contaminants. Currently for GWSs, there are only requirements for distribution system monitoring of
total coliforms, periodic sanitary surveys of small GWSs, and a maximum contaminant level (MCL) for
total coliforms.
The GWR will address the human risks of illness and death due to fecal contamination of ground
water and improve upon the protection provided by existing sanitary survey requirements for GWSs. The
reduction in risk will be accomplished through implementation of the GWR's risk-targeted approach.
Because of the difficulties involved in monitoring for the wide range of specific pathogenic bacteria and
viruses that could occur in ground water, one of the key provisions of the risk-targeted approach is
monitoring for a more easily measured bacterial or viral fecal indicator microorganism. Based on source
water sampling results, as well as sanitary survey results, PWSs will be required to take action to
minimize the possible presence of pathogenic bacteria and viruses that pose threats to human health.
Economic Analysis for the ES-1 October 2006
Final Ground Water Rule
-------
ES.2 Consideration of Regulatory Alternatives
EPA considered several regulatory alternatives when the GWR was proposed on May 10, 2000
(65 FR 30194). The proposed GWR and accompanying EA evaluated four regulatory alternatives, the
multi-barrier approach (Alternative 3) was the preferred alternative.
After the proposal, EPA considered comments and revised the occurrence analysis underlying the
cost-benefit analysis for the rule (as discussed in Chapter 4 of this EA) and made modifications to
Alternatives 2 and 3. This resulted in the choice of a different rule alternative, Alternative 2, termed the
risk-targeted approach, for the final GWR. .: The four alternatives considered for the final GWR and
analyzed in this EA are as follows:
Alternative 1 - Sanitary surveys and corrective action.
Alternative 2 - Risk-targeted approach: sanitary surveys, triggered monitoring, optional
assessment monitoring, corrective action, and compliance monitoring.
Alternative 3 - Multi-barrier approach: sanitary surveys, triggered monitoring, optional
hydrogeologic sensitivity assessment (HSA), assessment monitoring (a derivation of the proposed
routine monitoring), corrective action, and compliance monitoring.
Alternative 4 - Across-the-board disinfection.
As shown in Chapter 8, the risk-targeted approach is cost-effective (using either the Enhanced or the
Traditional cost-of-illness approach) and, considering the nonqualified benefits, the benefits justify the
costs. The final rule provides public health benefits while apportioning costs in a more flexible targeted
manner.
ES.3 Summary of the Final GWR Requirements
The GWR applies to all community and noncommunity PWSs that use ground water as a water
source. The final GWR targets GWSs that are susceptible to fecal contamination. A flowchart that
illustrates the compliance steps for systems is illustrated in Exhibit ES.l. The components of the risk-
targeted strategy are used to identify source contamination and significant deficiencies that may lead to
source contamination. Each component is described below. Exhibit ES.2 presents the schedule for these
activities.
Sanitary Surveys
The final GWR requires regular (every three years for CWSs and every five years for NCWSs)
comprehensive sanitary surveys of 8 critical components: (1) source; (2) treatment; (3) distribution
system; (4) finished water storage; (5) pumps, pump facilities, and controls; (6) monitoring and reporting,
and data verification; (7) system management and operation; and (8) operator compliance with State
requirements. The State may reduce the frequency of sanitary surveys for CWSs to at least once every
five years if the water system has an outstanding performance record as determined by the State (e.g., no
significant deficiencies documented in previous assessments and no history of total coliform MCL or
1 Modifications to Alternatives 2 and 3 since proposal are described in further detail in Chapter 3.
Economic Analysis for the ES-2 October 2006
Final Ground Water Rule
-------
monitoring violations under the TCR or the system maintains 4-log treatment of viruses using
inactivation, removal, or State-approved combination of virus inactivation and removal). If a significant
deficiency is identified, corrective action is required or a treatment technique violation is incurred.
Source Water Monitoring
In the final GWR, systems not achieving 4-log treatment of viruses (using inactivation, removal,
or a State-approved combination of these technologies) must conduct triggered source water monitoring
for the presence of at least one of the following State-specified fecal indicators: E. coll, enterococci, or
somatic coliphage. The triggered monitoring requirements apply to systems that are notified that a Total
Coliform Rule (TCR) routine sample is total coliform-positive. Within 24 hours of receiving the total
coliform-positive notice, GWSs must collect a source water sample and test it for the presence of the
State-specified fecal indicator.
If the State does not require corrective action (see Corrective Action section below) for the initial
fecal indicator-positive source water sample immediately, the system must collect five additional source
water samples within 24 hours of being notified of the initial fecal indicator-positive source water sample.
The GWR provides States with the option to require systems to conduct assessment source water
monitoring as needed and require systems to take corrective action. The purpose of this optional
assessment source water monitoring requirement is to target source water monitoring to systems that the
State determines are at higher risk for fecal contamination.
Corrective Action
The GWR requires a system with a significant deficiency or source water fecal contamination to
fix the problem by implementing a corrective action. The system must implement at least one of the
following corrective actions: correct all significant deficiencies; provide an alternate source of water;
eliminate the source of contamination; or provide treatment that reliably achieves at least 4-log treatment
of viruses. Furthermore, the system is required to notify the public served by the water system of any
uncorrected significant deficiencies and/or source water contamination. (The State may also require
notification of corrected significant deficiencies).
Compliance Monitoring
Compliance monitoring requirements are the final defense against microbial contaminants
provided by the final GWR. All GWSs that provide at least 4-log treatment of viruses using chemical
disinfection, membrane filtration, or a State-approved alternative treatment technology must conduct
compliance monitoring to demonstrate continual treatment effectiveness.
Economic Analysis for the ES-3 October 2006
Final Ground Water Rule
-------
Exhibit ES.1 Flowchart of Compliance with GWR Requirements for Systems
Sanitary surveys for all
ground water PWSs
All ground water PWSs «
Conduct TCR routine
monitoring
Does system
provide
treatment >4
logs?
Did survey
find
deficiencies?
Must correct deficiencies (4>
Continue GWR
compliance: sanitary
surveys, compliance
monitoring, TCR
compliance
Compliance momtorm
Is TCR routine
sample TC-positive?
(1)
(2)
(3)
(4)
Conduct triggered
monitoring ®
Continue GWR
compliance: sanitary
surveys, triggered
monitoring, TCR
compliance
Fecal Coliform
Indicator Positive?
Take non-treatment
corrective actions
NO
YES
Per State's direction, take
corrective action or take
5 additional samples (4)
Systems are not required
to take corrective action
if all 5 repeat samples are
fecal negative.
Includes mixed systems with ground water entry points
Treatment using inactivation, removal, or State-approved
combination of these technologies.
For those systems that do not receive a triggered monitoring
waiver from the State
The State may determine that the source of contamination has
been eliminated..
Did system
install
treatment as
corrective
action? <2
Economic Analysis for the
Final Ground Water Rule
ES-4
October 2006
-------
Exhibit ES.2 Implementation Timeline for the GWR
State
CWSs
NCWSs
YeaM
Year 2
YearS
Final GWR publication
date (2006); final rule
takes effect 3 years
after published in
Federal Register
Implementation
Primacy
Application
Year 4
Possible
Extension for
Primacy
Implementation
Implementation
YearS
Year 6
Year 7
YearS
Year 9
Year 10
Year 11
Conduct Sanitary Surveys
Review Triggered Source Water Monitoring
Review & Approve Corrective Action
Review Compliance Monitoring
1st round Sanitary Surveys
2nd round Sanitary Surveys
3rd round Sanitary
Surveys
Triggered Source Water Monitoring
Perform Corrective Action
Perform Compliance Monitoring
1st round Sanitary Surveys
2nd round Sanitary Surveys
Triggered Source Water Monitoring
Perform Corrective Action
Perform Compliance Monitoring
Economic Analysis for the
Final Ground Water Rule
ES-5
October 2006
-------
ES.4 Systems Subject to the GWR
Exhibit ES.3 shows the baseline number of systems and entry points subject to the rule and the
estimated number that will perform or undergo various rule activities (implementation, sanitary surveys,
triggered monitoring, corrective actions, and compliance monitoring). This baseline is derived from
EPA's Safe Drinking Water Information System (SDWIS) inventory, 4th quarter 2003 data.2 The systems
are subdivided by type [community water systems (CWS), nontransient noncommunity water systems
(NTNCWS), and transient noncommunity water systems (TNCWS)], and size (nine size categories).
All GWSs will undergo sanitary surveys (column A). EPA estimates that all GWSs will also
have to perform implementation activities (reading and understanding the rule, training, etc.) which is
also represented by column A. Column B presents the number of systems estimated to find and be
required to correct a significant deficiency based on conducting a sanitary survey. The remaining
activities are evaluated at the entry point into the distribution system. Column C indicates that all entry
points that do not achieve 4-log treatment of viruses will perform triggered source water monitoring for
the GWR. Corrective actions predicted as a result of triggered monitoring are presented in column D,
with estimates (cumulative) of the different levels of disinfection (i.e., increased disinfection or
installation of new disinfection) resulting from those corrective actions presented in columns E and F.
Finally, column G shows the number of entry points incurring costs for additional compliance monitoring
to ensure the effectiveness and reliability of treatment.3
2 SDWIS-Federal Version (SDWIS/FED) is a database created by EPA containing data submitted by States
and regions regarding compliance with SDWA.
3 The column G figures pertain only to entry points with newly installed treatment to achieve 4-log viral
inactivation or removal. The number of entry points performing non-treatment corrective actions can be calculated
by subtracting Column G figures from those presented in column F.
Economic Analysis for the ES-6 October 2006
Final Ground Water Rule
-------
Exhibit ES.3 Summary of Rule Implications
System Size
Systems
Receiving
Sanitary
Survey
A
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
100,001-1 Million
> 1 Million
12,843
14,358
4,649
5,910
2,884
1,444
167
103
3
Systems with
Corrective Actions
for Significant
Deficiencies
B
2,181
2,444
789
1,001
492
245
28
18
-
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
9,456
6,758
1,894
715
73
10
1
1
-
1,608
1,148
322
121
12
2
0
0
-
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
10,001-SOK
50.001-100K
100,001-1 Million
> 1 Million
64,448
18,993
1,940
585
74
19
1
1
-
10,990
3,234
329
99
13
3
0
0
-
Entry Points with
Triggered
Monitoring
c
12,797
14,819
5,578
8,910
5,638
4,357
1,295
749
-
8,609
6,149
1,724
651
66
9
1
1
-
63,295
18,648
1,905
574
73
19
1
1
-
Entry Points with
Corrective Actions
for Triggered
Monitoring
D
1,249
1,625
608
712
617
655
226
136
-
687
533
149
86
10
2
0
0
-
6,915
2,026
208
76
12
3
0
0
-
Entry Points with Viral
Disinfection Increased
from less than 4 logs
to 4 logs
E
358
917
360
396
353
548
93
94
-
150
119
33
19
2
0
0
0
-
1,143
337
35
12
2
1
0
0
-
Previously Non-
disinfecting Entry
Points Taking
Corrective Actions
F
891
709
248
317
264
107
133
42
-
537
415
117
67
8
1
0
0
-
5,772
1,689
174
63
10
3
0
0
-
Entry Points with
Incremental
Compliance
Monitoring
G
248
292
105
130
111
54
46
20
-
149
170
50
27
3
1
0
0
-
1,602
696
73
26
4
1
0
0
-
Sources: Exhibit 6.5b
Notes:
(G) indicates number of entry points with treatment corrective actions.
(F) - (G) indicates non treatment corrective actions.
Economic Analysis for the
Final Ground Water Rule
ES-7
October 2006
-------
ES.5 National Benefits and Costs of the GWR
EPA has determined from its analysis of the available human exposure and epidemiological
studies that the GWR will result in benefits in terms of reduced incidence of illnesses and deaths
associated with fecal contamination of ground water. There are substantial benefits attributable to the
GWR that are not quantified within this EA as part of the main analyses because of data limitations.
Beneficial aspects of the rule not quantified are characterized as either health benefits or non-health
benefits. Nonquantified health-related benefits include reducing other acute viral illness (other than those
caused by rotavirus and enterovirus), endemic acute bacterial illnesses and deaths, epidemic bacterial and
viral acute illness and death (associated with outbreaks, disinfection failures, and distribution system
contamination). Chronic illnesses, both bacterial and viral, are also not quantified. The rule will also
result in many non-health benefits such as reduced costs for responding to outbreaks, costs for averting
behavior, and reduced uncertainly regarding drinking water safety. Chapter 5, Section 5.4.3.2 presents a
discussion of nonqualified benefits and estimates a portion of their value, based only on bacterial
illnesses avoided, at four times the primary analysis benefits (resulting in total benefits that are five times
the primary benefits). This includes consideration of the value of deaths and hospitalization costs avoided
for waterborne cases of bacterial illness prevented by the rule. This does not include indirect (non-
medical) costs associated with waterborne bacterial illness or the value of avoiding other chronic or viral
illnesses.
Sections ES.5.1 and ES.5.2 summarize the methods used to derive the benefits and costs of the
rule. These sections describe the analyses of the total number of illnesses and deaths avoided, the
monetized benefits resulting from those cases and deaths avoided, a summary of nonqualified benefits,
and the total national costs (both one-time and annualized) for the GWR. The cases and monetized
benefits are based on reductions in microbial contamination that result from corrective actions. EPA's
national cost estimate includes rule implementation, sanitary surveys, triggered monitoring, corrective
actions (which account for more than half of the national costs), and compliance monitoring. Note that
two estimates for national costs and monetized benefits are presented depending upon the discount rate
used for present value calculations and annualizing costs.4 Benefits are also further divided into two
additional estimates based on different methodologies for estimating cost of illness (COI) attributed to
illnesses avoided. Chapters 4 through 6 and the appendices provide a more complete discussion of all the
analyses discussed in the sections below.
The GWR EA does not include estimates of costs and benefits of assessment source water
monitoring because it is an optional requirement. EPA does not know the extent to which States will use
the option or the manner in which they will implement it. This provision could potentially increase both
benefits and costs.
4 There is much discussion among economists of the proper social discount rate to use for policy analysis.
For GWR cost analyses, calculations are made using two social discount rates (3 and 7 percent) thought to best
represent current policy evaluation methodologies. Historically, the use of 3 percent is based on rates of return on
relatively risk-free investments, as described in the Guidelines for Preparing Economic Analyses (USEPA, 2000J).
The rate of 7 percent is a recommendation of the Office of Management and Budget (OMB) as an estimate of
"before-tax rate of return to incremental private investment" (USEPA, 1996b).
Economic Analysis for the ES-8 October 2006
Final Ground Water Rule
-------
ES.5.1 Derivation of Benefits
The GWR is expected to reduce the current incidence of acute and chronic illness caused by a
wide variety of viral and bacterial pathogens that are associated with fecal contamination of ground water.
Determining the economic value of the health benefits involves a sequential two-step process. First, a
risk assessment is prepared to quantify a health endpoint. Following this is the valuation of benefits.
Chapter 5 of this EA presents a complete discussion of these steps. Additional benefits accrue but are not
quantified.
EPA developed a risk assessment model to quantify a subset of the total number of predicted
illnesses and deaths avoided. Medical research has not isolated all waterborne pathogens, nor has it
thoroughly characterized the ability of these organisms to infect humans and cause illness and death.
Risk from bacterial pathogens was not characterized in the risk model because occurrence data for
bacterial pathogens in drinking water are limited (however, the discussion of nonqualified benefits
assesses the potential impact). For the risk assessment, EPA has selected two well-characterized viral
pathogens having different infectivity, morbidity, and mortality rates to represent a wide range of
waterborne viral pathogens to estimate the number of acute illnesses and deaths avoided by implementing
the GWR. The selected representative viruses are rotavirus and enteric viruses represented by echovirus
data.
Rotavirus is a highly infectious virus, but generally does not result in life-threatening illness. It is
similar to a large group of viruses that cause widespread, but usually not very serious, cases of
gastroenteritis. This large group includes noroviruses (i.e., Norwalk-like viruses), sapoviruses,
adenovirus, astrovirus and others. Echovirus is an enterovirus that is not highly infectious, but can cause
severe health effects if illness occurs (e.g., encephalitis or myocarditis). This economic analysis identifies
these two representative types as Type A and Type B viruses, respectively. EPA used these two viruses
to represent the range of possible acute illness risk that could result from all viruses.
The Agency, having identified representative viruses for modeling purposes, used probability of
virus occurrence in wells and samples along with estimated concentrations of enteric viruses to estimate
the population's exposure to Type A and Type B viruses. Combining these occurrence and concentration
data with daily drinking water intake data from the 1994-1996 USDA, Continuing Survey of Food Intakes
by Individuals (USEPA, 2000) yields estimates of risk of pathogen exposure. EPA combines these data
with viral dose-response functions and health effects data to estimate infections, morbidity, and mortality
to derive the baseline number of viral illnesses and deaths.
The baseline viral illnesses (185,186) and deaths (3.2) reflect conditions prior to the
implementation of the GWR. To determine the number of viral illnesses and deaths avoided by the final
GWR, EPA estimates the percentage reduction in viral occurrence as a result of systems implementing
changes to meet all rule requirements. The estimated numbers of avoided viral illnesses and deaths are
shown in Exhibit ES.4, and include the confidence bounds, reflecting uncertainty in those estimates.
Economic Analysis for the ES-9 October 2006
Final Ground Water Rule
-------
Exhibit ES.4 Summary of Annual Avoided Viral Illnesses and Deaths
by System Type
CWSs
NTNCWSs
TNCWSs
Total
Annual Illnesses Avoided
Mean
32,031
2,094
7,743
41,868
90% Confidence Bounds
5th
8,704
533
1,037
10,274
95th
68,994
4,308
14,738
88,039
Annual Deaths Avoided
Mean
0.62
0.03
0.09
0.74
90%Confidence Bounds
5th
0.07
0.00
0.01
0.08
95th
1.81
0.09
0.21
2.11
Note: Detail may not add due to independent forecasting. Values presented are average annual illnesses and
deaths avoided over the 25 year period of analysis following rule promulgation.
Source: Exhibit 8.1
The final step in the benefit calculation is to monetize the estimated reduction in cases by
applying economic values for avoided viral illness and deaths. The value of avoiding cases of viral
illness is based on estimates of the direct and indirect costs of becoming ill. Due to lack of adequate data
on the willingness-to-pay (WTP) to avoid becoming ill, EPA uses a COI estimate. EPA has chosen to
present two different estimates of COI, referred to in this EA as Enhanced and Traditional. Both
approaches include the value of the direct medical costs and of lost work time, but differ in the
assessment of value of lost nonmarket work time. The Enhanced COI values nonmarket work time based
on opportunity costs. The other approach, the Traditional COI, includes nonmarket (unpaid) work time
based on replacement costs. In addition, the Enhanced COI also includes the value of lost leisure time
and lost productivity—the reduced utility (or sense of well-being) associated with decreased enjoyment of
time spent in both market and nonmarket activities. For deaths due to viral infection, the Value of a
Statistical Life (VSL) is used to capture the value of benefits. The VSL represents an estimate of the
monetary value of reducing risks of premature death. The VSL, therefore, is not an estimate of the value
of saving a particular individual's life. Rather, the value of a "statistical" life represents the sum of the
values placed on small individual risk reductions across an exposed population. Other economic factors
are taken into consideration when calculating benefits over time, such as income growth, income
elasticity of demand, and social discount rates.
EPA estimates the quantified benefits of avoided illnesses and deaths from the GWRto be $8.6
million to $19.7 million depending on the discount rate and COI value used. These figures are presented
by system type in Exhibit ES.5.
Economic Analysis for the
Final Ground Water Rule
ES-10
October 2006
-------
Exhibit ES.5 Summary of Annualized Present Value Quantified Benefits
(SMillions, 2003$)
System Type
Annualized Benefits
at 3% Discount Rate
Mean
90% Confidence Bounds
5th
95th
Annualized Benefits
at 7% Discount Rate
Mean
90%Confidence Bounds
5th
95th
Enhanced COI
CWSs
NTNCWSs
TNCWSs
Total
$ 16.0
$ 0.9
$ 2.7
$ 19.7
$ 5.4
$ 0.3
$ 0.8
$ 6.5
$ 37.0
$ 2.2
$ 6.2
$ 45.4
$ 13.7
$ 0.8
$ 2.3
$ 16.8
$ 4.6
$ 0.2
$ 0.7
$ 5.5
$ 31.6
$ 1.8
$ 5.1
$ 38.6
Traditional COI
CWSs
NTNCWSs
TNCWSs
Total
$ 8.2
$ 0.5
$ 1.3
$ 10.0
$ 1.9
$ 0.1
$ 0.3
$ 2.2
$ 22.3
$ 1.3
$ 3.4
$ 27.0
$ 7.1
$ 0.4
$ 1.1
$ 8.6
$ 1.6
$ 0.1
$ 0.2
$ 1.9
$ 19.1
$ 1.0
$ 2.8
$ 22.9
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical costs and
lost work time (including some portion of unpaid household production). The Enhanced COI also factors in valuations for
lost personal time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI),
time with family, and recreation, and lost productivity at work on days when workers are ill but go to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unqualified benefits are
expected to comprise a significant portion of the overall benefits of the Rule and are presented in Section 5.4 of Ch. 5 of the
EA.
Source: Exhibit 5.23a, 5.23b
As mentioned above, there are substantial benefits attributable to the GWR that are not quantified
within this EA as part of the main analyses. These nonqualified benefits are shown in relation to
quantified benefits as part of the total benefits of the GWR in Exhibit ES.6. The nonquantified benefits
result from multiple factors. First, the quantified benefits are based on limited, well-defined data and key
assumptions that restrict the input parameters in the quantified benefit calculation. Typically, these
assumptions resulted in low mean values and narrow uncertainty ranges in the benefit analysis. This EA,
where applicable, discusses alternative assumptions. For example, the enterovirus morbidity fractions
are, by assumption, not determined using coxsackievirus (an enterovirus) data although the enterovirus
severity data use all enterovirus data. If coxsackievirus data were available, the mean morbidity values
would be greater. Choosing alternative values and ranges and differing key assumptions, which might
also be deemed reasonable, would increase the quantified benefits in this EA.
Second, the quantified benefits are based on data and assumptions that pertain to only partial
representation of Type A and Type B viruses potentially found in PWS wells with fecal contamination.
Due to limited available data, only rotavirus and some enterovirus data were used to calculate the
quantified benefits. As is more completely discussed in Section 5.4. other viruses as well as pathogenic
bacteria may contribute to the disease burden, both acute and chronic, associated with PWS wells with
fecal contamination. Most importantly, bacterial illnesses can result in more frequent and lengthier
hospitalization and more frequently have fatal outcomes. If bacterial diseases were considered in the
quantified benefits, the monetized benefits could be substantially greater because bacterial disease can be
more severe and can result in higher mortality rates.
Economic Analysis for the
Final Ground Water Rule
ES-11
October 2006
-------
Third, the quantified benefits are based on data and assumptions that limit the characterization of
acute disease. For rotavirus, only acute gastroenteritis illness and fatal dehydration associated with that
illness are monetized. Norovirus disease is not considered. For the enteroviruses, all acute disease
endpoints are considered, but the prevalence of severe endemic cases may be substantially diluted by the
large number of hand, foot, and mouth disease cases that are not likely to be waterborne. Thus, the
proportion of severe cases in the quantitative benefits is likely to be underestimated. As is discussed more
completely in Section 5.4, in neither instance, either for rotavirus or the enteroviruses, are chronic
diseases identified or monetized in the quantitative benefits calculation.
Fourth, the quantified benefits are based explicitly on what has been directly measured in PWS
wells, yet there is great difficulty in identifying and counting all infectious viral pathogens in dilute
drinking water samples. Indeed, some viral pathogens like infectious norovirus can never be identified in
any sample. Section 4.3.2 discusses these difficulties in more detail. Standard fecal indicator data such
as total coliforms and E. coll, commonly used to identify water treatment deficiencies and potential
human health hazards, are explicitly not used to determine human exposure for the purposes of
quantifying the benefits in this EA.
Fifth, the quantified benefits are assumed to be based only on one contamination scenario, fecal
contamination of source water. Other contamination scenarios are thoroughly documented in the ground
water contamination and outbreak scientific literature. However, these scenarios, such as inadequate
disinfection, are not explicitly considered in calculating the quantified benefits in this EA.
Sixth, the quantified benefits are assumed to be based only on avoidance of endemic disease. The
GWR will likely also decrease the incidence of epidemic disease (outbreaks). If epidemic illnesses and
the avoided non-health-related costs of ground waterborne disease outbreaks were included, the
quantified benefits would increase.
In summary, this EA quantifies a subset of the total health and non-health related benefits. In a
sample calculation, discussed in Section 5.4.3.2, EPA estimated that the total benefits could increase by a
factor of five by only accounting for additional deaths and hospitalizations caused by bacterial illness
being avoided. While EPA recognizes that this estimate includes substantial uncertainty, given all the
other nonquantified factors described above, EPA believes that the total benefits from the GWR are likely
to be more than five times those which have been quantified.
Economic Analysis for the ES-12 October 2006
Final Ground Water Rule
-------
Exhibit ES.6 Summary of Benefits of the GWR
Benefit
Category
Total Benefits
GWR EA Quantified Benefits
Health Benefits
Reduction in
endemic illness
incidence
Reduction in epi-
demic (outbreak)
illness incidence
Reduction in
treatment failures
viral exposure risk reduction (morbidity
and mortality)
bacterial exposure risk reduction
(morbidity and mortality)
chronic sequelae reduction
reduction in secondary transmission of
viral or bacterial illness from
symptomatic and asymptomatic
individuals
viral exposure risk reduction (morbidity
and mortality)
bacterial exposure risk reduction
(morbidity and mortality)
chronic sequelae reduction
reduction in secondary transmission of
viral or bacterial illness to susceptible
populations
Decreased illness through minimizing
treatment failures or fewer episodes
with inadequate treatment
acute rotavirus (Type A) illnesses and
deaths avoided
acute enterovirus (Type B) illnesses
and deaths avoided
reduction in secondary transmission of
viral illness from symptomatic
individuals
Not quantified
Not quantified
Non-Health Benefits
Outbreak
responses avoided
Avoided costs of
averting behavior
Increased
confidence
Avoided costs to affected water
systems, local governments (provision
of alternate water, issuing warnings
and alerts), and community (decreased
tourism due to bad press).
reduced need or perceived need to
use bottled water, point-of-use
devices, etc. (includes time and
material costs)
less time spent on averting behavior:
hauling/boiling water, etc.
Perceived reduction in risk associated
with perceived improvement in drinking
water quality
Not quantified
Not quantified
Not quantified
Economic Analysis for the
Final Ground Water Rule
ES-13
October 2006
-------
ES.5.2 Derivation of Costs
To estimate the total national costs of the GWR, EPA calculated the costs to be incurred by PWSs
and States for the rule activities. Cost analyses for PWSs include estimating the costs to implement the
rule, assist with sanitary surveys, perform triggered source water monitoring, undergo corrective actions,
and perform compliance monitoring. State cost analyses include estimates of the labor burdens that States
would face for implementation and other annual administrative tasks (e.g., recordkeeping), conducting
sanitary surveys, responding to PWS reports for source water and compliance monitoring, and reviewing
corrective action plans. The methodology for estimating corrective action and noncorrective action-
related costs for systems is discussed in the next several paragraphs, followed by a discussion of
uncertainties. A complete discussion of the cost analysis is provided in Chapter 6.
Noncorrective action costs for implementation, sanitary surveys, triggered monitoring, and
compliance monitoring are based on estimates of labor hours for performing these activities and on
additional laboratory costs. Some systems also incur capital costs for compliance monitoring. For all
noncorrective action cost calculations, EPA used the appropriate baseline (for systems or entry points)
shown in Exhibit ES.3.
Corrective action costs are based on unit cost estimates for a number of treatment technologies.
Technology unit cost estimates are in the form of "dollars per entry point" for initial capital and yearly
operations and maintenance (O&M) activities. Derivation of unit costs for a wide range of plant sizes,
represented by different design and average daily flow rates, is described in detail in the document,
Technology and Cost Document for the Final Ground Water Rule (USEPA, 2006b). EPA uses mean
population per system for each of the nine system size categories (derived from SDWIS) combined with
regression equations and entry point per system estimates to estimate mean design and average daily
flows per entry point. Technology unit costs per entry point are calculated using these mean flow values
for each of the nine system size categories. The technology unit costs are then combined with the
predicted number of entry points assumed to select each technology to produce national treatment cost
estimates.
EPA recognizes that systems vary with respect to many of the input parameters to the GWR Cost
Model (e.g., entry points per system, population served, flow per population, and labor rates). In many
cases, there is insufficient information to characterize fully the variability on a national scale. EPA
believes that the mean values for the various input parameters are adequate to generate EPA's best
estimate of national costs for the rule while still recognizing that impacts on specific systems may differ
substantially from these averages.
EPA also recognizes that there is uncertainty in the national cost estimates related to the national
average unit capital and O&M costs for the various technologies expected to be implemented in response
to the GWR. This uncertainty has been incorporated into the cost model using a Monte Carlo simulation.
The national costs of the GWR summarized below in Exhibits ES.7 and ES.8 show both the expected
values and the 90 percent confidence bounds on the national cost estimates obtained from the cost model
reflecting this uncertainty. Further discussion of uncertainty analysis is presented in Section 6.7.
Although EPA has quantified the significant costs of the GWR, there are some costs that the Agency
did not quantify. Overall, EPA believes that these nonquantified costs are much smaller than the
nonqualified benefits. These nonquantified costs result from uncertainties surrounding rule assumptions
and from modeling assumptions. Detailed discussion of nonquantified costs is presented in Section 6.6.
Economic Analysis for the ES-14 October 2006
Final Ground Water Rule
-------
Exhibit ES.7 Total Initial Capital and One-Time Costs ($Millions, 2003$)
CWS Total Initial Capital
NTNCWS Total Initial Capital
TNCWS Total Initial Capital
Total Initial PWS Capital Costs
CWS Start-Up Costs
NTNCWS Start-Up Costs
TNCWS Start-Up Costs
Total One-Time PWS Costs
State Start-Up Cost
Total State One-Time Costs
PWSsServing<10,000
Mean
Value
$ 79
$ 31
$ 173
$ 283
$ 5
$ 2
$ 9
$ 16
90 Percent
Confidence Bound
Lower
(5th % Me)
$ 29
$ 12
$ 64
$ 105
$ 5
$ 2
$ 9
$ 16
Upper
(95th %ile)
$ 158
$ 60
$ 339
$ 556
$ 5
$ 2
$ 9
$ 16
PWSs Serving > 10,000
Mean
Value
$ 62
$ 0
$ 1
$ 64
$ 0
$ 0
$ 0
$ 0
90 Percent
Confidence Bound
Lower
(5th % Me)
$ 24
$ 0
$ 0
$ 24
$ 0
$ 0
$ 0
$ 0
Upper
(95th %ile)
$ 117
$ 1
$ 2
$ 120
$ 0
$ 0
$ 0
$ 0
Total
Mean
Value
$ 141
$ 31
$ 174
$ 346
$ 5
$ 2
$ 9
$ 17
$ 13
$ 13
90 Percent
Confidence Bound
Lower
(5th % Me)
$ 53
$ 12
$ 64
$ 129
$ 5
$ 2
$ 9
$ 17
$ 13
$ 13
Upper
(95th %ile)
$ 275
$ 61
$ 341
$ 676
$ 5
$ 2
$ 9
$ 17
$ 13
$ 13
Notes: Detail may not add to totals due to independent rounding.
Source: Exhibit 6.30
Exhibit ES.8 Total Annualized Present Value Costs of the GWR ($Millions, 2003$)
Discount
Rate
3 percent
7 percent
Systems
Mean
Value
$ 50.0
$ 50.6
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 34.3
$ 35.2
Upper
(95th %ile)
$ 68.8
$ 69.0
States
Mean
Value
$ 11.8
$ 11.7
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 10.9
$ 10.9
Upper
(95th %ile)
$ 12.6
$ 12.6
Total
Mean
Value
$ 61.8
$ 62.3
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 45.2
$ 46.1
Upper
(95th %ile)
$ 81.4
$ 81.6
Notes: Detail may not add to totals due to independent rounding.
Source: Exhibit 6.33
ES.6 Projected Impacts on Household Costs
The household cost analysis assumes that systems may pass some or all costs of a new regulation on
to their consumers in the form of rate increases. Exhibit ES.9 presents estimated annual household cost
increases. Only CWSs are included in this analysis because they are the only systems that serve
households directly. The top half of the exhibit shows summary statistics for all households affected by
the rule, including households that will incur minimal costs (e.g., those served by systems incurring only
implementation costs). The bottom half shows statistics only for households served by entry points
adding treatment technologies to comply with the rule (see Exhibit ES.3 for estimates of entry points
subject to corrective actions). Because treatment changes represent the majority of rule costs, this
provides insight into how the rule will affect the segment of the population most impacted by the rule.
As shown in Exhibit ES.9, mean annual household costs based on all GWSs (including those that do
not add treatment) range from $0.21 to $16.54, depending on system size. Mean household costs
reflecting the subset of GWSs that undertake corrective actions range from $0.45 to $52.38, depending on
system size. EPA estimates that, as a whole, households subject to the GWR face minimal increases in
their annual costs. Approximately 66 percent of the households potentially subject to the rule are served
by systems serving at least 10,000 people; these systems experience the lowest increases in costs due to
Economic Analysis for the
Final Ground Water Rule
ES-15
October 2006
-------
significant economies of scale. Households served by small systems that undertake corrective actions
will face the greatest increases in annual costs.
Exhibit ES.9 Summary of Annual Per-Household Costs for the GWR (2003$Year)
Systems Size
(Population
Served)
Households
Mean
Median
90th
Percentile
All Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
>1 00,000
Total
289,222
1,303,890
1,278,081
4,196,105
6,271,380
11,468,813
4,204,584
9,755,817
38,767,890
$ 16.54
$ 3.51
$ 0.97
$ 0.37
$ 0.27
$ 0.21
$ 0.34
$ 0.21
$ 0.51
$ 2.81
$ 0.64
$ 0.16
$ 0.04
$ 0.03
$ 0.04
$ 0.10
$ 0.04
$ 0.09
$ 9.31
$ 6.11
$ 1.70
$ 0.61
$ 0.43
$ 0.49
$ 1.02
$ 0.62
$ 0.88
Corrective Action Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
>1 00,000
Total
70,563
312,484
302,557
919,133
1,487,159
2,871,250
1,215,544
2,283,144
9,461,833
$ 52.38
$ 12.00
$ 3.23
$ 1.33
$ 0.80
$ 0.45
$ 0.53
$ 0.68
$ 1.51
$ 18.99
$ 4.52
$ 1.33
$ 0.47
$ 0.25
$ 0.18
$ 0.26
$ 0.39
$ 0.60
$ 82.21
$ 25.76
$ 6.56
$ 2.59
$ 2.18
$ 1.18
$ 1.36
$ 1.65
$ 3.20
Source: GWR model output.
ES.7 Comparison of Benefits and Costs, and of Regulatory Alternatives of the GWR
Exhibit ES. 10 compares estimated quantified benefits with estimated costs. Based on the
comparison of these values, the estimated quantified benefits of the rule range from approximately 14% to
32% of the costs, depending on the discount rate and COI approach. The estimated quantified benefits for
the Enhanced COI approach are greater than the corresponding estimated benefits for the Traditional COI
approach. The quantified estimate of the benefits significantly understates the true benefit of the rule. As
discussed in Section ES.5.1 and Exhibit ES.6, the nonqualified health and non-health benefits far exceed
those that EPA was able to quantify.
Economic Analysis for the
Final Ground Water Rule
ES-16
October 2006
-------
Exhibit ES.10 Estimated Annualized National Benefits and Costs for the GWR
(SMillions, 2003$)
System Type
3% Discount Rate
Mean
90% Confidence Bounds
5th
Benefits
Costs
Net Benefits
$ 19.7
$ 61.8
$ (42.1)
$ 6.5
$ 45.2
Note 1
Benefits
Costs
Net Benefits
Nonquantified
Benefits
Nonquantified
Costs
$ 10.0
$ 61.8
$ (51.8)
$ 2.2
$ 45.2
Note 1
95th
7% Discount Rate
90
Mean 5
Enhanced COI
$ 45.4
$ 81.4
Note 1
$ 16.8 $
$ 62.3 $
$ (45.5) Nc
Traditional COI
$ 27.0
$ 81.4
Notel
$ 8.6 $
$ 62.3 $
$ (53.7) Nc
'/oConfidence Bounds
th 95th
5.5 $ 38.6
46.1 $ 81.6
te 1 Note 1
1.9 $ 22.9
46.1 $ 81.6
te 1 Note 1
Decreased incidence of other acute viral disease endpoints
Decreased incidence of bacterial illness and death
Decreased incidence of chronic bacterial and viral illness sequelae
Decreased incidence of waterborne disease outbreaks and epidemic illness
Decreased illness through minimizing treatment failures or fewer episodes with inadequate treatment
Decreased use of bottle water and point-of-use devices (material costs)
Decreased time spent on averting behavior
Avoided costs associated with outbreak response
Perceived improvement in drinking water quality and reduction in risk associated with ingestion
Benefits from optional Assessment Monitoring
Benefits from correction of sanitary survey deficiencies identified in the distribution systems and treatment
plant
Costs for optional Assessment Monitoring
Costs from correction of sanitary survey deficiencies identified in the distribution systems and
treatment plant
Costs for compliance monitoring for some systems that already disinfect
Some land costs depending on the treatment technology
Cost for five repeat samples but this is small compared to the overestimate of cost for the initial
fecal-indicator sample that systems would take.
Note 1: Because benefits and costs are calculated using different model modules, bounds are not calculated on net benefits.
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical costs and lost work time (including
some portion of unpaid household production). The Enhanced COI also factors in valuations for lost personal time (non-work time) such as childcare and
homemaking (to the extent not covered by the Traditional COI), time with family, and recreation, and lost productivity at work on days when workers are ill
but go to work anyway.
Source: Exhibit 8.5a
The Agency also performed a number of other analyses related to the final rule. This process
included an analysis of net benefits, as well as cost effectiveness and efficiency analyses. In addition, the
Agency performed a number of comparisons among the four regulatory alternatives that are described in
more detail in Chapter 8. The following is a summary of these analyses.
Economic Analysis for the
Final Ground Water Rule
ES-17
October 2006
-------
The GWR likely passes economic threshold criteria:
" The GWR has positive net benefits when both quantified and nonqualified benefits are
considered. For the Enhanced COI approach, the quantified benefits alone are approximately
27 to 32 percent of the costs of the GWR depending on the discount rate. For the Traditional
COI approach, the quantified benefits are approximately 14 to 16 percent of the costs of the
GWR. Considering that nonqualified benefits are expected to be significantly larger than
the quantified benefits, it appears likely that the final GWR would have positive net benefits
regardless of the discount rate or cost of illness approach used. Section 5.4.3.2 presents a
discussion of nonqualified benefits and estimates a portion of their value, based only on
bacterial illnesses avoided, at four times the primary analysis benefits (resulting in total
benefits that are five times the primary benefits). This includes consideration of the value of
deaths and hospitalization costs avoided for ground water borne cases of bacterial illness
prevented by the rule. Including only these estimated bacterial illness and death benefits, the
total net benefits of the GWR would be positive using the Enhanced COI approach. Total net
benefits would still be slightly negative using the Traditional COI approach, however, other
nonqualified benefits such as indirect (non-medical) costs associated with waterborne
bacterial illness or the value of avoiding other chronic (either bacterial or viral) or other viral
illnesses (not accounted for in this analysis) would most likely make this value positive.
•• The number of illnesses that must be avoided to break even with costs is well above the
estimated number of viral cases avoided, but is most likely within the bounds of cases
avoided once nonquantified cases (both bacterial and viral) are considered. The number of
deaths that must be avoided to break even, while outside the bounds of the quantitative
analysis, is small in absolute terms. Consideration of all nonquantified benefits is predicted
to result in favorable break even results.
•• The GWR is cost-effective (using either the Enhanced or the Traditional COI approach): no
other alternative achieves greater benefits at the same cost or the same benefits at lower cost.
Final GWR determinations:
" The economic analysis for this rule, considering quantified and nonquantified benefits,
supports the basis for selecting the final GWR over other alternatives. However, the
distinction between Alternative 2 and 3 on an economic basis, is not great.
•• EPA chose the final GWR because EPA believes it is more flexible, targeted, and cost-
effectively protective than Alternative 3. Optional assessment monitoring allows States to
most effectively target those systems at greatest risk and minimize unnecessary monitoring.
EPA took the following considerations into account in making this judgment:
1) Under Alternative 3, some States may not be able to conduct HSAs and thereby require
systems in nonsensitive aquifers to conduct assessment monitoring unnecessarily. For
systems not at risk this additional monitoring would provide no benefit.
2) Systems with frequent TC positives in the distributions system (and subsequent frequent
triggered monitoring) would benefit little from assessment monitoring regardless if they
were located in sensitive aquifers or not because the source water would already be
thoroughly evaluated. Under Alternative 3, such systems in sensitive (or undetermined)
aquifers would be required to do assessment monitoring.
Economic Analysis for the ES-18 October 2006
Final Ground Water Rule
-------
3) Systems identified as having significant risk factors pertaining to potential fecal
contamination at their source (e.g., aquifer condition, well characteristics, proximity to
sewage or septic), but infrequent triggered monitoring source water samples, would
benefit from assessment monitoring. States will be able to identify such systems on an
ongoing basis through a variety of tools and information readily available to them.
The EPA believes that the final rule is a logical outgrowth of the proposed rule, that it is
supported by comments, and that it provides public health benefits while apportioning costs
in a more flexible targeted manner.
ES.8 Conclusions
Pursuing its mandate to protect public health, EPA has promulgated the GWR to reduce the risk
that microbial contamination of ground water poses to consumers of drinking water. Over 114 million
people in the United States use ground water as a source of drinking water. It is, however, a largely
unprotected source, with considerable risk of fecal contamination. This contamination includes both viral
and bacterial pathogens causing illnesses of varying severity. EPA has determined that the final GWR
will provide important protection against illnesses and deaths attributable to ground water contamination.
EPA also believes that the GWR will provide this desired protection from ground water pathogen
contamination at a justifiable cost.
Economic Analysis for the ES-19 October 2006
Final Ground Water Rule
-------
1. Introduction
This Economic Analysis (EA) presents the evaluation of the benefits and costs of the Ground
Water Rule (GWR). The analysis is performed in compliance with Executive Order 12866, Regulatory
Planning and Review (58 FR 51735), which requires the United States Environmental Protection Agency
(EPA or Agency) to estimate the economic impact of rules that have an annual effect on the economy of
over $100 million and make that analysis available to the public in conjunction with publication of the
final rule. Although EPA's analysis of the GWR has determined that its annual costs are most likely
below this threshold, EPA has chosen to publish a complete EA for this rule. EPA also prepared an EA
(formerly known as the Regulatory Impact Analysis) that accompanied the May 2000 proposed GWR.
EPA developed the GWR in collaboration with States and other interested stakeholders. The
primary goal of the GWR is to improve public health by identifying public ground water systems (GWSs)
that are susceptible to fecal contamination and to ensure that they take adequate measures to remove or
inactivate pathogens in drinking water they provide to the public.
This chapter provides a summary of the GWR in section 1.1. Section 1.2 outlines the
organization of this EA, and section 1.3 provides information regarding supporting calculations and
citations.
1.1 Summary of the Ground Water Rule
The GWR applies to all community and noncommunity public water systems (PWSs) that serve
ground water (referred to as GWSs in this document) as a water source, including mixed systems with any
ground water entry points to distribution systems. The GWR does not apply to ground water determined
by the State to be under the direct influence of surface water, nor does the rule apply to public water
systems that combine all of their ground water with surface water prior to treatment. These systems are
already regulated under surface water treatment rules.
The Risk-Targeted Approach of the final GWR targets ground water systems that are susceptible
to fecal contamination and requires corrective action. Key components of the strategy are:
1. Sanitary surveys and corrective action,
2. Triggered source water monitoring,
3. Corrective actions, and
4. Compliance monitoring.
In addition to the mandatory rule components listed above, the GWR provides a mechanism for
States to adopt an optional assessment source water monitoring provision, hereafter referred to as
assessment monitoring. See Chapter 3 (section 3.4.2) of this EA for details.
Economic Analysis for the 1-1 October 2006
Final Ground Water Rule
-------
1.2 Document Organization
The remainder of this EA is organized into the following chapters:
Chapter 2 summarizes the technical, regulatory, and public health issues addressed by the rule
and provides an overview of National Primary Drinking Water Regulations (NPDWRs)
relevant to the GWR. It also explains the statutory authority for the GWR and the economic
rationale for the regulatory approach.
• Chapter 3 reviews alternative approaches EPA considered during the development of the rule
and presents the rationale for selecting the final rule requirements.
Chapter 4 characterizes conditions that exist (including system inventory, treatment, and
water quality data) before systems make changes to meet the GWR requirements.
• Chapter 5 presents a summary of the risk assessment performed to estimate the public health
benefits of the GWR. The economic benefits of the rule are also presented. The benefits of
other regulatory alternatives considered are compared.
• Chapter 6 presents an estimate of the costs of implementing the rule to industry, households,
and States. The costs of other regulatory alternatives considered are compared.
Chapter 7 discusses distributional analyses performed to evaluate the effects of the rule on
different segments of the population, and considers various executive orders and
requirements, including the Regulatory Flexibility Act (RFA) and Unfunded Mandates
Reform Act (UMRA).
• Chapter 8 compares the rule's benefits and costs to evaluate the potential net benefits and
cost-effectiveness of the rule. The results are discussed and compared to other regulatory
alternatives considered.
1.3 Calculations and Citations
This EA presents results from detailed and complex analyses. To help the reader track the
various calculations and analyses, the following are provided:
• A reference section.
• Appendices.
- Appendix A provides additional discussion of the derivation of the cost of illness values
used in the benefits model.
- Appendix B presents detailed cost of illness and value of statistical life estimates by
illness and year as used in the benefits model.
- Appendix C presents detailed benefits estimates by illness and system type.
— Appendix D presents detailed cost estimates by systems size and type.
Economic Analysis for the 1-2 October 2006
Final Ground Water Rule
-------
— Appendix E discusses the potential implications of using a population dynamic modeling
to estimate secondary spread in the benefits model.
— Appendix F describes analyses conducted to select model forms and estimate model
parameters for infectivity dose response relationships.
— Appendix G provides summary flowcharts for the baseline risk and benefits models.
- Appendix H presents the detailed cost effectiveness analysis of the rule alternatives using
a quality-adjusted life years approach.
- Appendix I provides detailed analysis of total coliform hit rates in ground water systems.
— Appendix J discusses changes in the cost and benefits modeling approaches between the
proposed and final rules.
— Appendix K discusses costing detail for regulatory alternatives not presented in the main
text.
— Appendix L provides summary flowcharts for the cost model.
Exhibits. Most tabular exhibits include a row that provides the formulas used to compute the
contents of each column.
Sources for information used, but not calculated within the exhibits.
Supporting electronic file outputs (i.e., GWR cost and benefits model outputs).
Flowcharts that illustrate methodologies of analyses as well as rule requirements.
Economic Analysis for the 1-3 October 2006
Final Ground Water Rule
-------
2. Need for the Rule
2.1 Introduction
The United States Environmental Protection Agency (EPA or Agency) is promulgating the
Ground Water Rule (GWR) to address microbial contamination of ground water-supplied drinking water
systems in accordance with the Safe Drinking Water Act (SDWA) of 1974, as amended in 1986 and again
in 1996. The 1986 SDWA Amendments directed EPA to establish National Primary Drinking Water
Regulations (NPDWRs) requiring disinfection for the inactivation of microbiological contaminants for all
public water systems (PWSs), including systems supplied by ground water sources. The 1996
Amendments included more specific language regarding ground water disinfection, specifying that the
Administrator must publish NPDWRs requiring disinfection as a treatment technique for all ground water
PWSs only "as necessary."
This chapter summarizes the technical, regulatory, and public health issues addressed by the rule
and provides an overview of other NPDWRs relevant to the GWR. It also explains the statutory authority
for promulgating the GWR and the economic rationale for choosing a regulatory approach for
implementing the rule.
2.1.1 Description of the Issue
An estimated 147,330 PWSs in the United States, serving over 114 million people, use ground
water as their primary water source. EPA is concerned about any potential adverse health risks that may
be associated with ground water sources and, in particular, the risks associated with fecal contamination.
Fecal contamination includes all of the bacteria and viruses—both pathogenic (disease-causing) and
nonpathogenic—found in feces. Under certain circumstances, these can make their way into ground
water sources. Unlike surface water sources, no federal regulations currently require filtration or
disinfection of ground water sources to remove microbial contaminants. There are, however, existing
requirements for distribution system monitoring and periodic inspection of ground water systems
(GWSs).
Monitoring for specific pathogenic bacteria and viruses is often difficult. Many methods for
monitoring pathogenic bacteria and viruses are not reliable or as sensitive as those for nonpathogenic
indicators of fecal contamination. Additionally, since pathogenic bacteria and viruses are shed in low
numbers by a few infected individuals and for a limited time, their numbers are low and difficult to detect
through periodic sampling compared to detection of nonpathogenic fecal microorganisms. Therefore,
because of their widespread presence in fecal material, nonpathogenic fecal organisms are often used as
indicators of fecal contamination. These indicators include strains of Escherichia coli (E. coli),
coliphage, coliform, and other bacteria. Coliphage are bacteriophages (viruses that infect bacteria) that
primarily infect E. coli. Coliforms include many bacteria that are free-living in the environment as well
as fecal coliform (bacteria more commonly found in human feces). Other bacteria that are used as
indicators of fecal contamination include fecal streptococci (enterococci) and Clostridium perfringens, a
spore-forming anaerobic organism that can persist for long periods of time in the environment.
The basic public health issue the GWR addresses is that people served by ground water sources
may face an increased risk of illness due to fecal contamination of those sources. EPA has evaluated data
on outbreaks and the occurrence of waterborne pathogens and indicators of fecal contamination in ground
water supplying PWS wells. These data indicate that there is a subset of GWSs that are susceptible to
Economic Analysis for the 2-1 October 2006
Final Ground Water Rule
-------
fecal contamination where risk management strategies are needed to protect public health. Specifically,
the Centers for Disease Control and Prevention (CDC) reports that between 1991, the year in which the
TCR became effective, and 2000, GWSs were associated with 68 outbreaks that caused 10,926 illnesses
(Lee et al., 2002). These accounted for 51 percent of all waterborne disease outbreaks in the United
States. The major deficiencies identified by the CDC report were source water contamination and
inadequate treatment (or treatment failures). Distribution system deficiencies were also important.
Studies of pathogen and/or fecal indicator occurrence in ground waters that supply PWSs show that a
subset of PWSs utilize contaminated wells. Using cell culture methods, one large national survey of 448
wells in 35 States found five percent of the wells positive for enteroviruses, 15 percent of the wells
positive for bacterial indicators, and 20 percent of the wells positive for viral indicators. Using another
method designed to detect viral DNA or RNA, this same study found over 30 percent of the wells positive
for viruses.
Based on outbreak and occurrence data, inadequate sanitary surveys, difficulties in monitoring
directly for pathogenic bacteria and viruses, and uncorrected system deficiencies, EPA has concluded that
PWSs need to implement targeted risk management strategies to protect public health. Consequently,
EPA is promulgating requirements that provide a flexible, risk-based approach to achieve public health
protection. The final GWR builds on existing State programs - some which emphasize the importance of
monitoring and treatment and others which emphasize inspections and technical assistance - to identify
susceptible GWSs. PWSs will be required to take action to minimize the presence of pathogenic bacteria
and viruses, as necessary, based on the results of sampling or sanitary surveys. The GWR establishes
treatment technique requirements that allow multiple options to address significant deficiencies and fecal
contamination. Furthermore, the final GWR establishes compliance monitoring requirements to ensure
that treatment effectiveness is maintained.
2.2 Public Health Concerns to Be Addressed
EPA's primary mission is to protect human health and the environment. The GWR requirements
are intended to achieve this mission with regard to fecal contamination of drinking water drawn from
ground water sources. This section describes the potential adverse health effects associated with
consuming fecally contaminated ground water.
2.2.1 Contaminants and Their Health Effects
Pathogenic enteric viral and bacterial microorganisms are excreted in the feces of infected
humans and animals. The word enteric (relating to the intestines or, more specifically, the human gut)
indicates that the natural habitat of these microorganisms is the intestinal tract of animals and humans.
Enteric microorganisms, sometimes referred to as intestinal microflora, can survive in sewage and
leachate derived from septic tanks (septage). When sewage and septage are released into the
environment, they are sources of intestinal microflora and potential sources of viral and bacterial
pathogens. Once in the environment, fecal matter from infected humans or animals may make its way
into ground water sources. If an enteric pathogen is ingested, the likelihood of infection varies depending
on the number and pathogenicity of the organism. The likelihood and severity of symptomatic illness
also vary with the type of pathogen, the level of acquired immunity, and the general resistance of the
person who is exposed.
Examples of common fecal viral pathogens include enteroviruses (e.g., echoviruses and
coxsackieviruses), rotavirus, and hepatitis A virus (HAV). Viruses cannot reproduce outside of a host,
Economic Analysis for the 2-2 October 2006
Final Ground Water Rule
-------
although they can survive and remain infectious. Also, with a few exceptions, viruses that can infect
human cells typically cannot infect the cells of other animals, and vice versa. Viruses that infect gut cells
become capable of reproducing when they infect humans. Once infected, humans shed viruses in stools,
typically for a period lasting between a few weeks to a few months. Thus, regardless of whether
individuals infected by the waterborne pathogen have actual symptoms of illness (such as diarrhea), they
still shed the virus, which may infect other people. This is called secondary spread, and it can result from
person-to-person contact or contact with contaminated surfaces. As a result, viral pathogens may infect
others via a variety of routes.
Some enteric viruses may infect cells in tissues outside the gut causing mild or serious secondary
effects ("sequelae") such as myocarditis, conjunctivitis, meningitis, or hepatitis. There is also increasing
evidence that the human body reacts to foreign invasion by viruses in ways that may also be detrimental.
For example, one hypothesis for the cause of adult onset (Type 2) diabetes is that the human body,
responding to coxsackie B5 virus infection, attacks both pathogenic cells and healthy pancreas cells in an
autoimmune reaction because of similarities between the two (Solimena and De Camilli, 1995).
Examples of illnesses caused by known or suspected waterborne fecal viral pathogens are shown
in Exhibit 2.1.
Examples of common fecal bacterial pathogens include E. coll, Salmonella, Shigella, and
Campylobacter jejuni. Some waterborne bacterial pathogens cause disease by rapid growth and
dissemination (e.g., Salmonella) while others primarily cause disease via toxin production (e.g., Shigella,
E. coli O157, Campylobacter jejuni}. Campylobacter jejuni, E. coli, and Salmonella have a host range
that includes both animals and humans; Shigella is associated only with humans (Geldreich 1996).
Unlike viruses, bacteria are able to reproduce outside of the host.
Most of the waterborne bacterial pathogens cause gastrointestinal illness, but some can cause
other severe illnesses as well. For example, Legionella causes Legionnaires Disease, a form of
pneumonia that has a fatality rate of about 15 percent. It can also cause Pontiac Fever, which is a milder
respiratory infection form of Legionnaires Disease. Several strains of E. coli can cause severe disease,
including kidney failure.
Some bacterial pathogens are opportunistic (i.e., they are only infectious in the presence of
another, preexisting condition or weakness). Opportunistic pathogens usually cause illness only in
immunocompromised persons or in other sensitive subpopulations, such as the very young or the elderly.
Other pathogens, such as Salmonella, Shigella, and Campylobacter jejuni, are not entirely opportunistic
but result in certain diseases with greater frequency and severity in immunocompromised persons (Framm
and Soave, 1997).
Examples of illnesses caused by major waterborne bacterial pathogens are shown in Exhibit 2.2.
Economic Analysis for the 2-3 October 2006
Final Ground Water Rule
-------
Exhibit 2.1 Examples of Illnesses Caused by Known or Suspected Waterborne
Fecal Viral Pathogens
Enteric Virus
Poliovirus
Coxsackievirus A
Coxsackievirus B
Echovirus
Norovirus
Hepatitis A virus
Hepatitis E virus
Rota virus
Enteric Adenovirus
Astrovirus
Illness
Paralysis
Meningitis, fever, respiratory disease
Myocarditis, meningitis, pleurodynia, eye
infections, congenital heart disease, rash, fever,
encephalitis, associated with diabetes
Meningitis, rash, fever, gastroenteritis,
encephalitis, flaccid paralysis
Gastroenteritis
Hepatitis
Hepatitis
Gastroenteritis
Eye infections, gastroenteritis, respiratory disease
Gastroenteritis
Note: Bold highlights indicate diseases directly caused by the enteric virus; other illnesses represent secondary
effects ("sequelae").
Source: Adapted from Irving et al. 1996, Salvato 1992.
Exhibit 2.2 Examples of Illnesses Caused by Common Waterborne Bacterial
Pathogens
Bacterial Pathogen
Campylobacterjejuni
Shigella species
Salmonella species
Vibrio cholerae
Escherichia coli (several species, including E. coli
O157:H7)
Yersinia enterocolitica
Legionella species
Illness
Gastroenteritis, meningitis, associated with
reactive arthritis and Guillain-Barre paralysis
Gastroenteritis, dysentery, hemolytic uremic
syndrome, convulsions in young children,
associated with Reiters Disease (reactive
arthropathy)
Gastroenteritis, septicemia, anorexia, arthritis,
cholecystitis, meningitis, pericarditis, pneumonia,
typhoid fever
Cholera (dehydration and kidney failure)
Gastroenteritis, hemolytic uremic syndrome
(kidney failure)
Gastroenteritis, acute mesenteric lymphadenitis,
joint pain
Legionnaires Disease, Pontiac Fever
Source: Adapted from Irving et al. 1996, Salvato 1992.
Economic Analysis for the
Final Ground Water Rule
2-4
October 2006
-------
EPA and Centers for Disease Control and Prevention (CDC) data provide an indication of the
types of pathogens that have caused waterborne disease outbreaks. Exhibit 2.3 identifies the etiology of
bacterial and viral waterborne outbreaks in GWSs reported to the CDC from 1991 through 2000. Of the
68 outbreaks in GWSs, 14 (21 percent) were associated with specific bacterial pathogens. The fecal
bacterial pathogen, Shigella, caused more reported outbreaks (five-seven percent) than any other single
agent. Identified viral pathogens were associated with four (six percent) reported outbreaks. Etiologic
agents were not identified in 39 (57 percent) outbreaks; however, EPA suspects that many of these were
caused by viruses, given that it is generally more difficult to analyze for viral pathogens than bacterial
pathogens.
Exhibit 2.3 Etiology of Waterborne Outbreaks in Ground Water Systems,
1991-2000
Causative Agent
Protozoa
Giardia
Cryptosporidium
Virus
Hepatitis A
Norwalk Virus
Bacteria
Shigella
Campy lob acter
Salmonella, non-typhoid
S. typhimurium
E. coli
Vibrio
Undetermined
Total
CWSs
Outbreaks
8
5
3
-
-
-
6
1
1
1
1
1
1
5
19
Cases of
Illness
1,675
136
1,539
-
-
-
1,037
83
172
625
124
22
11
65
2,777
Percent of
Total
Outbreaks
42.1%
26.3%
15.8%
0.0%
0.0%
0.0%
31.6%
5.3%
5.3%
5.3%
5.3%
5.3%
5.3%
26.3%
100.0%
NCWSs
Outbreaks
3
2
1
4
-
4
8
4
2
-
-
2
-
34
49
Cases of
Illness
576
25
551
1,806
-
1,806
1,309
473
51
-
-
785
-
4,458
8,149
Percent of
Total
Outbreaks
6.1%
4.1%
2.0%
8.2%
0.0%
8.2%
16.3%
8.2%
4.1%
0.0%
0.0%
4.1%
0.0%
69.4%
100.0%
TOTAL
Outbreaks
11
7
4
4
-
4
14
5
3
1
1
3
1
39
68
Cases of
Illness
2,251
161
2,090
1,806
-
1,806
2,346
556
223
625
124
807
11
4,523
10,926
Percent of
Total
Outbreaks
16.2%
10.3%
5.9%
5.9%
0.0%
5.9%
20.6%
7.4%
4.4%
1.5%
1.5%
4.4%
1.5%
57.4%
100.0%
Note: Detail may not add to totals due to independent rounding.
Sources: Compiled from CDC 1993, Kramer et al. 1996, Levy et al. 1998, Berwick et al. 2000, and Lee et al. 2002.
2.2.2 Sources of Contaminants
As discussed in section 2.2.1, water from ground water sources can contain microbial
contaminants. Fecal contamination of ground water can occur by several routes, including through the
subsurface, wellhead, or the distribution system. This section discusses these potential sources of viral
and bacterial fecal contamination of ground water supplies. Because pathogens are associated with
human and animal waste, the following sections discuss fecal contamination in general and not
necessarily the specific types of microbes associated with each source.
Economic Analysis for the
Final Ground Water Rule
2-5
October 2006
-------
2.2.2.1 Ground Water Contamination through the Subsurface
Many factors control the fate and transport of viruses and bacteria in subsurface media. Because
these factors are often interrelated, defining the processes involved in the survival and migration of
viruses and bacteria is a complex task. Factors such as pH, hydrogeologic conditions, soil types,
inorganic ion content, organic matter content, microbial activity, moisture content, and type of pathogen
all affect pathogenic fate and transport. Other factors, such as climatic changes and agriculture and land
use practices, influence, and may alter, the complex soil environment. For example, wetter climatic
conditions may result in high water tables, thereby potentially reducing the distance and time required for
contaminants to enter the now-shallower aquifers. In addition, sewage and sludge application to land may
alter the physical and chemical properties of soils and affect their capacity to impact viral migration and
survival (Bitton and Gerba, 1984). These factors are likely to have a direct or indirect effect on pathogen
survival.1
Frequently, the subsurface conditions provide adequate natural attenuation of microbial
contaminants to ensure protection of the source water. However, certain hydrogeologic features make
aquifers more sensitive to microbial contamination. For example, karst, fractured bedrock, and gravel
aquifers are considered sensitive aquifers. In these hydrogeologic settings, contaminants that are
introduced into the environment are more likely to reach the drinking water production well than in
localities with greater natural attenuation capabilities (e.g., sand aquifers).
Given the right conditions, fecal contamination from a variety of sources can reach an aquifer.
Normal septage and sewage practices release great amounts of human waste into the subsurface. Canter
and Knox (1984) estimated the volume of septic tank waste that is released into the subsurface in the
United States to be one trillion gallons per year. Contaminants from failed septic systems or sewage
lagoons, leaking sewer lines, and overflowing cesspools can enter ground water sources through the
subsurface. Other sources of fecal contamination include improperly treated wastewater used to recharge
ground water or to irrigate crop land and improper land application of raw septage or treated sewage.
Solid wastes contaminated with human or animal bacteria and viruses may contaminate ground water
through individual waste disposal practices, open dumping practices, and landfills (Washington State
Department of Health, 1995). Improper land application of waste waters associated with food processing
or animal slaughter may also contribute to the contamination of ground water sources of drinking water.
Microbial pathogens found in animal wastes may enter ground water from unlined or leaky manure
lagoons, spread manure, and concentrated animal feeding operations (USEPA, 1993; Washington State
Department of Health, 1995). Given the right subsurface conditions, contaminated water from any of
these sources may reach the aquifer and, possibly, the intake zone of a drinking water well.
2.2.2.2 Ground Water Contamination through the Wellhead
Conditions at or near wells may contribute to the occurrence of ground water contamination.
Contamination may occur at the wellhead in several ways. The main causes are poor well location and/or
construction, improperly abandoned wells, and the presence of test holes or exploratory wells.
1 For a more detailed discussion of the factors affecting bacterial and viral fate and transport, see the
Occurrence and Monitoring Document for the Final Ground Water Rule (USEPA, 2006b).
Economic Analysis for the 2-6 October 2006
Final Ground Water Rule
-------
Well Location and/or Construction
Fecal contamination can enter inappropriately located or improperly constructed wells in several
ways. A well located in a low-lying area or within a well pit is susceptible to flooding. A well may be
particularly vulnerable to surface water contamination if it is not adequately cased and grouted. An
improperly constructed water-supply well may allow surface runoff or surface waters to enter through a
non-existent or broken well seal. Ground water contamination may also result from water infiltrating
through a contaminated gravelpack or the fill surrounding the intake point.
Many old wells were built before the institution of strict construction guidelines, sometimes in a
manner that could allow contamination. In some cases, newer wells, built after the institution of stricter
well construction guidelines, do not adhere to those guidelines. Such wells, if constructed in or near
potential sources of contamination, may be vulnerable to the contamination. Even wells that are built and
sited correctly may be exposed to contaminants through an improperly constructed well penetrating the
same aquifer as the properly constructed well.
Abandoned Wells
Historically, well abandonment and plugging have generally not been properly planned, designed,
and executed (USEPA, 1990; Canter et al., 1987). In many cases, the well casing was pulled out if it was
not too worn or corroded, thereby linking aquifers at different depths through the well shaft. Such wells
could then serve as conduits for contaminated ground water to spread more rapidly to other zones within
an aquifer or allow contaminants to enter adjacent aquifers at lower hydraulic pressures (USEPA, 1990).
Other wells may not have been adequately plugged, providing a pathway for contamination from the
surface. Occasionally, abandoned wells have also been used as disposal sites for a variety of wastes,
resulting in the direct contamination of an aquifer.
Test Holes, Exploratory Wells, and Monitoring Wells
Many test holes and exploratory wells have been dug or drilled into the subsurface to search for
oil, gas, coal, minerals, and water. Other holes have been drilled for testing, including soil boreholes and
seismic shot holes. Monitoring wells are often drilled to sample ground water quality. When these holes
are not backfilled, or when the wells are not properly constructed or abandoned, they provide potential
conduits for contamination to enter ground water sources.
2.2.2.3 Contamination of Drinking Water in Distribution Systems
Even if the ground water source for a water system remains clean, contamination may occur
within the distribution system. Numerous contamination incidents in distribution systems have been
reported from systems using ground water sources. For example, if proper precautions are not taken,
water in the distribution system can become contaminated following routine maintenance or emergency
repairs (e.g., flushing or chlorination). Inadequately disinfected distribution systems, including storage
towers, can develop microbial mats or biofilms. Initially, biofilms may function as a filter, adsorbing
pathogens (Seunghyun and Corapcioglu, 1997), but the pathogens may ultimately be shed (sloughed)
from the system, potentially contaminating the drinking water at the tap.
Economic Analysis for the 2-7 October 2006
Final Ground Water Rule
-------
2.3 Statutory Authority for Promulgating the Rule
EPA has the primary responsibility for regulating the quality of drinking water. The SDWA
establishes this responsibility and defines the mechanisms at the Agency's disposal to protect public
health. EPA sets standards by identifying which contaminants should be regulated and by establishing
the maximum levels of the contaminants allowed in drinking water.
The 1986 Amendments to the Safe Drinking Water Act directed EPA to promulgate regulations
requiring disinfection at all public water systems using either surface water or ground water. The Surface
Water Treatment Rule (40 CFR Part 141 subpart H) implemented that requirement for surface water
systems (and systems using ground water under the direct influence of surface water), but when Congress
amended the Safe Drinking Water Act again in 1996, EPA had not promulgated regulations requiring
disinfection for systems that use ground water. In the legislative history of the 1996 Amendments to the
Safe Drinking Water Act, Congress identified several reasons for the delay, including the recognition that
not all GWSs are at risk of contamination and the high cost of across-the-board disinfection. In light of
this recognition, Sectionl412(b)(8) of the Safe Drinking Water Act, as amended on August 6, 1996,
requires EPA to promulgate national primary drinking water regulations (NPDWRS) requiring
disinfection as a treatment technique for all ground water systems only as necessary. In addition, Section
1412(b)(8) requires EPA to promulgate criteria as part of the regulations for determining whether
disinfection should be required as a treatment technique for any public water system served by ground
water.
Section 1413(a)(l) allows EPA to grant a State primary enforcement responsibility (primacy) for
NPDWRs when EPA has determined that the State has adopted regulations that are no less stringent than
EPA's. To obtain primacy for the final GWR, States must adopt comparable regulations within two years
of EPA's promulgation of the final rule, unless atwo year extension is granted. State primacy requires,
among other things, adequate enforcement (including monitoring and inspections) and reporting. EPA
must approve or deny State primacy applications within 90 days of submission to EPA (section
1413(b)(2)). In some cases, a State submitting revisions to adopt an NPDWR has primacy enforcement
authority for the new regulation while EPA action on the revision is pending (section 1413(c)). Section
1445 authorizes the Administrator to establish monitoring, record keeping and reporting regulations to
assist the Administrator in determining compliance with the Safe Drinking Water Act and in advising the
public of the risks of unregulated contaminants. Section 1450 of the Safe Drinking Water Act authorizes
the Administrator to prescribe such regulations as are necessary or appropriate to carry out his functions
under the Act.
2.4 Regulatory History
The following sections summarize the development of NPDWRs and programs most relevant to
the GWR over the past 20 years. These include rules addressing microbial contaminants, rules addressing
disinfectants and disinfection byproducts (DBFs), and other rules that were considered (e.g., to determine
conflict or overlap of requirements) during the evaluation of GWR regulatory requirements. Drinking
water regulations that apply only to surface water systems, including the 1989 Surface Water treatment
Rule (SWTR), the 1998 Interim Enhanced Surface Water Treatment Rule (IESWTR), the 2001 Filter
Backwash Recycling Rule (FBRR), the 2002 Long Term 1 Enhanced Surface Water Treatment Rule
(LT1ESWTR), and the 2006 Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR), are
not summarized.
Economic Analysis for the 2-8 October 2006
Final Ground Water Rule
-------
2.4.1 1979 Total Trihalomethane Rule
Under the Total Trihalomethane (TTHM) Rule (44 FR 68624, November 1979), EPA set an MCL
for TTHM, the sum of the concentrations of chloroform, bromoform, bromodichloromethane, and
dibromochloromethane, of 0.10 milligrams per liter (mg/L) as a running annual average (RAA) of
quarterly measurements. This standard applies to community water systems (CWSs) using surface or
ground water that serve at least 10,000 people and that add a disinfectant to the drinking water during any
part of the treatment process. This 1979 rule has been superceded by the 1998 Stage 1 DBPR (section
2.4.4). Compliance with the Stage 1 DBPR began in January 2002.
2.4.2 1989 Total Coliform Rule
The Total Coliform Rule (TCR) (54 FR 27544, June 1989) applies to all PWSs. Because
monitoring PWSs for every possible pathogenic organism is not feasible, coliform organisms are used as
indicators of possible system contamination. Coliforms are easily detected in water and are used to
indicate a water system's vulnerability to pathogens. In the TCR, EPA set a Maximum Contaminant
Level Goal (MCLG) of zero for total coliforms. EPA also set a monthly MCL for total coliforms and
required testing of total coliform-positive cultures for the presence of E. coll or fecal coliforms. E. coll
and fecal coliforms indicate more immediate health risks from sewage or fecal contamination and are
used as a trigger of an acute MCL violation. Coliform monitoring frequency is determined by population
served, the type of system (community or noncommunity systems), and the type of source water (surface
water, ground water under the direct influence of surface water (GWUDI), or ground water). In addition,
the TCR required sanitary surveys every five years (ten years for noncommunity systems using
disinfected ground water) for systems that collect fewer than five total coliform samples per month (those
serving 4,100 people or fewer).
2.4.3 1996 Information Collection Rule
The Information Collection Rule (ICR) (61 FR 24354, May 1996) applied to PWSs serving at
least 100,000 people. A more limited set of ICR requirements cover GWSs serving 50,000 to 99,999
people. The ICR authorized EPA to collect occurrence and treatment information from water treatment
plants to help evaluate the possible need for changes to the current microbial requirements and existing
microbial treatment practices and to help evaluate the need for future regulation of disinfectants and
DBFs. The ICR provided EPA with information on the national occurrence of (1) chemical byproducts
that form when disinfectants used for microbial control react with naturally occurring compounds and
ions present in source water; and (2) disease-causing microorganisms including Cryptosporidium,
Giardia, viruses, and coliform bacteria. The ICR also mandated the collection of treatment train data on
how water systems currently control for contaminants. The ICR monthly sampling data provided 18
months of information on the quality of the influent and treated water, including pH, alkalinity, turbidity,
temperature, calcium, total hardness, total organic carbon, UV254 absorbency, bromide, ammonia, and
disinfectant residual. These data provide some indication of the "treatability" of the water, the occurrence
of contaminants, and the potential for DBP formation. The data collected under the ICR are continuing to
be analyzed to help develop current and future NPDWRs.
Economic Analysis for the 2-9 October 2006
Final Ground Water Rule
-------
2.4.4 1998 Stage 1 Disinfectants and Disinfection Byproducts Rule
The Stage 1 DBPR (63 FR 69390, December 1998) applies to all CWSs and nontransient
noncommunity water systems (NTNCWSs) that add a chemical disinfectant to their water. Certain
requirements designed to provide protection against acute health effects from chlorine dioxide also apply
to transient noncommunity water systems (TNCWSs). Surface water and GWUDI systems serving at
least 10,000 people were required to begin compliance with the rule by January 2002. Surface water and
GWUDI systems serving fewer than 10,000 people and all GWSs must comply beginning January 2004.
The Stage 1 DBPR sets Maximum Disinfectant Residual Level Goals (MRDLGs) for chlorine (4
mg/L as C12), chloramines (4 mg/L as C12), and chlorine dioxide (0.8 mg/L as C1O2) and MCLGs for
bromodichloromethane (0 mg/L), bromoform (0 mg/L), dibromochloromethane (0.06 mg/L),
dichloroacetic acid (0 mg/L), trichloroacetic acid (0.3 mg/L), bromate (0 mg/L), and chlorite (0.8 mg/L).
The rule sets Maximum Residual Disifectant Leevels (MRDLs) for chlorine (4.0 mg/L as C12),
chloramines (4.0 mg/L as C12), and chlorine dioxide (0.8 mg/L as C1O2) and MCLs for TTHM (0.080
mg/L), haloacetic acids [total of five] (HAAS) (0.060 mg/L), bromate (0.010 mg/L), and chlorite (1.0
mg/L). The MRDLs and MCLs, except those for chlorite and chlorine dioxide, are calculated as RAAs of
quarterly measurements. For surface water and GWUDI systems using conventional filtration treatment,
a treatment technique—enhanced coagulation/softening—is specified for the removal of DBF precursors.
2.4.5 2001 Arsenic Rule
The Arsenic Rule (66 FR 6976, January 2001) increases the level of public health protection
against exposure to arsenic in drinking water. The rule revises the MCL for arsenic in drinking water
from 0.05 mg/L to 0.010 mg/L and sets an MCLG of 0 mg/L for all CWSs and NTNCWSs. Clarification
on how compliance is demonstrated for many inorganic and organic contaminants in drinking water is
also given. All existing CWSs and NTNCWSs were required to comply with the Arsenic Rule by
January 23, 2006.
2.4.6 2006 Stage 2 Disinfectants and Disinfection Byproducts Rule
The Stage 2 DBPR tightens certain DBF compliance standards set by the Stage 1 DBPR. The
Stage 2 DBPR was promulgated concurrently with the LT2ESWTR and is designed to reduce DBP
occurrence peaks in the distribution system by changing compliance monitoring provisions. The
requirements in the Stage 2 DBPR apply to all CWSs and NTNCWSs that add a primary or residual
disinfectant other than ultraviolet light (UV) or that deliver water that has been treated with a disinfectant
other than UV.
For the Stage 2 DBPR, the MCLs will remain at the Stage 1 DBPR levels (0.080 mg/L for TTHM
and 0.060 mg/L for HAAS), but compliance will be measured based on locational running annual
averages (LRAAs) instead of the RAAs used in the Stage 1 DBPR. Most systems will also be required to
conduct Initial Distribution System Evaluations (IDSEs) to identify monitoring locations that represent
locations with the highest concentrations of TTHM and HAAS. In addition, the Stage 2 DBPR addresses
systems that are in full compliance with the Stage 2 DBPR LRAA MCLs but have individual DBP
measurements that exceed the MCLs. The rule provides a formula for determining averages for TTHM
and HAAS. If these averages exceed certain levels, the system must conduct an operational evaluation
and submit a written report or the operational evaluation to the State.
Economic Analysis for the 2-10 October 2006
Final Ground Water Rule
-------
2.4.7 Underground Injection Control Program
The EPA's Underground Injection Control (UIC) program was established to protect sources of
drinking water from underground injection of fluids through wells. Owners and operators of injection
wells are prohibited from operating their wells in a manner that causes the movement of fluid into
underground sources of drinking water if it may cause a violation of any primary drinking water
regulation or otherwise adversely affect human health. To prevent such fluid movement, EPA or the
appropriate State regulatory agency may require certain construction criteria, corrective action, operation,
monitoring, reporting, or plugging and abandonment. These regulations are designed to recognize
varying geologic, hydrological, or historical conditions among different States or areas within a State.
The regulations included in 40 CFR 144.6 define five classes of injection wells. These wells may
inject fluids that are associated with hazardous waste or radioactive waste sites, natural gas or oil
production, extraction of minerals, or other purposes. Class V wells are the most prevalent of the
injection well types and are most often associated with ground water contamination relevant to the GWR.
They include:
Untreated sewage waste disposal wells
Large-capacity cesspools
Large-capacity septic systems (undifferentiated disposal method)
Large-capacity septic systems (well disposal method)
Large-capacity septic systems (drainfield disposal method)
Domestic wastewater treatment plant effluent disposal wells
EPA regulates only multiple-dwelling, community, or regional septic systems, as opposed to individual or
single-family residential septic systems, as Class V wells (40 CFR 144.1(g)(l)(2)).
In November 1999, EPA finalized new UIC regulations that added requirements for two
categories of Class V wells (USEPA, 1999b). The regulation bans new large-capacity cesspools
nationwide and phases out existing large-capacity cesspools by April 2005. It also bans new motor
vehicle waste disposal wells nationwide. Operation of existing motor vehicle waste disposal wells in
ground water source water areas or other sensitive ground water areas is banned but with a provision that
owners and operators of such wells may seek a waiver from closing if they obtain a permit. The permit
conditions include requirements for meeting MCLs and other health-based standards at the point of
injection, injectate and sludge monitoring, and implementing best management practices. EPA expects to
achieve substantial protection of underground sources of drinking water by focusing the requirements on
these particular wells.
Economic Analysis for the 2-11 October 2006
Final Ground Water Rule
-------
2.5 Economic Rationale
This section addresses the economic rationale for choosing a regulatory approach rather than non-
regulatory alternatives. An economic rationale for the rule is required by Executive Order 12866,
Regulatory Planning and Review (58 FR51735), which states:
"[EJach agency shall identify the problem that it intends to address (including, where
applicable, the failures of the private markets or public institutions that warrant new
agency action) as well as assess the significance of that problem." (Section 1, b(l))
In addition, Office of Management and Budget (OMB) guidance, dated January 11, 1996, states
that "in order to establish the need for the proposed action, the analysis should discuss whether the
problem constitutes a significant market failure" (USEPA, 1996a).
In a perfectly competitive market, prices and quantities are determined solely by the aggregated
decisions of buyers and sellers. Such a market occurs when many producers of a product are selling to
many buyers and where both producers and consumers have perfect information on the characteristics and
prices of each firm's products. Barriers to entry in the industry cannot exist, and individual buyers and
sellers must be "price takers" (i.e., their individual decisions cannot affect the price). Several properties
of the public water supply do not satisfy the conditions for a perfectly competitive market and thus lead to
market failures that require regulation.
Many water systems are natural monopolies. A natural monopoly exists when it is impossible for
more than one firm in each area to recover the costs of production and survive. There are high fixed costs
associated with reservoirs and wells, transmission and distribution systems, treatment plants, and other
facilities. For other potential suppliers to enter the market, they would have to provide the same extensive
infrastructure to realize similar economies of scale and be competitive. A splitting of the market with
increased fixed costs (e.g., two supplier networks in a single market) usually makes this situation
unprofitable. The result is a market suitable for a single supplier and hostile to alternative suppliers. In
such natural monopolies, suppliers have fewer incentives for providing quality services or maintaining
competitive prices. In these situations, governments often intervene to help protect the public interest.
For example, because PWSs are legal, as well as natural, monopolies, they are often subject to
price controls, if not outright public ownership. While customers may demand improvements in water
quality, the regulatory structure may not facilitate the transmission of that demand to the water supplier or
allow the supplier to raise its price to recover the cost of the improvements. If consumers do not believe
that their drinking water is safe enough, they cannot simply switch to another water utility. Other options
for obtaining safe drinking water (e.g., buying bottled water or installing point of use filtration) most
often represent a higher water cost to consumers than the purchase from PWSs. Therefore, the water
supplier may have little incentive to improve water quality.
The public may also not understand the health and safety issues associated with poor drinking
water quality. Understanding the health risks posed by trace quantities of drinking water contaminants
involves analysis and synthesis of complex toxicological and health sciences data. Therefore, the public
may not be aware of the risks it faces. EPA has implemented a Consumer Confidence Report (CCR) Rule
(63 FR 44512, August 1998) that makes water quality information more easily available to consumers.
This rule requires CWSs to publish an annual report on local drinking water quality. Consumers,
however, still have to analyze this information for its health risk implications. Furthermore, even if
informed consumers are able to engage water systems in a dialogue about health issues, the transaction
Economic Analysis for the 2-12 October 2006
Final Ground Water Rule
-------
costs of such interaction (measured in personal time and monetary outlays) present another significant
impediment to consumer expression of risk reduction preferences.
SDWA regulations are intended to provide a level of protection from exposure to drinking water
contaminants. The regulations set minimum performance requirements to protect consumers from
exposure to contaminants. SDWA regulations are not intended to restructure market mechanisms or to
establish competition in supply; rather, they establish the level of service to be provided that best reflects
public preference for safety. The federal regulations reduce the high information and transaction costs by
acting on behalf of consumers in balancing risk reduction and the social costs of achieving this risk
reduction.
Economic Analysis for the 2-13 October 2006
Final Ground Water Rule
-------
3. Consideration of Regulatory Alternatives
3.1 Introduction
The 1996 Amendments to the Safe Drinking Water Act (SDWA) require the United States
Environmental Protection Agency (EPA or Agency) to develop a national primary drinking water
regulation that requires disinfection as a treatment technique for all ground water PWSs only "as
necessary." To address this mandate and the public health concerns presented in Chapter 2 of this EA,
EPA developed the Ground Water Rule (GWR). The Agency convened workgroups, held stakeholder
meetings, published a proposed rule, and evaluated public comments to develop a final regulation. This
chapter describes the process used to evaluate regulatory alternatives considered during the development
of the GWR and evaluated as part of this Economic Analysis (EA).
3.2 Process for Development of Regulatory Alternatives
In 1992, EPA circulated a draft proposal for review and comment, which began the process of
developing regulatory alternatives for addressing microbial contamination in ground water systems
(GWSs). In 1993, EPA published a preliminary draft of the Ground Water Disinfection Rule (later
renamed the Ground Water Rule). After review of the public comments, EPA recognized that additional
information needed to be gathered and, in 1995, convened a GWR regulatory workgroup. EPA used the
workgroup to obtain comments and additional information regarding the GWR. In 1996, EPA published
a report, Ground Water Disinfection and Protective Practices in the United States, on ground water-
related statutes, regulations, guidance, and disinfection practices gathered from 50 State drinking water
programs (USEPA, 1996b). In 1997, EPA initiated another workgroup, including members from EPA,
other Federal agencies, and State agencies, to cooperate in the development of a proposed GWR.
In December 1997, EPA initiated stakeholder meetings. The Agency published public meeting
announcements in the Federal Register. EPA involved citizens, environmental groups, small businesses,
and water suppliers early in the rule development process and held public meetings in different regions
(Washington, DC; Portland, OR; Madison, WI; and Dallas, TX) to facilitate rule development and allow
stakeholders the opportunity to comment on the regulation.
In addition to the public meetings with stakeholders, EPA, as part of the consultation process
required by the Small Business Regulatory Enforcement Fairness Act (SBREFA), met with
representatives of small systems (i.e., those serving fewer than 10,000 people) in March and April 1998.
EPA presented possible regulatory requirements and requested comments from the representatives during
these meetings.
In January 1999, EPA published a preliminary draft preamble for the GWR and solicited
comment. The preliminary draft preamble described regulatory alternatives and requested public
comment on a number of potential modifications. EPA received 80 comments on the preliminary draft
preamble.
Economic Analysis for the 3-1 October 2006
Final Ground Water Rule
-------
3.3 Regulatory Alternatives Considered
EPA proposed four regulatory alternatives. The multi-barrier approach was the preferred
alternative in the proposed GWR. The primary elements of the multi-barrier approach were sanitary
surveys, triggered monitoring, HSAs, routine monitoring, corrective action, and compliance monitoring.
After the proposal, EPA considered comments and chose a different final rule alternative, Alternative 2,
termed the risk-targeted approach. The primary elements of the risk-targeted approach include sanitary
surveys, triggered monitoring, optional assessment monitoring, corrective action, and compliance
monitoring.
The following discussion provides details of the four main regulatory alternatives considered. A
detailed comparison of the quantified benefits and costs of each of the four regulatory alternatives for the
GWR is found in Chapter 8 of this EA.
Alternative 1: Sanitary survey and corrective action
Sanitary surveys are on-site assessments of the source water, treatment facilities, distribution
systems, finished water storage tanks, monitoring records, and the management and operation of a public
water system (PWS). This alternative requires sanitary surveys to be conducted by the State at least once
every 3 years for community water systems (CWSs) (with a provision that States may reduce frequency to
every five years for CWSs with outstanding performance and no TCR violations or that provide 4-log
treatment) and every 5 years for noncommunity water systems (NCWSs). Sanitary surveys must address
various elements set out in the EPA/State Joint Guidance on sanitary surveys. Operators would be
required to correct any significant deficiencies within 120 days of receiving the State sanitary survey
report or in accordance with a State-approved plan and schedule for correcting these deficiencies. All
systems that perform treatment (including those already achieving 4-log treatment of viruses before or at
the first customer) must perform compliance monitoring to ensure that the treatment is effective.
Alternative 2: Risk-Targeted Approach (Final Rule Alternative)
In addition to all the sanitary survey components of the first alternative, the Risk-Targeted
Approach includes a triggered source water microbial monitoring requirement for systems. Systems that
do not already achieve at least 4-log (99.99 percent) treatment of viruses (using inactivation, removal, or
State-approved combination of these technologies) before or at the first customer and that have a total
coliform (TC) positive result for any routine sample taken under the Total Coliform Rule (TCR) must
collect a ground water source sample. The ground water system must test the source water sample for a
fecal indicator (E. coll, enterococci, or coliphage) determined by the State or primacy agency. Unless the
State requires immediate corrective action for the initial fecal indicator-positive source water sample, the
system must collect and analyze five additional ground water source samples for the presence of a fecal
indicator. If any one of the five source water samples tests fecal indicator-positive, the system must take
correction action. Systems that are required to take corrective action must consult with the State within
30 days and complete the action within 120 days (or be in compliance with the State or primacy agency
approved plan and schedule) of receiving notice of the fecal indicator-positive sample in one of the five
additional samples or when the State determines a corrective action is necessary, whichever occurs first.
The final GWR provides States with the option to require systems to conduct assessment monitoring any
time and require systems to take corrective actions. See section 3.4 for more details on this alternative.
Economic Analysis for the 3-2 October 2006
Final Ground Water Rule
-------
Alternative 3: Multi-barrier Approach
The multi-barrier approach described in this EA includes modifications to two proposed
provisions: hydrogeologic sensitivity assessment and routine monitoring provisions. The changes made
to the multi-barrier approach stemmed from public comments received on the proposed GWR regulatory
alternative.
The multi-barrier approach builds on the sanitary survey and triggered monitoring requirements
of the first two alternatives. This alternative includes an optional HSA provision for States rather than a
required HSA provision as proposed. An HSA is a tool used to identify those systems in aquifers where
water can move quickly through the subsurface, thereby increasing the possibility of fecal contamination.
Under this alternative, each State has the option to complete the hydrogeologic sensitivity assessment
within six and eight years from the GWR publication date for CWSs and NCWSs, respectively. HSAs
would be conducted for all existing and new systems that do not maintain at least a 4-log treatment of
viruses (using inactivation, removal, or State-approved combination of these technologies) before or at
the first customer. If the State did not conduct an HSA to determined aquifer sensitivity, the system
would be required to conduct twelve months of assessment monitoring.
If an aquifer that a system uses as a source is identified as sensitive through an HSA or if the
State chooses not to conduct an HSA, the system would have to conduct assessment monitoring (a
derivation of routine monitoring in the proposed rule). The assessment monitoring provision involves the
system collecting a ground water source sample and testing it for the State-specified fecal indicator each
month that the system serves water to the public until a total of 12 ground water source samples is
collected. The proposed rule required a system to collect a source water sample each month that the
system serves water to the public, or otherwise specified by the State. Assessment monitoring is
performed in addition to any triggered source water monitoring samples that may be required.
Under the multi-barrier approach, systems found to have source water contamination (either
through assessment or triggered monitoring) would be required to consult with the State within 30 days
and take corrective action within 120 days (or longer if the State or primacy agency approves a plan and
schedule) from the date the system receives notice of the fecal indicator-positive sample.
Alternative 4: Across-the-Board Disinfection
This alternative requires all public ground water systems to install or operate disinfection
treatment processes capable of achieving a 4-log treatment of viruses (using inactivation, removal, or
State-approved combination of these technologies) before or at the first customer on a continuing basis.
Systems treating to less than 4 logs would be required to upgrade their treatment. Unlike the other
alternatives, the across-the-board disinfection alternative does not consider the quality of a system's
source water or potential for contamination. Similar to Alternatives 2 and 3, all systems (including those
already achieving 4-log treatment of viruses before or at the first customer) would have to conduct
compliance monitoring to ensure the treatment is effective. Also, States would be required to perform
sanitary surveys of ground water systems to ensure the treatment practices are being properly operated
and to evaluate the ground water system for potential source contamination.
3.4 Final Rule Requirements
Following publication of the proposed GWR, EPA accepted public comments for 90 days. EPA
received approximately 3,300 comments from over 250 individuals and organizations representing a wide
Economic Analysis for the 3-3 October 2006
Final Ground Water Rule
-------
range of stakeholders, including public water systems, States, Tribes, other organizations, and private
citizens. Each comment was read and considered as part of the process for selecting and, where
appropriate, modifying the final GWR regulatory alternative. A record of every comment received on the
proposal, as well as EPA's response to each, can be found in the Public Comment and Response
Document for the Final Ground Water Rule (USEPA, 2006a). Copies of individual comments are also
available as part of the public record and can be accessed through EPA's Water Docket.
Based on public comments, EPA reevaluated the regulatory alternatives, the assumptions, and
data underlying the GWR proposal. As stated in section 3.3, EPA selected Alternative 2 as the final
GWR alternative. Alternative 2, the risk-targeted approach, includes mandatory rule components and an
optional assessment monitoring provision. Exhibit 3.1 provides a flowchart of the mandatory rule
components.
Economic Analysis for the 3-4 October 2006
Final Ground Water Rule
-------
Exhibit 3.1 Flowchart of Mandatory PWS Ground Water Rule Requirements
Sanitary surveys for all
ground water PWSs
All ground water PWSs
Conduct TCR routine
monitoring
Does system
provide
treatment >4
logs?
Did survey
find
deficiencies?
Must correct deficiencies (4->
Continue GWR
compliance: sanitary
surveys, compliance
monitoring, TCR
compliance
Compliance monitorm
Is TCR routine
sample TC-positive?
Continue GWR
compliance: sanitary
surveys, triggered
monitoring, TCR
compliance
Conduct triggered
Take non-treatment
corrective actions
Fecal Coliform
Indicator Positive?
Did system
install
(1)
(2)
(3)
(4)
Per State's direction, take
corrective action or take
5 additional samples (4'
Systems are not required
to take corrective action
if all 5 repeat samples are
fecal negative.
Includes mixed systems with ground water entry points
Treatment using inactivation, removal, or State-approved
combination of these technologies.
For those systems that do not receive a triggered monitoring
waiver from the State
The State may determine that the source of contamination has
been eliminated..
treatment as
corrective
action?
Economic Analysis for the
Final Ground Water Rule
3-5
October 2006
-------
3.4.1 Mandatory Rule Components
The following subsections describe the GWR regulatory requirements with which ground water
systems and States must comply. Systems meeting certain criteria (i.e., 4-log disinfection or inactivation
of viruses) may not be required to implement all provisions.
Sanitary Surveys
The final GWR requires States to perform sanitary surveys for all GWSs. Ground water systems
must provide the State with any pertinent, existing information that will enable the State to perform the
sanitary survey. The final GWR goes beyond the existing definition of sanitary survey at 40 CFR 141.2,
explicitly references the use and relevance of source water assessments required under the 1996 SDWA
Amendments, and specifies in more detail the scope of a sanitary survey. Specifically, the final GWR
requires that States evaluate the eight components outlined in the EPA/State Joint Guidance as part of the
sanitary survey to the extent that they apply to an individual system:
(1) source;
(2) treatment;
(3) distribution system;
(4) finished water storage;
(5) pumps, pump facilities, and controls;
(6) monitoring, reporting, and data verification;
(7) system management and operation; and
(8) operator compliance with State requirements.
The final GWR requires States to conduct sanitary surveys of ground water CWSs every three
years (every five years for CWSs that meet performance criteria as described in the following paragraph)
and of ground water NCWSs every five years. States are required to complete the initial sanitary survey
cycle by December 31, 2012 for CWSs, except those that meet performance criteria (e.g., 4-log treatment
or outstanding performance and no TCR violations), and December 31, 2014 for all NCWSs and CWSs
that meet performance criteria. States may conduct more frequent sanitary survey cycles for any GWS as
appropriate.
The final GWR allows individual components of a sanitary survey to be conducted according to a
phased review process (e.g., as part of ongoing State assessment programs). While all applicable
components need not be evaluated at the same time, they must be evaluated within the required three- or
five-year frequency interval. Also, the final GWR allows the three-year CWS schedule to be extended to
a five-year frequency if the system meets performance criteria.
Finally, the final GWR requires that GWSs correct any significant deficiencies identified in
sanitary surveys. Significant deficiencies, as determined by the State, include, but are not limited to,
defects in design, operation, or maintenance, or a failure or malfunction of the sources, treatment, storage,
or distribution system that the State determines to be causing, or have the potential for causing, the
introduction of contamination into the water delivered to consumers.
The State must provide the GWS with written notification, which describes any significant
deficiencies found, no later than 30 days after the State identifies the significant deficiency. The notice
may be sent to the PWS, or it may be provided on-site either at the time the sanitary survey is conducted
or the significant deficiency is identified. The State may specify appropriate follow-up corrective action
steps in the notice or may notify the GWS of appropriate corrective actions during the consultation
Economic Analysis for the 3-6 October 2006
Final Ground Water Rule
-------
period. After receiving the written notification, the GWS has 30 days to consult with the State regarding
corrective actions. The GWS must correct the significant deficiency within 120 days or be on a
State-approved corrective action plan and schedule. States must confirm that the deficiencies have been
addressed within 30 days after the scheduled correction date. (See also the discussion of corrective
actions below.)
Triggered Source Water Monitoring
A GWS must conduct triggered source water monitoring within 24 hours of receiving notification
that a sample collected in accordance with 40 CFR 141.21(a) (TCR) is total coliform-positive. A GWS
must collect at least one ground water source sample from each ground water source (e.g., a well or
spring) in use at the time the total coliform-positive sample was collected. Triggered source water
monitoring is required unless:
(1) the GWS provides at least 4-log treatment of viruses (using inactivation, removal, or a
State-approved combination of 4-log virus inactivation and removal) before or at the first
customer for each ground water source;
(2) the GWS is notified that a positive sample collected in accordance with 40 CFR
141.21(a) (TCR) has been invalidated under 40 CFR 141.21(c); or
(3) the cause of the total coliform-positive collected under 40 CFR 141.21(a) directly relates
to the distribution system according to State criteria or a State determination.
The State may extend the 24-hour limit on a case-by-case basis if the State determines that the system
cannot collect the ground water source water sample within 24 hours due to circumstances beyond its
control. In the case of an extension, the State must specify how much time the system has to collect the
sample.
Systems are not required to conduct triggered source water monitoring if, according to State
criteria or a State determination, the cause of the total coliform-positive sample collected under 40 CFR
141.21(a) directly relates to the distribution system. If the decision is made according to State criteria, the
GWS must document the decision in writing; if the decision is made by the State, the State must
document the decision in writing. In the primacy application, the State must include criteria that will be
used to determine that the cause of a total coliform-positive sample collected under 40 CFR 141.21 (a) is
directly related to the distribution system.
If the State approves the use of E. coli as a fecal indicator for triggered source water monitoring,
GWSs serving 1,000 people or fewer may use a TCR repeat sample collected from a ground water source
to simultaneously meet the requirements of 40 CFR 141.21(b) and satisfy the GWR's triggered source
water monitoring requirements for that ground water source only.
If approved by the State, GWSs with more than one ground water source may conduct triggered
source water monitoring at a representative ground water source or sources. The State may require
systems with more than one ground water source to submit for approval a triggered source water
monitoring plan that the system will use for representative sampling. A triggered source water
monitoring plan must identify ground water sources that are representative of each monitoring site in the
system's TCR sample siting plan.
Economic Analysis for the 3-7 October 2006
Final Ground Water Rule
-------
If any initial triggered source water sample is fecal indicator-positive, the system must collect five
additional source water samples within 24 hours at that site, unless the State requires immediate
corrective action to address contamination at that site. The samples must be tested for the same fecal
indicator for which the initial source water sample tested positive.
Ground water systems that purchase or sell finished drinking water (referred to as consecutive or
wholesale systems, respectively) must comply with triggered source water monitoring provisions for their
own sources.
Consecutive and wholesale systems must also comply with other triggered source water
monitoring requirements. A consecutive GWS that has a total coliform-positive sample collected under
40 CFR 141.21(a) (TCR) must notify the wholesale system(s) within 24 hours of being notified of the
total coliform-positive sample. If a wholesale GWS receives notice from a consecutive system it serves
that a sample collected under 40 CFR 141.21(a) (TCR) is total coliform-positive, the wholesale GWS
must conduct triggered source water monitoring. If the sample is fecal indicator-positive, in addition to
notifying its own customers, the wholesale GWS must notify all consecutive systems served by that
ground water source. The consecutive system is responsible for providing any required public notice to
the persons it serves.
Corrective Action
When a GWS has a significant deficiency, it must consult with the State regarding appropriate
corrective action within 30 days of receiving a written notice of the significant deficiency. When a GWS
receives a written notice from a laboratory indicating a fecal indicator positive result in one of the five
additional triggered source water monitoring samples, the GWS must consult with the State regarding
appropriate corrective action. When a GWS receives a written notice from a laboratory indicating a fecal
indicator positive result and the State has determined that corrective action is necessary, the GWS must
consult with the State regarding appropriate corrective action. Consultation must take place within 30
days. In any event, the State may specify corrective action without consultation. In the consultation
process, the State may approve and/or modify corrective actions and completion schedules proposed by
the system, or the State may specify alternatives. The State may also specify interim corrective action
measures.
The final GWR rule requires that within 120 days (or earlier if directed by the State) of receiving
the notification from the State or laboratory described in the preceding paragraph, the GWS must either
(i) complete appropriate corrective actions in accordance with applicable State plan review
processes or other State guidance or direction, or
(ii) be in compliance with a State-approved corrective action plan and schedule.
If a GWS is unable to complete corrective action within 120 days or on the schedule specified by the
State, then the GWS is in violation of the treatment technique requirement.
Systems must notify the State within 30 days of completing any State approved or specified
corrective action. As a condition of primacy, States must verify that the corrective action has been
completed within the next 30 days. States may verify that the corrective action has been completed and
has successfully addressed the significant deficiency and/or fecal contamination in the ground water
source either by a site visit or by written documentation from the system, which could consist of the
system's notification to the State.
Economic Analysis for the 3-8 October 2006
Final Ground Water Rule
-------
Corrective Action Alternatives
When a system has a significant deficiency or a fecal indicator-positive ground water source
sample (either by the initial triggered sample, or positive additional sample, as determined by the State),
the GWS must implement one or more of the following corrective action options:
(1) correct all significant deficiencies (e.g., repairs to well pads and sanitary seals, repairs to
piping tanks and treatment equipment, control of cross-connections);
(2) provide an alternate source of water (e.g., new well, connection to another PWS);
(3) eliminate the source of contamination (e.g., remove point sources, relocate pipelines and
waste disposal, redirect drainage or run-off, provide or fix existing fencing or housing of
the wellhead); or
(4) provide treatment that reliably achieves at least 4-log treatment of viruses (using
inactivation, removal, or a State-approved combination of 4-log virus inactivation and
removal) before or at the first customer for each ground water source.
Compliance Monitoring for Systems Providing At Least 4-log Treatment of Viruses
The final GWR also establishes compliance monitoring requirements for GWSs that provide at
least 4-log treatment of viruses as a corrective action. The final GWR also establishes compliance
monitoring requirements for those systems that have notified the State that they provide at least 4-log
treatment of viruses for their ground water sources before the first customer and are therefore not required
to meet the triggered source water monitoring requirement of this rule.
Treatment technologies capable of providing at least a 4-log treatment of viruses include the
following:
•• Chemical Disinfection. Inactivation, with a sufficient disinfection concentration and
contact time, through disinfection with chlorine, chlorine dioxide, ozone, or through
anodic oxidation. Disinfectant concentration and contact time (CT) can be based on
existing CT tables (USEPA, 1991) or State-approved alternatives.
•• Membrane Filtration. Removal with membrane technologies with an absolute molecular
weight cut-off (MWCO), or an alternate parameter that describes the exclusion
characteristics of the membrane, that can reliably achieve at least a 4-log removal of
viruses.
•• Alternative Treatment. Inactivation, removal or combination of inactivation and removal
through alternative treatment technologies (e.g., ultraviolet radiation (UV)) approved by
the State, if the alternative treatment technology, alone or in combination (e.g., UV with
filtration, chlorination with filtration), can reliably provide at least 4-log treatment of
viruses.
Under the final GWR, GWSs providing 4-log treatment of viruses using chemical disinfection
must monitor for and must meet and maintain a State-determined residual disinfectant concentration (i.e.,
4-log inactivation of viruses based on CT tables) or State-approved alternatives every day the GWS
serves from the ground water source to the public. If the State has not approved compliance criteria for
Economic Analysis for the 3-9 October 2006
Final Ground Water Rule
-------
the system to use to demonstrate 4-log treatment by the time that the system is required to conduct
compliance monitoring, the system must comply with ground water source monitoring in '141.402 until
the State approves compliance criteria for the system to use to demonstrate 4-log treatment.
Systems serving greater than 3,300 people and using chemical disinfection (e.g., chlorine) to
provide 4-log inactivation must continuously monitor the residual disinfectant concentration using
analytical methods specified in 40 CFR 141.74(a)(2) (Analytical and monitoring requirements) at a
location approved by the State, and record the lowest residual disinfectant level each day that the GWS
serves water from the ground water source to the public. The GWS must maintain the State-determined
residual disinfectant concentration every day the GWS serves from the ground water source.
Systems serving 3,300 people or fewer that use chemical disinfection must monitor the residual
disinfectant concentration using analytical methods specified in 40 CFR 141.74(a)(2) (Analytical and
monitoring requirements) at a location approved by the State either by taking at least one grab sample
every day the GWS serves water to the public or by continuously monitoring the disinfectant residual.
Systems collecting grab samples must record the disinfectant residual level each day that the GWS serves
water from the ground water source to the public. The GWS must take a grab sample during the hour of
peak flow or at another time specified by the State. Systems serving 3,300 people or fewer that use
continuous residual monitoring equipment must record the lowest residual disinfectant level each day that
the GWS serves water from the ground water source to the public.
If a GWS taking grab samples has a sample measurement that falls below the State-specified
residual disinfectant concentration, then the system must take follow-up samples at least every four hours
until the State specified residual disinfectant level is restored. If a system using continuous monitoring
equipment fails to maintain the State specified disinfectant residual level necessary to achieve 4 log
inactivation of viruses, the system must restore the disinfectant residual level to the State specified level
within four hours. If continuous disinfectant monitoring equipment fails, the GWS must take a grab
sample at least every four hours until the equipment is back on-line. The system has 14 days to resume
continuous monitoring. Failure to restore the residual disinfectant level to that required for 4-log
inactivation of viruses within four hours, using either continuous monitoring or grab sampling, is a
treatment technique violation.
Ground water systems that use a membrane filtration treatment technology must maintain the
integrity of the membrane and monitor and operate the membrane filtration system in accordance with
State-specified monitoring and compliance requirements (e.g., membrane performance parameters and
integrity testing). If a system fails to meet these requirements or maintain the integrity of the membrane,
it must correct the problem within four hours or be in violation of the treatment technique requirement.
Systems that use a State-approved alternative treatment technology must monitor and operate the
alternative treatment in accordance with all compliance requirements that the State determines to be
necessary to demonstrate that at least 4-log treatment of viruses is achieved. If the system does not
comply with these requirements, fails to maintain at least 4-log treatment of viruses, and does not restore
proper operation within four hours, the system is in violation of the treatment technique requirement.
GWSs providing at least 4-log treatment of viruses may discontinue treatment if the State
determines (e.g., based on source water monitoring or replacement of the source) and documents in
writing that the need for 4-log treatment of viruses no longer exists for that ground water source. GWSs
that discontinue treatment with State approval must comply with the triggered source water requirements
of this rule. GWSs that provide 4-log treatment of viruses and notify the State that they are not subject to
the source water monitoring requirements of this rule but subsequently discontinue 4-log treatment of
Economic Analysis for the 3-10 October 2006
Final Ground Water Rule
-------
viruses must have State approval and must comply with the triggered source water requirements of the
final GWR.
3.4.2 Optional Provision
The final GWR provides States with the option to require systems to conduct assessment source
water monitoring at any time and require systems to take corrective action. EPA believes that this
optional provision is an important tool that should be used by States to protect public health. States may
elect to require assessment source water monitoring on a case-by-case basis. EPA recommends that
States require GWSs that are most susceptible to fecal contamination to conduct assessment monitoring.
States may use hydrogeologic sensitivity assessments (HSAs) as a tool to identify high risk systems for
assessment source water monitoring. States have other information available to them to target high risk
systems, such as source water assessments, wellhead protection plans, and historical monitoring data.
Data on past indications of source water fecal contamination, particularly from TCR monitoring, in
combination with GWR triggered source water monitoring results, can be another important tool.
EPA recommends that States require GWSs that are conducting assessment source water
monitoring to collect a total of 12 ground water source samples that represent each month the GWS
provides ground water to the public. For seasonal systems, EPA recommends equally distributing 12
samples or sampling during consecutive years. EPA recommends that States require corrective action for
sources that are fecally contaminated.
3.5 Other Changes Since Proposal
In addition to modifications made to Alternative 2 since proposal, updates were made based on
comments to the assumptions, data, and analytical processes used to support the economic analysis of the
alternatives. Wherever possible, assumptions and data were updated to reflect the latest information on
the characteristics and number of entities affected, occurrence of contaminants, costs of items used as
modeling inputs, and public comment. In particular, extensive consideration was given to the appropriate
use of occurrence data to inform the benefits and cost analyses. EPA published a Notice of Data
Availability in the March 27, 2006 Federal Register to present additional occurrence studies the Agency
considered using in the final GWR economic analysis.
See Appendix J of this EA for a detailed discussion on changes in the GWR economic analysis
since proposal.
Economic Analysis for the 3-11 October 2006
Final Ground Water Rule
-------
4. Baseline Analysis
4.1 Introduction
The baseline analysis is a characterization of the industry and its operations under the conditions
expected to exist before systems make changes to meet requirements of the GWR. The baseline allows a
consistent comparison of public health impacts (developed in Chapter 5) and the economic and financial
impacts (developed in Chapters 6 and 7) of the rule. Development of the GWR baseline consists of the
following processes:
•• Compiling an industry profile—identifying and collecting information on the segment(s)
of the water supply industry subject to the GWR.
•• Characterizing current disinfection practices of ground water systems—summarizing the
status of disinfection practices currently employed to ensure public health protection
from ground water contaminants.
•• Characterizing current ground water quality—summarizing the relevant characteristics of
ground water sources.
Section 4.2 characterizes the water industry, including the baseline estimates of systems, entry
points, and population subject to the GWR. This section also includes an assessment of the current status
of disinfection practices among the systems potentially affected by the rule as well as estimates of annual
positive total coliform samples under the Total Coliform Rule (TCR). Source water quality is
summarized in section 4.3, and the recent history of outbreaks is presented in section 4.4. Lastly, section
4.5 itemizes and estimates the effects of significant uncertainties in the baseline analysis.
This chapter presents an analysis that is at a level of detail and precision appropriate to support
subsequent analyses and regulatory decisions under consideration for the GWR. Therefore, it does not
give an exhaustive review of the water supply industry, source waters, or industry practices.
4.2 Industry Profile
This section provides a water industry characterization that is used to derive costs and benefits for
the GWR. It is organized as follows:
Section 4.2.1 describes the data sources used to characterize the industry baseline.
Section 4.2.2 is a background section describing the various ways in which water systems can be
classified and identifies distinctions that are important for regulatory analysis.
Section 4.2.3 presents the baseline numbers of systems, entry points, and population according to
disinfection practices used for estimating treatment costs and subsequent benefits of the GWR.
Economic Analysis for the 4-1 October 2006
Final Ground Water Rule
-------
Section 4.2.4 presents the mean plant design and average daily flow for each of the nine system
size categories.
Section 4.2.5 presents the treatment practices baseline.
Section 4.2.6 estimates the total number of households in each of the nine size categories of
systems subject to the GWR.
Section 4.2.7 estimates the annual number of triggered monitoring samples that systems will have
to take as a result of positive total coliform samples taken under the Total Coliform Rule.
4.2.1 Data Sources
Several data sources were used to characterize the GWR baseline. Data from the Safe Drinking
Water Information System-Federal Version (SDWIS/FED or SDWIS) are used to create system and
population baselines (USEPA, 2003)1. SDWIS is the United States Environmental Protection Agency's
(EPA or Agency) national regulatory compliance database for the drinking water program. It includes
information on the nation's 170,000 public water systems (PWSs) and on violations of drinking water
regulations. For more information on SDWIS, refer to EPA's website
(http://www.epa.gov/safewater/sdwisfed/sdwis.htm). A second key source of data used to develop the
industry profile is the Third Edition of the Water Industry Baseline Handbook (Baseline Handbook)
(USEPA, 200la) published in May 2001, which compiles data derived from the 1995 Community Water
System Survey (CWSS) and SDWIS. For certain analyses, CWSS raw data were used to develop
modeling inputs. The 1995 CWSS was a mail survey that covered ground and surface water systems of
all sizes (based on population served). The survey was based on a two-phase, stratified, random sample
design. Phase 1 was a telephone screening survey that provided a sampling frame for the main data
collection in Phase 2. The survey sample in Phase 2 was stratified according to water system size
(residential population served), ownership (public, private, or ancillary), and primary water source
(ground or surface). A total of 3,681 systems covering a range of source water types and system sizes
were selected to receive the main survey questionnaire. Of these, 1,980 systems responded. See the EPA
Report, "Community Water System Survey, Volume 2" (USEPA, 1997a), for more information on the
1995 CWSS sample design and data evaluation.
EPA also used the December 2000 document, "Geometries and Characteristics of Water Systems
Report" (Model Systems Report) (USEPA, 2000a). In this document, EPA analyzed 1995 CWSS data to
create equations relating flow and population, among other things.
4.2.2 Water System Characterization
Categorization of water systems is important because system size, ownership, and
consecutive/wholesale relationships affect the way in which costs and benefits are estimated. This section
explains the classifications of water systems, as defined by EPA's National Primary Drinking Water
Regulations (NPDWRs) and describes further subdivisions according to water source, size (population
served), and ownership for regulatory analysis purposes.
Data used are from the 4th quarter freeze of the 2003 database.
Economic Analysis for the 4-2 October 2006
Final Ground Water Rule
-------
PWS Type
NPDWRs apply to all PWSs. A PWS is defined as a system that provides water for human
consumption through pipes or other constructed conveyances if such a system has at least 15 service
connections or regularly serves an average of at least 25 individuals per day for at least 60 days per year.
PWSs are categorized as follows:
•• Community Water Systems (CWSs) are PWSs that have at least 15 service connections
used by year-round residents or that regularly serve at least 25 year-round residents.
•• Noncommunity Water Systems (NCWSs) are PWSs that are not classified as CWSs.
NCWSs are subdivided into two categories:
•• Nontransient Noncommunity Water Systems (NTNCWSs) are NCWSs that regularly
serve at least 25 of the same people more than 6 months per year.
•• Transient Noncommunity Water Systems (TNCWSs) are NCWSs that do not regularly
serve at least 25 of the same people more than 6 months per year.
Source Water Type
Systems are classified by the source from which they draw water. Systems that use either surface
water or ground water under the direct influence of surface water (GWUDI) are classified as surface
water systems. Ground water systems are, by default, systems that draw from ground water that are not
GWUDI.
Some systems may obtain water from both ground water and surface water and are referred to as
"mixed systems." In SDWIS and the Baseline Handbook, a mixed system is categorized as a surface
water system because it gets some portion of its flow from surface water (i.e., all mixed systems are
considered surface water systems). Based on an analysis in the Geometries and Characteristics of Water
Systems Report (USEPA 2000a), it is estimated that 21 percent of systems classified as surface water
obtain some of their water from ground water sources. Furthermore approximately one-third of these, or
8 percent of all surface water systems in SDWIS and the Baseline Handbook, receive the majority of their
flow from ground water. The 1995 CWSS data are classified by primary source (i.e., if a system receives
more than 50 percent of its flow from ground water sources, it is considered a ground water system).
Population Served
Small systems are those serving fewer than 10,000 people. Systems are categorized in SDWIS
and the Baseline Handbook by retail population served (i.e., not including population of wholesale
customers). In the analyses that follow, nine size categories are most often used. System size is especially
important because smaller systems are expected to take different approaches to meet rule provisions than
large systems. Additionally, smaller systems are not able to achieve the same economies of scale as
larger systems for a given treatment technology. To account for these differences, both the compliance
decision tree and the unit costs for selected technologies use assumptions that are dependent on system
size.
Economic Analysis for the 4-3 October 2006
Final Ground Water Rule
-------
Ownership
Systems are categorized in SDWIS and in the Baseline Handbook according to three ownership
types: "private," "public," and "other." Private systems are owned by private corporations or individuals.
Public systems are owned by public entities such as municipalities, counties, or special districts. The
"other" category contains systems where ownership is not reported in SDWIS. Ownership distinctions
are important to the analysis because public systems have access to capital and other means of financing
that may not be available to private systems. This distinction becomes important in calculating household
costs (see Chapter 6) and in assessing the Unfunded Mandates Reform Act (UMRA) requirements (see
Chapter?).
Consecutive and Wholesale System Types
Systems are categorized according to whether they treat water themselves or purchase treated
water from other systems. The GWR defines a consecutive system as a PWS that buys or otherwise
receives some or all of its finished water from one or more wholesale systems for at least 60 days per
year. A wholesale system is defined as a PWS that treats and then sells or otherwise delivers finished
water to another PWS at least 60 days per year. Treatment modifications are generally not made by
consecutive water systems, but are instead made by the associated wholesale systems. Costs of these
treatment modifications are typically passed on to the consecutive systems in the form of water rate
increases.
4.2.3 Baseline Number of Systems, Entry Points, and Population
The GWR applies to all PWSs, regardless of their size, that use ground water as a source.
Further, because a person may need only ingest a small number of certain microbial pathogens (e.g.,
Shigella, enterovirus, rotavirus, norovirus) to become ill, this EA also considers the impact across all
types of ground water systems, including those noncommunity systems that provide drinking water only
part of the time or to a transient population. This section estimates the baseline number of systems, the
number of entry points, and, the size of the population subject to the GWR. These will be used to
estimate costs and benefits of the rule later in this EA.
Number of Systems
Estimates of the number of ground water PWSs subject to the GWR are presented in Exhibit 4.1.
Most of these use ground water as their only source and reflect the SDWIS ground water system
inventory. In addition to the systems served solely by ground water, PWSs served by multiple sources
(i.e., those using both ground and surface water) may be subject to rule requirements. Entry points
delivering only ground water are often present in PWSs that are classified as surface water systems in
accordance with the SDWIS classification scheme (i.e., SDWIS classifies a system as surface water if any
portion of its source water comes from a surface source). These "mixed water" systems, and associated
ground water entry points and individuals served by them, would be excluded from the ground water
system baseline if only the SDWIS ground water inventory were used, resulting in a potential
underestimate of rule costs and benefits. To account for ground water entry points in mixed systems,
EPA derived an inventory of "primarily ground water" mixed systems that is added to the ground water-
only system inventory.
Economic Analysis for the 4-4 October 2006
Final Ground Water Rule
-------
To derive the primarily ground water mixed system inventory, EPA identified surface water
CWSs that use primarily (more than 50 percent) ground water based on CWSS data, as presented in
Geometries and Characteristics of Public Water Systems (USEPA 2000a). NCWSs were not included in
the mixed system analysis due to lack of data on the existence of multiple sources within such systems.
Because NTNCWSs and TNCWSs are typically a single building or located in a small area, a simplifying
assumption was made for this analysis that all NCWSs draw from a single source. The primarily ground
water mixed CWSs identified by this calculation (862 systems, as shown in Exhibit 4.1, column P) were
added to the ground water inventory to produce the baseline number of ground water systems used in this
EA (Exhibit 4.1, columns Q through U).
The resulting baseline number of ground water systems are all treated as ground water-only
systems throughout subsequent analyses. This methodology, treating mixed systems as ground water-
only systems, may overestimate costs and benefits (i.e., some surface water entry points are now counted
as ground water entry points). However, the ground water entry points in the excluded mixed surface
water inventory (those mixed systems using less than 50 percent ground water) are not included in the
analysis, potentially underestimating costs and benefits. The contrasting over- and under-accounting for
ground water entry points are expected to offset one another in the cost and benefit analyses. Data are not
available to quantify the direction or magnitude of the final effect on overall national cost estimates, but
the effect is expected to be minimal.
In addition to the baseline number of systems presented in Exhibit 4.1, the national cost model
requires system-by-system data. For ground water-only systems, systems and their associated attributes
are taken directly from SDWIS. System identity data (e.g., PWS Identification (ID) and address) are not
included in the attributes because the data are used for national level analysis and are not meant for
analysis of specific systems. To derive the system-by-system data for primarily ground water mixed
systems, a representative sample of systems was selected from the SDWIS surface water inventory. This
was done as follows:
•• For each surface water system size category, all of the SDWIS systems were placed in
ascending order by population served.
•• Within this ordered system list, systems were selected at equal intervals based on the
overall number of primarily ground water mixed systems. For example, the SDWIS
surface water inventory includes a total of 1,163 systems serving 100 or fewer people.
Based on CWSS data, as presented in Geometries and Characteristics of Public Water
Systems (USEPA 2000a), 3.7 percent of these systems (43 systems) are primarily ground
water mixed systems. To select a representative sample of systems from the SDWIS
inventory for this size category, every 27th system (1,163/43) was selected from the
SDWIS surface water system list (ordered by population served) and assigned, along with
its attributes (minus system identity data), to the primarily ground water mixed system
data set.
Selection in this manner ensures that the systems chosen represent the full range of population served
within any given size category. The resulting system data were added to the SDWIS ground water data
for use in the cost model.
Economic Analysis for the 4-5 October 2006
Final Ground Water Rule
-------
Exhibit 4.1 Ground Water Rule System Baseline (continued on next page)
System Size
(population served)
Number of Ground Water Only Systems
Public
Purchased
Systems
A
Non-
Purchased
Systems
B
Pr vate
Purchased
Systems
C
Non-
Purchased
Systems
D
Total
Ground
Water Only
Systems
E
Number of Surface Water Systems
Public
Purchased
Systems
F
Non-
Purchased
Systems
G
Private
Purchased
Systems
H
Non-
Purchased
Systems
1
Total
Surface
Water
Systems
J
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Total
174
604
324
325
114
33
1
1
1,576
1,402
4,284
2,735
4,135
2,105
1,032
110
52
3
15,858
120
252
92
81
18
6
-
-
569
11,104
9,022
1,498
1,224
408
201
27
12
23,496
12,800
14,162
4,649
5,765
2,645
1,272
138
65
3
41,499
459
848
629
1,169
868
708
111
58
4,850
220
357
274
864
910
823
172
180
13
3,813
324
607
367
301
137
92
17
9
1,854
160
226
76
128
77
98
31
24
3
823
1,163
2,038
1,346
2,462
1,992
1,721
331
271
16
11,340
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Total
11
18
7
8
4
3
-
-
51
1,913
3,076
1,162
374
28
5
1
-
-
6,559
13
14
3
1
1
-
-
32
7,519
3,650
722
332
40
2
1
-
12,266
9,456
6,758
1,894
715
73
10
1
1
-
18,908
TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Total
Grand Total
71
33
18
9
1
1
1
-
-
134
1,761
51,730
3,773
622
268
41
9
1
-
56,444
78,861
521
45
9
5
2
1
-
-
583
1,184
12,126
15,142
1,291
303
30
8
-
-
28,900
64,662
64,448
18,993
1,940
585
74
19
1
1
-
86,061
146,468
4,850
3,813
1,854
823
11,340
Economic Analysis for the
Final Ground Water Rule
4-6
October 2006
-------
Exhibit 4.1 Ground Water Rule System Baseline (continued)
System Size
(population served)
Percentage
of SW that is
Primarily
(>50%) GW
K
Number of Primarily Ground Water Systems
Public
Purchased
Systems
L=F*K
Non-
Purchased
Systems
M=G*K
Private
Purchased
Systems
N=H*K
Non-
Purchased
Systems
O=I*K
Total
Primarily
Ground
Water
Systems
P=J*K
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Total
3.7%
9.6%
0.0%
5.9%
12.0%
10.0%
8.9%
14.0%
0.0%
17
81
69
104
71
10
8
-
360
8
34
51
109
82
15
25
-
325
12
58
18
16
9
2
1
-
116
6
22
8
9
10
3
3
-
60
43
196
145
239
172
29
38
-
862
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Total
TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Total
Grand Total
-
360
325
116
60
862
Ground Water System Baseline
Public
Purchased
Systems
Q=A+L
Non-
Purchased
Systems
R=B+M
Private
Purchased
Systems
S=C+N
Non-
Purchased
Systems
T=D+O
Total
Number of
Ground
Water
Systems
U=E+P
191
685
324
394
218
104
11
9
-
1,936
1,410
4,318
2,735
4,186
2,214
1,114
125
77
3
16,183
132
310
92
99
34
15
2
1
-
685
11,110
9,044
1,498
1,232
417
211
30
15
-
23,556
12,843
14,358
4,649
5,910
2,884
1,444
167
103
3
42,361
11
18
7
8
4
3
-
51
1,913
3,076
1,162
374
28
5
1
-
6,559
13
14
3
1
1
-
-
32
7,519
3,650
722
332
40
2
1
12,266
9,456
6,758
1,894
715
73
10
1
1
18,908
71
33
18
9
1
1
1
-
134
2,121
51,730
3,773
622
268
41
9
1
56,444
79,186
521
45
9
5
2
1
-
583
1,300
12,126
15,142
1,291
303
30
8
-
28,900
64,722
64,448
18,993
1,940
585
74
19
1
1
86,061
147,330
Notes: Surface water systems include mixed systems. Detail may not add to totals due to independent rounding.
Sources: (A-J) Ground water system inventories for CWSs, NTNCWSs, and TNCWSs: SDWIS (USEPA, 2003a).
(F-J) Surface water system inventory for CWSs: SDWIS (USEPA, 2003a).
(K) Percent of surface water CWSs served by more than 50% ground water from Geometries and Characteristics of Public Water Systems
(USEPA 2000a), Exhibit 2.9.
Number of Disinfecting Systems
The system inventory presented in Exhibit 4.1 represents the baseline for rule activities that are
applicable to PWSs on a system level (i.e., rule implementation and performance of sanitary surveys),
regardless of any other factors. The applicability of other rule requirements (i.e., triggered source water
monitoring and compliance monitoring) are dependent upon whether a system (or entry point) achieves 4-
log treatment of viruses (using inactivation, removal, or State-approved combination of these
technologies) before or at the first customer. Exhibit 4.2 shows the system baseline stratified according to
disinfecting and nondisinfecting systems.
Economic Analysis for the
Final Ground Water Rule
4-7
October 2006
-------
Exhibit 4.2 Ground Water Rule System Baseline: Disinfecting1 and
Nondisinfecting Systems
System Size
(population
served)
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Mllion
> 1 Mllion
National Totals
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3301-10,000
10,001-50,000
50,001-100,000
100,001-1 Mllion
> 1 Mllion
National Totals
TNCVWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Mllion
> 1 Mllion
National Totals
Percentage2
of Systems
Dsinfectirg
A
52.8%
77.9%
84.0%
79.7%
86.8%
96.5%
86.3%
96.4%
100.0%
29%
29%
29%
29%
29%
29%
29%
29%
29%
18%
18%
18%
18%
18%
18%
18%
18%
18%
Grand Totals
Number of Dsinfecting Ground V\ater Systems
Public
Purchased
Systems
&Sys*A
101
534
272
314
189
100
9
9
-
1,529
3
5
2
2
1
1
-
-
-
15
13
6
3
2
0
0
0
-
-
24
1,568
Non-
Purchased
Systems
C=Sys*A
745
3,364
2,297
3,336
1,922
1,075
108
74
3
12,925
555
892
337
108
8
1
0
-
-
1,902
9,311
679
112
48
7
2
-
0
-
10,160
24,937
Private
Purchased
Systems
F>Sys*A
70
242
77
79
30
15
1
1
-
514
4
4
1
0
0
-
-
-
-
9
94
8
2
1
0
0
-
-
-
105
629
Non-
Purchased
Systems
&Sys*A
5,866
7,045
1,258
962
362
203
26
15
-
15,757
2,181
1,059
209
96
12
1
-
0
-
3,557
2,183
2,726
232
55
5
1
-
-
-
5,202
24,516
Total
P=Sys*A
6,781
11,185
3,905
4,710
2,503
1,394
145
99
3
30,725
2,742
1,960
549
207
21
3
0
0
-
5,483
11,601
3,419
349
105
13
3
0
0
-
15,491
51,699
Nurrber of Nondisinfecting Ground V\ater Systems
Public
Purchased
Systems
GSys*
(1-A)
90
151
52
80
29
4
1
0
-
408
8
13
5
6
3
2
-
-
-
36
58
27
15
7
1
1
1
-
-
110
554
Non-
Purchased
Systems
H^ys*
(1-A)
666
954
438
850
292
39
17
3
-
3,258
1,358
2,184
825
266
20
4
1
-
-
4,657
42,419
3,094
510
220
34
7
-
1
-
46,284
54,199
Private
Purchased
Systems
l=Sys*
(1-A)
62
69
15
20
5
1
0
0
-
171
9
10
2
1
1
-
-
-
-
23
427
37
7
4
2
1
-
-
-
478
672
Non-
Purchased
Systems
J=Sys*
(1-A)
5,244
1,999
240
250
55
7
4
1
-
7,799
5,338
2,592
513
236
28
1
-
1
-
8,709
9,943
12,416
1,059
248
25
7
-
-
-
23,698
40,206
Total
K=Sys*
(1-A)
6,062
3,173
744
1,200
381
51
23
4
-
11,636
6,714
4,798
1,345
508
52
7
1
1
-
13,425
52,847
15,574
1,591
480
61
16
1
1
-
70,570
95,631
Notes: Detail may not add to totals due to independent rounding.
Footnotes: 1) "Dsinfection" refers to primary disinfection that is intended for rricrobial inactivation; 2) Percentages shown are weighted for item-level nonresponse, as
indicated in Table 6.1 of "Gsorretries and Characteristics of V\ater Systems Report" (Model Systems Report) (USEPA 2000a)."
Sources:(A) CW3 percent disinfecting from Third Edition of the Baseline Handbook, Table B1.3.3, except for systems serving >1 million people. The three ground water
systems serving >1 trillion people all perform disinfection. NTNQA6 and TNQA6 percent disinfecting derived from Ground V\ater Dsinfection Practices in the United
States (USEPA 1996b).
(B-K) Sys = System inventory from Exhibit 4.1, columns Q-U.
Economic Analysis for the
Final Ground Water Rule
4-&
October 2006
-------
Number of Entry Points
The GWR benefits and cost models use entry points as a metric for exposure potential and for
estimating the number of points of treatment and monitoring in response to GWR requirements. Ground
water systems can consist of one entry point supplying all water to the population, or multiple entry
points treating water, possibly from different sources.
As with the system baseline, the GWR requirements at the entry point level also depend on
whether disinfection is applied. In order to estimate baseline disinfection rates, EPA analyzed data from a
survey of disinfection practices in CWSs serving fewer than 10,000 people (AWWA, 1998). In this
analysis, EPA used the data on the number of systems that apply disinfection prior to entry to the
distribution system, the flow rates, the volume of the distribution system prior to the first customer, and
the contact time (CT) value required for inactivation of Hepatitis A virus (HAV) at a temperature of 15
degrees C, and a pH of 6-9. EPA assumed that those systems providing insufficient information for the
CT calculation in the AWWA survey are not currently achieving 4-log virus inactivation. Where
additional assumptions were necessary, they were made to not overestimate the amount of 4-log treatment
in place (e.g., maximum flow rate), and any treatment provided prior to distribution was not accounted
for. Based on the evaluation, EPA found that 52 percent of small community ground water systems
applying disinfection met 4-log inactivation of viruses prior to the first customer. The number was also
applied to community ground water systems serving 10,000 or more people.
Because the AWWA survey did not have any data on NCWSs, EPA used best professional
judgement to estimate the percent of NCWSs that achieved 4-log inactivation of viruses before the first
customer. Because of their small size and usually simple design as well as the transient nature of the
populations they serve, TNCWSs are believed to be the least likely systems to achieve 4-log inactivation.
Based on this characterization, EPA estimates that only 10 percent of TNCWSs that apply disinfection
(see Exhibit 4.3) achieve 4-log inactivation of viruses before the first customer. Many NTNCWSs are
relatively simple systems that share many characteristics with TNCWSs. However, many also share
characteristics more like CWSs (e.g., systems serving large institutions). Because of this dichotomy, EPA
estimated that the percentage of NTNCWSs that apply disinfection that are achieving 4-log inactivation
of viruses before the first customer falls between the percentages used for CWSs and TNCWSs, or 31
percent. Exhibit 4.3 shows the entry point baseline used for GWR analyses.
Population
System population characteristics are important to this analysis for several reasons. It is
important to know the total population served by ground water systems as well as the average population
served by each entry point so that the distribution of costs and benefits of the GWR can be assessed.
Exhibit 4.4 presents the system-level population baseline and Exhibit 4.5 presents the entry point-level
baseline.
As presented in Exhibit 4.4, ground water CWSs serve over 100 million people, while ground
water NCWSs serve about 14 million people. Overlaps do occur because individuals may be served by
both types of systems. For example, a person may be served by a surface water CWS at home and by a
ground water NCWS at work or at a restaurant. It should be noted that there does not appear to be a
consistent reporting standard for populations served by transient systems. In addition, some States may
report the total population served by a system over a year, while others may report the average population
served each day.
Economic Analysis for the 4-9 October 2006
Final Ground Water Rule
-------
Exhibit 4.3 Ground Water Rule Entry Point Baseline (continued on next page)
System Size
(population
served)
Entry
Points
per
System
A
Percentage
of Entry
Points
Disinfecting
B
Disinfecting
Entry Points
Achieving
4-Log
C
Number of Ground Water Entry Points D sinfecting to 4-Log
Public
Purchased Entry
Points
D=Sys*A*B*C
Non-Purchased
Entry Points
E=Sys*A*B*C
Pr vate
Purchased Entry
Points
F=Sys*A*B*C
Non-Purchased
Entry Points
G=Sys*A*B*C
Total
H=Sys*A*B*C
Number of Ground Water Entry Points Not Disinfecting or
Dis
Public
Purchased
Entry Points
l=Sys*
A*(1-B*C)
Non-Purchased
Entry Points
J=Sys*
A*(1-B*C)
nfecting to<4-Log
Private
Purchased
Entry Points
K=Sys*
A*(1-B*C)
Non-
Purchased
Entry Points
L=Sys*
A*(1-B*C)
Total
IVl=Sys*
A*(1-B*C)
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
1.3
1.6
2.0
2.4
3.2
5.6
11.3
12.4
11.4
45.6%
72.9%
75.1%
72.1%
73.5%
91.4%
59.3%
82.2%
100.0%
52.0%
52.0%
52.0%
52.0%
52.0%
52.0%
52.0%
52.0%
100.0%
59
424
247
358
268
277
38
48
-
1,720
439
2,669
2,087
3,809
2,723
2,975
436
409
34
15,581
41
192
70
90
42
41
5
7
-
488
3,457
5,589
1,143
1,121
513
563
104
81
-
12,570
3,996
8,873
3,547
5,378
3,547
3,856
583
545
34
30,359
191
694
386
598
434
306
85
65
-
2,758
1,412
4,371
3,257
6,350
4,402
3,285
979
548
-
24,603
132
314
110
150
68
45
12
9
-
840
11,122
9,155
1,784
1,868
829
621
232
109
-
25,721
12,857
14,534
5,536
8,966
5,734
4,257
1,308
730
-
53,921
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
29%
29%
29%
29%
29%
29%
29%
29%
29%
31%
31%
31%
31%
31%
31%
31%
31%
31%
1
2
1
1
0
0
-
-
5
172
277
104
34
3
0
0
-
590
1
1
0
0
0
-
-
3
676
328
65
30
4
0
-
0
-
1,103
850
608
170
64
7
1
0
0
-
1,700
10
16
6
7
4
3
-
-
46
1,741
2,799
1,058
340
25
5
1
-
5,969
12
13
3
1
1
-
-
29
6,843
3,322
657
302
36
2
-
1
-
11,163
8,606
6,150
1,724
651
66
9
1
1
-
17,208
TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
18%
18%
18%
18%
18%
18%
18%
18%
18%
10%
10%
10%
10%
10%
10%
10%
10%
10%
Grand Totals
1
1
0
0
0
0
0
-
2
1,727
931
68
11
5
1
0
-
0
-
1,016
17,186
9
1
0
0
0
0
-
-
10
501
218
273
23
5
1
0
-
-
520
14,193
1,160
342
35
11
1
0
0
0
-
1,549
33,608
70
32
18
9
1
1
1
-
132
2,936
50,799
3,705
611
263
40
9
-
1
-
55,428
86,000
512
44
9
5
2
1
-
-
573
1,441
11,908
14,869
1,268
298
29
8
-
-
28,380
65,264
63,288
18,651
1,905
574
73
19
1
1
-
84,512
155,641
Economic Analysis for the
Final Ground Water Rule
4-10
October 2006
-------
Exhibit 4.3 Ground Water Rule Entry Point Baseline (continued)
Number of Ground Water Entry Points Not Disinfecting
Public
Purchased
Entry Points
N=Sys*
A*(1-B)
Non-
Purchased
Entry Points
O=Sys*
A*(1-B)
Private
Purchased
Entry Points
P=Sys*
A*(1-B)
Non-
Purchased
Entry Points
Q=Sys*
A*(1-B)
Total
R=Sys*
A*(1-B)
Number of Ground Water Entry Points Disinfecting to <4-Log
Public
Purchased
Entry Points
S=I-N
Non-
Purchased
Entry Points
T=J-O
Private
Purchased
Entry Points
U=K-P
Non-
Purchased
Entry Points
V=L-Q
Total
W=M-R
136
303
158
267
186
50
50
20
1,170
1,007
1,908
1,331
2,834
1,888
538
576
170
10,252
94
137
45
67
29
7
7
3
389
7,931
3,996
729
834
356
102
137
34
14,117
9,168
6,343
2,262
4,002
2,459
698
770
227
25,929
55
391
228
331
248
256
35
45
1,588
405
2,463
1,926
3,516
2,514
2,746
403
377
14,351
38
177
65
83
39
37
5
6
450
3,191
5,159
1,055
1,034
474
520
96
75
11,604
3,689
8,191
3,274
4,964
3,274
3,559
538
503
27,992
8
13
5
6
3
2
-
36
1,358
2,184
825
266
20
4
1
-
4,657
9
10
2
1
1
-
-
23
5,338
2,592
513
236
28
1
1
8,709
6,714
4,798
1,345
508
52
7
1
1
13,425
2
4
1
2
1
1
-
10
383
616
233
75
6
1
0
-
1,312
3
3
1
0
0
-
-
6
1,505
730
144
66
8
0
0
2,454
1,892
1,352
379
143
15
2
0
0
3,783
58
27
15
7
1
1
1
-
110
1,316
42,419
3,094
510
220
34
7
-
1
46,284
61,193
427
37
7
4
2
1
-
-
478
890
9,943
12,416
1,059
248
25
7
-
-
23,698
46,524
52,847
15,574
1,591
480
61
16
1
1
70,570
109,923
12
5
3
1
0
0
0
-
22
1,620
8,380
611
101
43
7
1
-
0
9,144
24,807
84
7
1
1
0
0
-
-
94
551
1,964
2,453
209
49
5
1
-
-
4,682
18,740
10,441
3,077
314
95
12
3
0
0
13,942
45,718
Note: Detail may not add to totals due to independent rounding.
Sources: (A) CWS entry points derived from Question 18 and 20 of the 1995CWSS. NTNCWS and TNCWS entry points to system ratio assumed to be 1:1
because these systems are most often housed in a single building or small area.
(B) CWS percent disinfecting from Third Edition of the Baseline Handbook, Table B1.3.5, except for systems serving >1 million people. Ground water systems
serving >1 million people (3) all perform disinfection. NTNCWS and TNCWS percent disinfecting derived from Ground Water Disinfection Practices in the
United States (USEPA 1996b).
(C) Percentage of disinfecting entry points achieving 4 log derived from AWWA data (1998).
(D-R) Sys = System inventory from Exhibit 4.1, columns T-X.
Economic Analysis for the
Final Ground Water Rule
4-11
October 2006
-------
Exhibit 4.4 Ground Water Rule System Population Baseline
System Size
(population
served)
Average
Population
per System
A
Disinfecting Systems Population
Public
Purchased
B=Sys*A
Non-
Purchased
C=Sys*A
Private
Purchased
D=Sys*A
Non-
Purchased
E=Sys*A
Total
F=Sys*A
Nondisinfecting Systems Population
Public
Purchased
G=Sys*A
Non-
Purchased
H=Sys*A
Private
Purchased
l=Sys*A
Non-
Purchased
J=Sys*A
Total
K=Sys*A
Total
Population
L=F+K
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
58
235
712
1,839
5,632
20,569
65,030
207,247
1,311,178
5,882
125,587
193,786
577,381
1,066,489
2,060,372
610,540
1,822,052
-
6,462,088
43,427
791,234
1,635,819
6,134,719
10,824,258
22,118,232
7,032,405
15,423,507
3,933,533
67,937,134
4,065
56,851
55,026
144,735
168,362
301,712
84,911
251,731
-
1,067,392
342,143
1,657,070
895,962
1,804,890
2,039,704
4,184,262
1,670,104
3,068,718
-
15,662,854
395,517
2,630,742
2,780,593
8,661,725
14,098,814
28,664,578
9,397,959
20,566,008
3,933,533
91,129,468
5,258
35,629
36,912
147,062
162,185
74,729
96,922
68,043
-
626,739
38,821
224,471
311,585
1,562,545
1,646,085
802,216
1,116,384
575,982
-
6,278,088
3,634
16,128
10,481
36,865
25,603
10,943
13,479
9,401
-
126,535
305,855
470,106
170,659
459,715
310,185
151,761
265,127
114,599
-
2,248,008
353,568
746,334
529,637
2,206,186
2,144,059
1,039,648
1,491,912
768,025
-
9,279,369
749,084
3,377,075
3,310,229
10,867,911
16,242,873
29,704,225
10,889,872
21,334,033
3,933,533
100,408,836
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
50
226
673
1,552
5,153
20,764
66,000
110,000
161
1,180
1,366
3,601
5,978
18,065
-
30,350
27,985
201,651
226,696
168,338
41,845
30,108
19,140
-
715,762
190
918
585
450
1,494
-
3,638
109,993
239,280
140,856
149,433
59,779
12,043
31,900
743,285
138,329
443,028
369,502
321,822
109,096
60,217
19,140
31,900
1,493,035
394
2,889
3,343
8,816
14,636
44,228
-
74,306
68,514
493,696
555,013
412,137
102,449
73,713
46,860
-
1,752,383
466
2,247
1,433
1,102
3,659
-
8,906
269,294
585,823
344,853
365,854
146,355
29,485
78,100
1,819,766
338,668
1,084,655
904,643
787,909
267,098
147,427
46,860
78,100
3,655,361
476,998
1,527,684
1,274,145
1,109,731
376,195
207,644
66,000
110,000
5,148,396
TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
38
177
641
1,416
5,017
15,770
51,850
125,000
-
Grand Totals
488
1,051
2,076
2,294
903
2,839
9,333
-
18,984
6,511,422
355,608
120,171
71,722
68,320
37,029
25,547
-
22,500
-
700,897
69,353,793
3,582
1,433
1,038
1,275
1,806
2,839
-
-
11,972
1,083,002
83,358
482,276
148,864
77,242
27,094
22,709
-
-
841,543
17,247,682
443,036
604,932
223,700
149,130
66,832
53,933
9,333
22,500
-
1,573,397
94,195,899
2,223
4,788
9,455
10,452
4,114
12,931
42,517
-
86,482
787,526
1,619,993
547,445
326,736
311,234
168,687
116,382
-
102,500
-
3,192,977
11,223,447
16,316
6,529
4,728
5,807
8,229
12,931
-
-
54,539
189,980
379,742
2,197,036
678,160
351,880
123,429
103,451
-
-
3,833,698
7,901,471
2,018,274
2,755,799
1,019,079
679,372
304,459
245,696
42,517
102,500
-
7,167,696
20,102,425
2,461,310
3,360,731
1,242,779
828,502
371,291
299,629
51,850
125,000
-
8,741,092
114,298,324
Note: Figures are derived using unrounded source data. Detail may not add to totals due to independent rounding.
Sources: (A) Derived from SDWIS (USEPA, 2003a).
(B-K) Sys = System inventory from Exhibt4.2, columns B-K.
Economic Analysis for the
Final Ground Water Rule
4-12
October 2006
-------
Exhibit 4.5 Ground Water Rule Entry Point Population Baseline
(continued on next page)
System Size
(population
served)
Average
Population
per Entry
Point
A
Population of Entry Points Disinfecting to 4-Log
Public
Purchased
B=EP*A
Non-
Purchased
C=EP*A
Private
Purchased
D=EP*A
Non-
Purchased
E=EP*A
Total
Disinfecting to
4-log
F=EP*A
Population of Entry Points Not Disinfecting or
Disinfecting to <4-Log
Public
Purchased
G=EP*A
Non-
Purchased
H=EP*A
Private
Purchased
I=EP*A
Non-
Purchased
J=EP*A
Total not
Disinfecting
to 4-log
K=EP*A
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
44
144
364
758
1,750
3,662
5,758
16,727
115,450
2,641
61,114
90,092
271,608
469,599
1,014,771
218,153
807,902
-
2,935,880
19,503
385,033
760,500
2,885,858
4,766,165
10,893,630
2,512,761
6,838,821
3,933,533
32,995,805
1,825
27,665
25,582
68,086
74,134
148,598
30,340
111,618
-
487,848
153,653
806,370
416,537
849,046
898,128
2,060,825
596,748
1,360,677
-
7,141,984
177,623
1,280,182
1,292,711
4,074,597
6,208,026
14,117,824
3,358,001
9,119,019
3,933,533
43,561,516
8,498
100,102
140,606
452,834
759,075
1,120,330
489,309
1,082,193
-
4,152,947
62,745
630,671
1,186,903
4,811,406
7,704,178
12,026,817
5,636,028
9,160,667
-
41,219,417
5,873
45,314
39,925
113,515
119,832
164,056
68,051
149,513
-
706,079
494,345
1,320,806
650,084
1,415,559
1,451,762
2,275,198
1,338,483
1,822,641
-
10,768,878
571,461
2,096,894
2,017,519
6,793,314
10,034,847
15,586,401
7,531,871
12,215,014
-
56,847,321
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
50
226
673
1,552
5,153
20,764
66,000
110,000
50
366
423
1,116
1,853
5,600
-
-
9,409
8,675
62,512
70,276
52,185
12,972
9,334
5,933
-
221,886
59
285
181
140
463
-
-
1,128
34,098
74,177
43,665
46,324
18,531
3,733
-
9,889
230,418
42,882
137,339
114,546
99,765
33,820
18,667
5,933
9,889
462,841
505
3,703
4,286
11,300
18,760
56,693
-
-
95,247
87,824
632,835
711,433
528,290
131,322
94,488
60,067
-
2,246,259
597
2,880
1,837
1,413
4,690
-
-
-
11,416
345,190
750,926
442,044
468,963
187,603
37,795
-
100,111
2,332,632
434,116
1,390,345
1,159,599
1,009,966
342,375
188,976
60,067
100,111
4,685,555
TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
38
177
641
1,416
5,017
15,770
51,850
125,000
-
Grand Totals
49
105
208
229
90
284
933
-
-
1,898
2,947,187
35,561
12,017
7,172
6,832
3,703
2,555
-
2,250
-
70,090
33,287,781
358
143
104
127
181
284
-
-
-
1,197
490,172
8,336
48,228
14,886
7,724
2,709
2,271
-
-
-
84,154
7,456,556
44,304
60,493
22,370
14,913
6,683
5,393
933
2,250
-
157,340
44,181,696
2,663
5,734
11,323
12,517
4,927
15,486
50,917
-
-
103,567
4,351,761
1,940,041
655,599
391,286
372,721
202,013
139,375
-
122,750
-
3,823,784
47,289,460
19,539
7,819
5,662
6,954
9,854
15,486
-
-
-
65,314
782,810
454,764
2,631,085
812,138
421,398
147,814
123,889
-
-
-
4,591,087
17,692,597
2,417,006
3,300,237
1,220,409
813,589
364,608
294,236
50,917
122,750
-
8,583,753
70,116,628
Economic Analysis for the
Final Ground Water Rule
4-13
October 2006
-------
Exhibit 4.5 Ground Water Rule Entry Point Population Baseline
(continued)
Population of Entry Points Not Disinfecting
Public
Purchased
L=EP*A
Non-
Purchased
M=EP*A
Private
Purchased
N=EP*A
Non-
Purchased
0=EP*A
Total Not
Disinfecting
P=EP*A
Population of Entry Points Disinfecting to <4-Log
Public
Purchased
Q=EP*A
Non-
Purchased
R=EP*A
Private
Purchased
S=EP*A
Non-
Purchased
T=EP*A
Total
Disinfecting
to <4-log
U=EP*A
6,060
43,689
57,444
202,119
325,599
183,619
287,937
336,437
1,442,904
44,743
275,256
484,903
2,147,537
3,304,641
1,971,158
3,316,557
2,847,909
14,392,704
4,188
19,777
16,311
50,666
51,401
26,888
40,045
46,481
255,758
352,511
576,465
265,589
631,825
622,721
372,898
787,639
566,631
4,176,277
407,502
915,187
824,247
3,032,147
4,304,361
2,554,563
4,432,178
3,797,458
20,267,644
2,438
56,413
83,162
250,715
433,476
936,711
201,372
745,756
2,710,043
18,002
355,415
702,000
2,663,869
4,399,537
10,055,659
2,319,471
6,312,758
26,826,712
1,685
25,537
23,614
62,848
68,431
137,168
28,006
103,032
450,321
141,834
744,341
384,496
783,734
829,041
1,902,300
550,844
1,256,010
6,592,600
163,960
1,181,706
1,193,271
3,761,167
5,730,486
13,031,838
3,099,693
8,417,556
36,579,676
394
2,889
3,343
8,816
14,636
44,228
-
;
74,306
68,514
493,696
555,013
412,137
102,449
73,713
46,860
;
1,752,383
466
2,247
1,433
1,102
3,659
-
-
;
8,906
269,294
585,823
344,853
365,854
146,355
29,485
-
78,100
1,819,766
338,668
1,084,655
904,643
787,909
267,098
147,427
46,860
78,100
3,655,361
111
814
942
2,485
4,125
12,465
-
;
20,942
19,309
139,139
156,420
116,153
28,873
20,775
13,207
;
493,876
131
633
404
311
1,031
-
-
;
2,510
75,896
165,103
97,190
103,109
41,247
8,310
-
22,011
512,866
95,447
305,689
254,956
222,057
75,277
41,549
13,207
22,011
1,030,194
2,223
4,788
9,455
10,452
4,114
12,931
42,517
;
86,482
1,603,691
1,619,993
547,445
326,736
311,234
168,687
116,382
-
102,500
3,192,977
19,338,064
16,316
6,529
4,728
5,807
8,229
12,931
-
;
54,539
319,204
379,742
2,197,036
678,160
351 ,880
123,429
103,451
-
;
3,833,698
9,829,741
2,018,274
2,755,799
1,019,079
679,372
304,459
245,696
42,517
102,500
7,167,696
31,090,701
439
946
1,868
2,065
813
2,555
8,400
;
17,085
2,748,070
320,047
108,154
64,550
61,488
33,326
22,993
-
20,250
630,808
27,951,396
3,223
1,290
934
1,147
1,626
2,555
-
;
10,775
463,606
75,022
434,049
133,978
69,518
24,385
20,438
-
;
757,389
7,862,856
398,732
544,438
201,330
134,217
60,149
48,540
8,400
20,250
1,416,057
39,025,927
Note: Figures are derived using unrounded source data. Detail may not add to totals due to independent rounding.
Sources: (A) Average population per system from Exhibit 4.4, column A divided by entry points per system from Exhibit 4.3, column A.
(B-U) EP = Entry points per system from Exhibit 4.3, columns D-Y.
Economic Analysis for the
Final Ground Water Rule
4-14
October 2006
-------
Uncertainty in Baseline Input Data
Although EPA recognizes that there is uncertainty related to the various data sources used to
define the system inventory for the GWR, the uncertainty in the system inventory data inputs is not
quantified in this EA. However, a qualitative discussion of the identified uncertainties follows below.
As noted above, SDWIS and the 1995 CWSS are the sources of system inventory data. SDWIS
is EPA's primary drinking water database, containing data for over 170,000 PWSs. SDWIS stores State-
reported information on each water system, including name, ID number, population served, type of
system, and source of water (ground water or surface water), along with monitoring and violation
information. In 1998, EPA began a major effort to assess the quality of its drinking water data in SDWIS.
The results of this effort, published in the report Data Reliability Analysis of the EPA SDWIS/FED, found
that the data quality of the required inventory data was high (USEPA 2000b). Thus, EPA believes that
uncertainty in the system inventory data from SDWIS with respect to numbers of systems, source
information, and size classification is low.
The 1995 CWSS was developed to gather data on water systems in the United States. A total of
3,681 systems covering a range of source water types and system sizes were selected statistically to
receive the main survey questionnaire. Of these, 1,980 systems responded. These responses were given a
weighting factor to maintain statistical representation of the total universe of CWSs. This weighting
factor was used in all evaluations of data. The EPA report, "Community Water System Survey" (USEPA
1997a) provides information on the 1995 CWSS survey design and data evaluation.
The 1995 CWSS was the primary data source used to estimate percentage of ground water
systems that disinfect, the number of entry points per system, and average and design flow based on
population served. Because the CWSS is a statistical sample, estimates based on the data will contain
uncertainty because of sampling and other errors. The resulting sampling error uncertainty in some of the
CWSS estimates were characterized in the published report by confidence bounds (95 percent) on the
means and proportions provided. While these confidence bounds were not directly used to quantify
uncertainty in the CWSS data elements used in this EA, most of the 95 percent confidence intervals for
CWSS data related to those used in this EA were with within +/- 10 percent of the best estimates
provided.
4.2.4 Water Treatment Plant Design and Average Daily Flows
Treatment technology costs are based on the volume of water treated per day. The cost analysis
described in Chapter 6 uses two types of treatment plant flow:
Design flow—the maximum capacity at which the plant was intended to operate, expressed in
millions of gallons per day (mgd).
Average daily flow—the flow produced by a treatment plant in one day, averaged over 365 days,
expressed in mgd.
Design flows are used to estimate the capital costs of the technology that will be installed to meet the
requirements of the GWR. Average daily flows are used to estimate the annual cost of continuing
operations and maintenance (O&M).
Economic Analysis for the 4-15 October 2006
Final Ground Water Rule
-------
To derive flow information for different sized systems, EPA developed the following regression
equations from the Baseline Handbook (USEPA 200la) relating design and average daily flow (MOD)
for ground water systems to population served (X), using data from the 1995 CWSS:
Design Flow (MOD) = (0.39639*X097708)/1,000
Average Daily Flow (MOD) = (0.06428*X107652)/1,000
The derivation of these equations is presented in detail in the Model Systems Report (USEPA, 2000a) and
summarized in the Baseline Handbook (USEPA, 200la). The equations are used in this EA to estimate
mean flows per entry point for each size category, using the average population served per entry point.
Exhibit 4.6 presents the average population per entry point and corresponding average daily and design
flows.
Exhibit 4.6 Design Flows and Average Daily Flows per Plant (MGD)
System Size
(population served)
Average Population
Served per Entry Point
X
Design Flows (MGD)
Per Entry Point
Y = 0.39639 x097708; 1,000
Average Daily Flow (MGD) Per
Entry Point
Y = 0.06428 x107652; 1,000
CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
44
144
364
758
1,750
3,662
5,758
16,727
115,450
0.016
0.051
0.126
0.258
0.585
1.203
1.872
5.306
35.034
0.004
0.014
0.037
0.081
0.199
0.441
0.718
2.263
18.107
NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
>1 Million
50
226
673
1,552
5,153
20,764
66,000
110,000
-
0.02
0.08
0.23
0.52
1.68
6.55
20.29
33.42
-
0.00
0.02
0.07
0.18
0.64
2.86
9.92
17.19
-
TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
>1 Million
38
177
641
1,416
5,017
15,770
51,850
125,000
0.01
0.06
0.22
0.48
1.64
5.01
16.03
37.86
0.00
0.02
0.07
0.16
0.62
2.12
7.65
19.72
Note: Flow rates are calculated from unrounded population data.
Source: Equations relating mean population to flow are from the Baseline Handbook (USEPA, 2001a).
Population per system (X) data are from Exhibit 4.5, column A.
Economic Analysis for the
Final Ground Water Rule
4-16
October 2006
-------
This EA uses a single regression equation to estimate flows for both publicly and privately owned
systems. There is, however, a slight difference in the flow characteristics for these two ownership types,
as discussed in the Model Systems Report (USEPA, 2000a). The use of different flow equations for
public and private systems would not affect total national costs, although per-household costs may be
slightly affected. EPA has evaluated the equations and believes that the differences are small and would
have a negligible effect on estimated household costs.
Comparable analyses relating average daily and design flow to population were not performed for
the NCWSs. Other drinking water rules have evaluated flows for NCWSs according to service categories
(e.g., schools, restaurants, hotels, and industry) instead of size. EPA considered using this method for
evaluating NTNCWSs for the GWR, but decided against it for the following reasons:
•• Service category flows are based on mean population served for all systems in that
category, regardless of source water type. EPA expects that ground water sources would
be less prevalent in larger NCWSs, but has no basis for developing revised population
estimates for each service category by source.
•• The prediction of technology selection in Chapter 6 is a function of population served
and does not directly apply to service categories that may include a wide range of water
system sizes and flows (e.g., schools can be very small local buildings or large
metropolitan high schools).
EPA, therefore, applied the CWS regression equations to NCWSs, recognizing that this may
over-estimate flows and, therefore, costs. This over-estimation is addressed as part of the uncertainties
summarized in section 4.5. Mean plant flows for CWSs and NCWSs may differ from each other because
of the difference in mean population per plant within each size category.
4.2.5 Treatment Practices Baseline
To properly estimate the cost of compliance with the GWR, the EA takes into account the
percentage of systems presently employing certain types of disinfection. These percentages are used for
determining what additional treatment technologies will be installed to comply with the rule. The GWR
compliance forecast for treatment technology selection (see section 6.3.6.2) was developed based on the
assumption that systems would install technologies in approximately the same proportions as the
technologies that are currently employed. Exhibit 4.7 displays the percentage of systems using various
treatment technologies. Because some systems perform no treatment and some may perform more than
one type of treatment, percentages do not total to 100 percent of systems by system size.
4.2.6 Number of Households Served
Because CWS costs are often passed onto customers in the form of water rate increases, the GWR
also conducts analyses to assess the impact of the rule provisions at a household level. The number of
households served by CWSs expected to be subject to the GWR is estimated by dividing the population
for each system size category by the average number of people per household (2.59, according to the
2000 U.S. Census) (U.S. Bureau of the Census, 200 la). As shown in Exhibit 4.8, CWSs serve almost 39
million households.
Economic Analysis for the 4-17 October 2006
Final Ground Water Rule
-------
Exhibit 4.7 Disinfection Treatment Practices for Disinfecting
Ground Water Systems
Treatment Type
Service Population Category (Population Served)
<100
101-500
501-
1,000
1,001-
3,300
3,301-
10,000
10,001-
50,000
50,001-
100,000
>1 00,000
Pre-Disinfection
Chlorine
Chlorine dioxide
Chloramines
Ozone
Pre-disinfection/
oxidation combinations
64.2%
1 .3%
0.0%
0.0%
0.3%
69.9%
0.0%
0.0%
0.0%
0.5%
56.7%
0.0%
0.0%
0.0%
0.0%
73.2%
0.0%
0.0%
0.0%
0.7%
60.6%
0.0%
0.0%
0.0%
1 .0%
57.4%
0.0%
0.6%
0.0%
2.6%
36.2%
3.1%
1.4%
0.0%
0.0%
38.1%
0.0%
0.7%
0.6%
0.0%
Filtration
Reverse Osmosis
0.0%
0.7%
0.0%
0.6%
0.6%
0.2%
0.3%
0.0%
Post-Disinfection
Chlorine/
Hypochlorination
Chlorine dioxide
Chloramines
Post-disinfection
combinations
23.0%
0.0%
0.0%
0.0%
23.4%
1 .0%
0.0%
0.0%
32.5%
0.0%
0.0%
0.0%
28.3%
0.0%
0.0%
0.0%
42.5%
0.0%
0.1%
0.1%
41 .9%
0.6%
1.1%
0.1%
54.5%
0.0%
3.9%
0.0%
65.8%
0.0%
4.3%
0.0%
Notes: Represents treatment practices for plants treating water that comes entirely or partly from ground sources.
Percentages may not add to 100% because systems may perform more than one treatment.
Sources: Community Water System Survey (CWSS), 1997. Volume II, Table 1-23.
Exhibit 4.8 Ground Water Rule CWS Household Baseline
System Size
(population
served)
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
National Totals
Number of Household Served by Disinfecting Systems
Public
Purchased
2,271
48,489
74,821
222,927
411,772
795,510
235,730
703,495
-
2,495,015
Non-
Purchased
16,767
305,496
631 ,590
2,368,617
4,179,250
8,539,858
2,715,214
5,955,022
1,518,739
26,230,554
Private
Purchased
1,569
21 ,950
21 ,245
55,882
65,005
116,491
32,784
97,193
-
412,121
Non-
Purchased
132,102
639,795
345,931
696,869
787,531
1,615,545
644,828
1,184,833
-
6,047,434
Total
152,709
1,015,730
1,073,588
3,344,295
5,443,557
11,067,404
3,628,556
7,940,544
1,518,739
35,185,123
Number of Households Served by Non-Disinfecting Systems
Public
Purchased
2,030
13,756
14,252
56,781
62,620
28,853
37,422
26,272
-
241 ,984
Non-
Purchased
14,989
86,668
120,303
603,299
635,554
309,736
431 ,036
222,387
-
2,423,972
Private
Purchased
1,403
6,227
4,047
14,234
9,886
4,225
5,204
3,630
-
48,855
Non-
Purchased
118,091
181,508
65,892
177,496
119,763
58,595
102,365
44,247
-
867,957
Total
136,513
288,160
204,493
851,809
827,822
401,408
576,028
296,535
-
3,582,768
Note: Detail may not add to totals due to independent rounding
Sources: System CWS population baseline (Exhibit 4.4) divided by 2.59 people per household (U.S. Bureau of the Census, 2001a).
4.2.7 Triggered Monitoring Baseline
The Ground Water Rule (GWR) requires specified systems to conduct triggered source water
monitoring with every sample that tests positive for total coliform (TC) collected in accordance with the
Total Coliform Rule (TCR) (40 CFR 141.21). An important variable used in the cost and benefits models
is the number of a system's TCR samples that would test positive for total coliform each year. The
Economic Analysis for the
Final Ground Water Rule
4-18
October 2006
-------
average varies by system type (CWS, NTNCWS, and TNCWS) and by system size (eight size categories).
This is an important variable because indicator monitoring is "triggered" by positive total coliform
(TC-positive) samples. Information from the Data Verification (DV) project were used as the primary
inputs for modeling the occurrence of TC-positive samples, while data from other sources were used for
weighting results. The following is a summary of the derivation of this baseline. More detailed
discussion and calculations are presented in Appendix I.
Data Verification (DV)
The DV study involves the comparison of 1 year's worth of PWS records in SDWIS/Fed with
State PWS records to identify any discrepancies between the two records. State files contain the water
systems' reports on the numbers of total coliform samples taken to comply with the TCR and the numbers
of TC-positive samples. Using these data, EPA derived national estimates of the percent of all TCR
samples that test positive for total coliform, and the number of total coliform samples per year that test
positive for all systems in the eight categories.
Description of Method
Several steps were taken to compile the data and weight them appropriately for each system type
and size category.
1) Estimate the fraction of samples that are TC positive by type (CWS, NTNCWS, and
TNCWS) and size of system (those serving at most 1,000 people, and those serving more
than 1,000 people).
2) Estimate the number of routine TC samples taken by each system according to its type
and size category. (Monitoring requirements under the TCR are related to system size.)
3) Multiply the fraction positive by the number of TC samples taken to estimate the number
of TC samples that will test positive, per system and per year, for each of the three
system types and eight system sizes.
Step 1: The raw annual fractions positive were computed for systems in the DV sample by
dividing the total number of routine samples that tested positive for TC by the total number of TC
samples that were taken for each type and size of system. The DV data includes information on systems'
treatment processes, including disinfection steps. This is important, since triggered monitoring under the
GWR applies only to systems that provide less than 4-log2 virus inactivation, removal, or State-approved
combination of these technologies. Thus, only systems listed as nondisinfecting in the DV database were
used in this step. This is a large sample, comprising 1,252 system-years3 of data and 18,467 TC samples
2 X-log virus inactivation or removal means that only one of every 10X viruses survives treatment in such a
condition that it is capable of initiating infection in a human host. 10"x is the fraction that is neither inactivated nor
removed.
3 System-years can be defined as the sum, across systems, of the number of years covered by their DV data.
Almost every DV system produces 1 year of DV data, so system-years are approximately equal to the number of
systems.
Economic Analysis for the 4-19 October 2006
Final Ground Water Rule
-------
from 43 States4 and is broadly representative of nondisinfecting ground water systems in the United
States. Weights based on the number of samples taken were used to extrapolate these raw rates to the
universe of nondisinfecting systems, as follows:
•• The number of nondisinfecting systems was estimated by state, type, and size using two
sources of data: a SDWIS inventory5 for the number of ground water systems and a
1996 EPA study6 for the percentages of systems by state that do not disinfect their
ground water. The percentage of systems that do not disinfect was multiplied by number
of systems and rounded to the nearest integer to estimate the total number of
nondisinfecting systems in each state, size, and system type category.7
•• The DV data provided values for: 1) the number of TC samples taken by each system
according to state and size category and 2) the associated numbers of systems. The
average number of samples taken per system was derived for each state and size category
by dividing its number of TC samples by its number of systems in the DV dataset. The
total national number of samples taken in a size category was then calculated by
multiplying the average number of DV samples per system by the total number of
systems of that size in the US.
•• A weighted TC-positive rate was derived for each size category and type of system by
multiplying its state-specific raw TC-positive rate by the appropriate weight (number of
samples assayed by the state, within the relevant size category and type of system),
summing the weighted TC-positive rates across states, and dividing by the sum of the
weights.
•• To improve consistency across size categories, the TC-positive data were combined into
two size categories: those systems serving at most 1,000 people and those serving more
than 1,000 people. The resulting approximate confidence intervals (using a 90 percent
confidence level, about estimates of 0.5) were +/- 4 percent for systems serving at most
1,000 people. For systems serving more than 1,000 people, the confidence intervals were
+/- 11 percent for CWS systems, +/- 29 percent for NTNCWS, and +/- 27 percent for
TNCWS.
4 There are several reasons why states could not be included in the analysis. Four States, (Alaska, Colorado,
Connecticut, and Illinois), the U.S. Territories and Regional DI programs are not included in this analysis because
data on the percent of systems that do not disinfect are not available, even though DV data are available. Arkansas
had no nondisinfecting systems in the DV sample. Two States, New York and North Dakota, could not be included
because the data verification forms for these States could not be located. Texas requires disinfection by all ground
water systems, but the DV project identified one wholesale system that took samples of source water before
disinfection, and so the one Texas system was included. Data were not included from systems if the systems reported
that no samples were taken, or if there was no information in SDWIS on whether disinfection was used.
5 2004, 4th Quarter
6 USEPA, 1996b
7 In a few cases, the rounding caused the estimated number of systems to drop below the number observed
in the DV inventory, and so in these cases, the DV inventory number was used.
Economic Analysis for the 4-20 October 2006
Final Ground Water Rule
-------
Exhibit 4.9 shows that TC-positive rates varied by size and type of system from 0.71
percent to 6.36 percent. These values are a good representation of the TC-positive rates
for systems in these size and type categories.
Exhibit 4.9 Total Coliform Positive Hit Rates
Type of
System
CWS
NTNCWS
TNCWS
Size of System
(Population Served)
•1,000
>1,000
•1,000
>1,000
•1,000
>1,000
TC-positive Hit
Rate (percentage
of samples)
2.72%
0.71%
2.98%
2.25%
6.36%
3.53%
Source: Appendix I.
Step 2: This step estimates the average number of routine TC samples taken annually by each
system, according to its size and type. The data set for this step was expanded to include disinfecting
systems and includes 2,774 system-years of data and 94,307 TC samples from 44 states8. This expanded
data set improves the estimate of the average number of samples taken. Systems were grouped according
to their baseline monitoring requirements under the TCR, and the number of samples taken was weighted
by the number of systems subject to triggered monitoring. The average number of samples per system
was calculated by dividing the sum of the weighted number of samples by the number of systems subject
to triggered monitoring. Exhibit 4.10 presents the results of this analysis, and Appendix I presents
additional detail for the calculations used.
Step 3: EPA estimated the frequency of total coliform-positives per year per system by
multiplying the number of TC samples per year (from Exhibit 4.10) by the probability of a TC-positive
(from Exhibit 4.9). So, for example, the estimated number of TC-positive samples per system per year
for CWSs serving less than 100 people is 0.0272 * 14 = 0.38; for TNCWs serving greater than 100,000
people it is 0.0353 * 1,496 = 52.8, Exhibit 4.11 shows these estimated frequencies for all system types
and sizes.
Arkansas was included because there were DV data for disinfecting systems.
Economic Analysis for the
Final Ground Water Rule
4-21
October 2006
-------
Exhibit 4.10 Estimated Number of Routine Total Coliform Samples
Taken Per System, Per Year, by Type and Size of System
System Type
cws
NTNCWS + TNCWS
CWS + NTNCWS + TNCWS
Population
Served
<100
101-500
500-1 K
<100
101-500
500-1 K
1011-3300
3301-10K
10,001-50K
50,001-100K
>100,001
TCR Baseline
Number of
Routine Samples
per System
12
12
12
4
4
4
24
84
360
960
2,520
Estimated Actual
Number of
Routine Samples
per System
14
15
18
7
8
9
31
82
311
924
1,496
Source: Appendix I.
Exhibit 4.11 Estimated Number of TC+ Samples Per System, Per Year,
by System Size and System Type
System
Type
CWS
NTNCWS
TNCWS
System Size (Population Served)
<100
0.38
0.22
0.47
101-
500
0.41
0.23
0.48
501 -1K
0.49
0.28
0.60
1,001-
3,300
0.22
0.70
1.1
3,301-
10K
0.58
1.8
2.9
10,001-
50K
2.2
7.0
11.0
50,001-
100K
6.6
20.8
32.6
>100K
10.6
33.7
52.8
Source: Derived from Exhibits 4.9 and 4.10.
The results of these analyses are used in both the benefits and cost models to estimate the number
of indicator samples that systems will have to take to comply with the triggered monitoring provisions of
the GWR. Further discussions of the application of these results are presented in Chapter 5 (section
5.2.5.5) for the benefits model and Chapter 6 (section 6.3.4) for the cost model.
Economic Analysis for the
Final Ground Water Rule
4-22
October 2006
-------
4.3 Water Quality Baseline
This section provides an overview of baseline water quality data that are used in the GWR risk
assessment model. It describes how existing historical, geological, and water quality (e.g., pathogen
occurrence) data are used to characterize the baseline conditions for ground water systems. Next it
describes how this information is used in a risk assessment model to estimate the baseline number of
illnesses and deaths associated with ingesting pathogenic viruses (only) in public ground water systems.
The standard framework is organized in accordance with EPA Policy for Risk Characterization (USEPA
1995a), EPA's Guidance for Risk Characterization (USEPA 1995b), EPA's Policy for Use of
Probabilistic Analysis in Risk Assessment (USEPA 1997b), and with EPA's developing guidance for
microbial risk assessment.
This standard framework requires the use of scientific data (or reasonable assumptions if data are
not available) to produce estimates of the nature, extent, and degree of a risk. Where there is uncertainty
in the data and assumptions used, that uncertainty is described and its impact on the risk estimates is
characterized. Where feasible, variability and uncertainty are mathematically modeled. The EA accounts
for different risk levels within the affected population (variability) and the confidence bounds on key
parameters of the risk assessment model (uncertainty). Variability arises from true heterogeneity across
people, places and time, and uncertainty represents the lack of knowledge of the true value of the factor
being considered (USEPA 1997b).
According to the 1995 EPA Policy for Risk Characterization (USEPA 1995a), health risk
assessments for environmental contaminants generally involve four components:
1. Hazard Identification addresses the nature of the potential adverse health
effects associated with exposure to the contaminant.
2. Exposure Assessment addresses both the number of people in the population
exposed to the contaminant and the distribution of levels of exposure within that
population.
3. Dose-Response Assessment addresses information concerning the relationships,
quantitatively where possible, between the magnitude of exposure to the
contaminant and the extent and severity of the adverse health effects that may
occur.
4. Risk Characterization combines the hazard identification, dose-response, and
exposure assessment information to describe overall risk to the exposed
population, both in terms of the distribution of risk levels in the population and
the total number of cases of adverse effects anticipated.
The exposure assessment step includes evaluation of the probability of pathogen occurrence in
wells and samples, pathogen concentrations in source water, disinfection treatment effectiveness, daily
drinking water consumption, number of days of exposure, and size of the exposed population. Pathogen
occurrence probabilities and concentration are of particular importance to the performance of the risk
assessment. Pathogen occurrence probabilities include (a) the fraction of ground water sources that have
some pathogen occurrence and (b) for sources with pathogens, the fraction of time that pathogens are
Economic Analysis for the 4-23 October 2006
Final Ground Water Rule
-------
present and can be detected. Pathogen concentration is the amount of a given virus in that source when
present. These parameters are used in conjunction to make the connection between an individual's
consumption of drinking water and the possibility of contracting an illness as a result of that consumption.
In terms of the risk assessment model, each contaminated well is assigned a probability that designates the
fraction of time, that its source water has a detectable virus contamination. (At other times the water is
assumed to be virus free.) This occurrence modeling is discussed in greater detail in the following
sections. The use of the baseline data to support the GWR risk assessment benefits analysis, and a
detailed description of the entire risk assessment methodology are presented in Chapter 5.
4.3.1 Background
Within the GWR risk assessment framework, two simplifying categorizations are made to
simplify the analysis. The first categorization is the placement of human viral pathogens into two
representative types for analysis (discussed in section 4.3.1.1). The second is the division of the universe
of wells within the US into "more" and "less" vulnerable categories to best represent the available
enterovirus concentration data (discussed in Section 4.3.3).
4.3.1.1 Representative Pathogens
Viruses
For purposes of conducting the GWR risk assessment, EPA has divided the universe of viruses
into two groups, Type A and Type B. These two types of viruses cause illnesses of different severity and
cause illness and death at different rates. A large number of viruses pathogenic to humans exist with
varying levels of occurrence, infectivity, morbidity, severity, and mortality rates. However, few data are
available on dose response relationship and occurrence in source water for most of these viral pathogens.
EPA has based its risk estimates on two viruses for which dose response data are available. Ground
water occurrence data are also available for some viral pathogens; however, only Type B occurrence data
are available from multiple studies using standardized methods.
The Type A group represents viruses that have high infectivity but generally have mild
symptoms. Examples of Type A viruses include but are not limited to the following: rotavirus, norovirus,
hepatitis A virus, and some other common viruses such as adenovirus and astrovirus that typically cause
outbreaks in schools and daycare centers. Such viruses generally do not result in life-threatening
illnesses. A common illness associated with Type A viruses is gastroenteritis, sometimes accompanied by
vomiting.
The Type B group represents viruses that have low to moderate infectivity but potentially more
severe health effects, which may result in death. Examples of Type B viruses include but are not limited
to the following: echovirus, coxsackievirus, and other enteroviruses. Illnesses associated with Type B
viruses range from gastroenteritis and meningitis to more severe illnesses such as myocarditis or flaccid
paralysis (due to non-polio enteroviruses).
Bacteria
Outbreak data show that bacterial pathogens occur in ground water (see Exhibit 2.3). Ideally,
EPA would estimate baseline illnesses and deaths for both bacterial and viral pathogens to support this
EA since the GWR will mitigate against risk from both concerns. While data (though limited) are
Economic Analysis for the 4-24 October 2006
Final Ground Water Rule
-------
available to characterize viral occurrence, insufficient data are available to characterize pathogenic
bacterial occurrence. Therefore, the primary analysis in this EA focuses on viral changes in viral
exposure attributed to the GWR. A discussion of the unquantified benefits related to reduction in
bacterial exposure is presented in Chapter 5 (section 5.5).
4.3.2 Enterovirus and E. coli Occurrence in PWS Well Source Ground Water
EPA evaluated all available relevant ground water occurrence studies (EPA, 2006b). This section
explains the rationale for selecting the 15 studies for use in the final GWR economic analyses. The
occurrence data are used to determine the probability that a well or sample will be positive for enterovirus
and/or E. coli. E.coli was selected because EPA expects that most States will select E.coli as their
monitoring target under the final rule. To assist with the analysis, EPA consulted with a group of
statisticians to discuss ways to make optimal use of these limited data. The statisticians strongly
recommended that EPA make use of all the available data unless there were known quality assurance
problems with a data set or the well contamination scenario was outside the normal operating range of US
PWS wells. Thus, EPA used all the available data on enterovirus occurrence in ground water from PWS
wells in the United States except for one data set of alluvial wells from Missouri that were substantially
affected by severe Mississippi River flooding (Vaughn, 1996). Data from the 15 studies selected as
described in the following section were combined into one complete data set.
Study Selection
EPA has reviewed data from 24 recent studies of pathogen and fecal indicator occurrence in
ground waters that supply PWSs (EPA, 2006b). Each study was conducted independently and with a
unique objective and scope. The available data indicate a wide range of enterovirus occurrence in water
drawn from wells across the United States. EPA selected 15 studies with results that are directly
applicable to evaluating GWR benefits. These studies include the largest data sets characterizing
enteroviral occurrence in the United States. One data set, Lieberman et al., 2002, targeted wells based on
presence of total coliforms and other indicators of vulnerability to fecal contamination. Another data set,
Abbaszadegan et al. 2003, targeted a representation of wells throughout the United States based on
hydrogeological conditions, but excluded any wells that were poorly constructed or without well logs.
Other studies sampled subsets of wells in particular states or in certain hydrogeologic settings within
states. Because most studies were designed to capture subsets of the total PWS well population, each
study description in the following is accompanied by a short discussion about the representativeness of
the subset as compared with the total population. Aside from recognizing the numbers of wells surveyed,
this analysis makes no attempt to weight any of the studies to compensate for any perceived over- or
under-representation of the subset as compared with the total population.
General Considerations for Interpreting Viral Occurrence Data (Including Uncertainty)
When evaluating enteroviral occurrence data, it is important to realize some of the fundamental
challenges in characterizing enterovirus occurrence, including detection, identification and concentration.
Key issues include recognizing the limitations of enterovirus measurement and the limitations of deriving
the probability of a well or sample being enterovirus-positive and the associated enterovirus concentration
estimates from the measured data.
Economic Analysis for the 4-25 October 2006
Final Ground Water Rule
-------
EPA relies only on the identification of enterovirus in PWS wells using cell culture methods since
they allow identification of infectious pathogenic viruses. However, these measurements are
underestimates of the actual occurrence because a) only a few of the viruses that can occur are detected by
the method9, b) each viral plaque that is counted as one virus originates from an infection by one or more
viruses, and c) virus recovery is variable depending on the water chemistry and the viral strains present.
Virus recovery can range from less than 20% to greater than 50% (Dahling, 2002; Denis-Mize et al.,
2004; Sobsey and Glass, 1984).
Viruses often aggregate in water or solution. Methods that count host cell infection cannot
differentiate between virus aggregates and solitary viruses and counts them all as solitary viruses (Teunis
et al., 2005; Young and Sharp, 1977). Where more than one virus strain co-occur in placque assays,
statistical analysis has shown that the actual concentration can be as much as 45% greater than the
concentration determined by the standard method, the plaque assay count (Teunis et al., 2005). Thus, the
probability that a well or sample will be positive for viruses pathogenic to humans will be an
underestimate of the true probability and varies depending on the water chemistry, the type of cell culture
used and the type of virus.
The standard host cell line used to recover human viruses is the Buffalo Green Monkey (BGM)
continuous cell line. This cell line is less sensitive than primary cell lines derived from freshly harvested
kidney cells (Ward et al., 1984). In addition, the age of the BGM cell line affects its sensitivity. BGM
cell lines should not be used if they are passaged more than 250 times or if it has been longer than one
week since the last passage.
Furthermore, concentration estimates derived from measured values of infectious viruses will be
underestimates of the actual concentration because some viruses (such as reovirus) may be favored in the
cell line used for testing and may out-compete other viruses (such as echovirus) (Carducci et al, 2002).
Among the enteroviruses, slower growing enteroviruses are not favored for recovery and identification.
For example, in BGM cells coxsackie B virus is a fast growing virus whereas Echo 11 grows slowly
(Lieberman et al., 2002). Thus, the probability that a sample will be positive for viruses pathogenic to
humans, given that such viruses are present in the sample, depends on water chemistry, the type of cell
culture used, and the type of virus.
Viral occurrence can be characterized by considering the probability that a well or sample is
virus-positive and the viral concentrations associated with those positive samples. However, the number
of samples taken at a site and the sensitivity of measurement can significantly influence the estimated
probabilities and concentrations per site for the anticipated exposure duration. Available data indicate that
viral concentration at one site taken at different points in time can be highly variable, ranging from below
detect to several orders of magnitude above detect (Lieberman et al., 2002). This is because some wells
9 While each cell culture method can detect pathogens such as poliovirus, some coxsackievirus, echovirus,
and reovirus (Type 3), many coxsackie A and other viruses are not detected. For example, no cell culture method
exists to recover noroviruses in stool or environmental samples, yet noroviruses are responsible for the greatest
proportion of water and food-borne disease outbreaks and therefore are most likely to be present in fecal
contamination. Similarly, the BGM cell culture method to detect enteroviruses is inefficient for detecting rotaviruses
in well water. There are few infectious rotavirus occurrence data available for making direct rotavirus health effect
predictions. Rotaviruses are ubiquitous in nature, exemplified by the fact that all adults in the United States are
seropositive for rotavirus, indicating previous infection. These detection method deficiencies minimize infectious
virus recovery for two of the most important viruses and therefore underestimate the health effects predictions.
Economic Analysis for the 4-26 October 2006
Final Ground Water Rule
-------
may never have a viral occurrence, others may have a short duration contamination, and still others may
have high levels of occurrence and concentration for sustained durations. The possibility of high variation
in viral occurrence at any given site makes estimates of probabilities and concentrations difficult.
In addition to the general issues regarding viral data characterization discussed above, specific
issues of data representativeness, bias, uncertainty, and variability exist for any individual study. The
interpretations of the data from the 15 primary studies used to support GWR analyses are presented in
detail below.
4.3.2.1 Lieberman et al. 2002 Study
Study Objectives
The major objectives of the Lieberman et al. 2002 study were: 1) to obtain occurrence data for
infectious human enteric viruses using the BGM cell line, 2) to assess the microbial indicators of fecal
contamination, and 3) to develop and to evaluate a molecular biology monitoring method (PCR) to
identify viral genomic material without consideration of the infectiousness of that material. The
objectives were accomplished by sampling wells to confirm total coliform presence and to establish the
presence of other fecal indicators, including somatic coliphage (Phase I) and by choosing a subset of these
for monthly sampling for 1 year (Phase II). Wells were nominated for sampling in Phase I by federal,
State and local drinking-water experts.
Well Selection
In Phase I, 180 wells were nominated, and 98 were selected. Each selected well was sampled
once for total coliform, E. coli, enterococci, Clostridium perfringens spores, and somatic coliphage.
Nominated wells were identified using historical total coliform occurrence data and any other available
information about the well. In choosing which wells to nominate, other information was considered such
as confirmed waterborne disease outbreaks, proximity to known sources of human fecal contamination
and, in some cases, siting in a sensitive hydrogeologic setting (e.g., karst). Selected wells were located in
22 States, Puerto Rico, and the U.S.Virgin Islands. The wells from Phase I served as the well selection
pool for 21 of the 30 wells chosen for Phase II sampling.
Twenty-seven of the thirty wells selected in Phase II had either a history of total or fecal coliform
occurrence or had any indicator occurrence during Phase I sampling. In aggregate, the 30 wells selected
for monthly sampling represent a group of wells considered to be vulnerable to fecal contamination
primarily due to historical indicator occurrence, but also due to positive results for somatic coliphage,
enterococci, or other indicators from a single sample during Phase I sampling. Proximity to fecal
contamination sources, high nitrate concentrations, and location in a sensitive hydrogeologic setting were
additional selection criteria for several additional wells. The 30 selected wells, located in 17 States and 2
U.S. territories, were sampled monthly for 1 year for total coliform, E. coli, enterococci, Legionella
species, Clostridium perfringens spores, somatic and male-specific coliphage, Bacteroides bacteriophage
and enteric viruses using BGM cell line.
Economic Analysis for the 4-27 October 2006
Final Ground Water Rule
-------
Sample Results
For viral analyses using cell culture assays, seven of the 30 wells (23 percent) were positive for
enterovirus and 20 samples (6 percent) were positive for enterovirus or reovirus. While 7 of the 30 wells
sampled had a cell culture positive among the twelve samples taken, most of the measurement were below
detects. One of the wells had 5 monthly viral positives, two of the wells had 4 monthly positives, one of
the wells had two monthly positives, and three wells had one positive. Viral strains identified by
serotyping included coxsackievirus and echovirus, as well as the enteric virus reovirus. Virus-positive
samples ranged in concentration from 0.9-212 PFU or MPN/100 liters with a mean infectious virus
concentration of 30.66 PFU or MPN/100 liters (PFU, or plaque forming units, and MPN, or most
probable number, are estimates of concentration) among all the positive samples.
Data Representativeness
Most of the wells selected as part of Phase II of the Lieberman et al. 2002 study had a history of
total coliform occurrence that was confirmed by Phase I sampling. Because most (but not all) of the wells
selected for inclusion in the study had a history of fecal contamination, these data are not representative
of all PWS wells in the United States because not all wells in the United States have a history of fecal
contamination.
The GWR is concerned primarily with ground water sources vulnerable to contamination,
especially the undisinfected sources. Most of the Lieberman et al. 2002 study wells, however, already
employ disinfection, which potentially introduces a bias to the data (i.e., the use of disinfection could be
considered an indication that the source is known to be contaminated). However, the use of disinfection
does not necessarily correlate with known contamination. One enterovirus-contaminated well in the
Lieberman et al. 2002 study was undisinfected and had the highest virus concentration for any single
monthly sample of the entire study. Another factor that mitigates against this potential bias is that many
States and some water systems require ground water disinfection as a matter of policy. For the Lieberman
et al 2002 study, 10 of the 30 wells are located in Alabama, Florida, or Texas; States that require
disinfection of all ground water sources. The existence of disinfection at a ground water system may not
be directly correlated with indicator occurrence at that facility and therefore any selection bias is
unknown.
Economic Analysis for the 4-28 October 2006
Final Ground Water Rule
-------
4.3.2.2 Abbaszadegan et al. 2003
Study Objectives
Among the objectives of the Abbaszadegan et al. 2003 study were: 1) to determine the
occurrence of virus contamination in source water of public ground water systems, 2) to investigate water
quality parameters and occurrence of microbial indicators in ground water and possible correlation with
human viruses, 3) to develop a statistically based screening method to identify wells at risk of fecal
contamination, and 4) to develop and evaluate a molecular biology monitoring method (PCR).
Well Selection
Wells were selected for the Abbaszadegan et al. 2003 study from a pool of 750 wells. The study
was initiated as a study of 150 samples from AWWSCo wells selected to test and evaluate the PCR
method (Abbaszadegan et al., 1999). With additional funding, the study was expanded to 539 samples
from 448 wells. The additional wells were nominated by State drinking water program or water utility
staff. Study personnel requested nominations of wells not known to be vulnerable to microbial
contamination. The researchers excluded 12 samples included in the first 150 AWWSCo well samples
because they were believed to be under the direct influence of surface water and therefore especially
vulnerable to contamination. Other nominated wells were excluded if well records were not available or if
the well was improperly constructed. All nominated wells were profiled by the well operators or their
designee using a questionaire that included a checklist of 11 different hydrogeologic settings. Researchers
selected wells for inclusion in the study based on the reported hydrogeologic setting information. Wells
were selected for inclusion if they were apparently located in a setting that was proportionately
under-represented as compared with a USGS national hydrogeologic profile derived using these same 11
hydrogeologic settings.
Sample Results
Source water samples were taken from each well and analyzed using a variety of methods to
detect pathogens and indicators. Samples were analyzed to determine the occurrence of viruses (using
both cell culture and polymerase chain reaction (PCR) methods) and total coliform (TC), enterococci, and
C. perfringens bacteria in ground waters of the United States. A total of 539 samples were obtained. Not
all analyses were conducted on all samples, and 25 wells were sampled two or more times. Information
was not available to identify which wells were sampled multiple times. Because the majority were
sampled once, and having no other recourse, EPA's data were reduced to single samples as though each
of these was the only one assayed for a well. PWSs performed the sampling and were given training on
procedures to collect at least 400 gallons (1,512 L) of water prior to disinfection. Exhibit 4.12 presents a
summary of the Abbaszadegan et al. 2003 study results.
Economic Analysis for the 4-29 October 2006
Final Ground Water Rule
-------
Exhibit 4.12 Results of the Abbaszadegan et al. 2003 Study
Assay
Enterovirus (cell culture)
Total coliform
Enterococci
Clostridium perfringens spores
Male specific coliphage (Salmonella WG-49 host)
Somatic Coliphage (E. coli C host)
Somatic and Male Specific Coliphage (E. coli C-3000 host)
Percent of Sites with
Positive Samples
(No. positive/samples analyzed)
4.8%
9.9%
8.7%
1 .8%
9.5%
4.1%
10.8%
Source: Abbaszadegan, 2002; Abbaszadegan et al., 1999, 2003
Data Representativeness
The Abbaszadegan et al. 2003 study included a large number of wells that were specifically
chosen to be representative of the range and proportion of the hydrogeological settings of the United
States. To further evaluate the representativeness of the wells with respect to hydrogeologic conditions,
EPA subsequently compared nitrate concentrations from a national database of nitrate concentrations in
ground water (Lanfear, 1992) with nitrate data measured in the Abbaszadegan et al. 2003 study wells to
determine if there was any statistically significant difference between the nitrate levels in the two data
sets. Nitrate was chosen for this comparison because a large, national database is available. The national
nitrate data were selected randomly from a database of more than 100,000 wells. Using U.S. Census data,
EPA stratified the nitrate data into rural and urban components and chose a small random subset of these,
comparable in size to the sample in the Abbaszadegan et al. 2003 study data (all available Abbaszadegan
et al. 2003 study data were used), for comparison. The analysis showed that the Abbaszadegan et al.
2003 study wells had nitrate concentrations that were not significantly different from the national data or
from the urban and rural components. Thus, using nitrate concentration as a surrogate, EPA further
verified that, by this measure, the Abbaszadegan et al. 2003 study wells data appear to be nationally
representative of hydrogeological conditions in the United States.
In the well selection process, the Abbaszadegan et al. 2003 study initially relied on wells that
were owned by AWWSCo and subsequently used wells that were volunteered for the study. Choosing
from among a restricted pool or using a volunteer process introduces a potential bias to the study. The
AWWSCo wells typically serve larger populations, have greater revenues and are more professionally
managed than most wells, and the volunteered wells were selected precisely because they appeared to be
at low risk. Thus, there may be a downward bias in the contamination levels found during the study. As
applied to the GWR risk analysis, this would result in an underestimate of benefits derived from the rule.
In addition, most wells were only sampled once, which also may underestimate the risks associated with
these wells.
Economic Analysis for the
Final Ground Water Rule
4-30
October 2006
-------
A potential bias of the Abbaszadegan et al. 2003 study, as with the Lieberman et al. 2002 study,
is that the majority of the study wells already employ disinfection (see discussion above for implications
of this bias). A mitigating factor is that many States and some water system companies require ground
water disinfection as a matter of policy. Fifty-two wells from the Abbaszadegan et al. 2003 study are
located in Alabama, Florida, or Texas; States that require disinfection of all ground water sources. In
addition, a large number of wells in the study are operated by AWWSCo, which also disinfects as a
matter of policy. Therefore, the existence of disinfection at a ground water system may not be directly
correlated to issues of contamination at that facility.
Because the description of the hydrogeologic setting was selected by the well operator or
designee from a checklist, there are potential uncertainties associated with the hydrogeologic setting data.
It is possible that the operator had insufficient data to determine the hydrogeologic setting and was unable
to easily consult with a hydrogeologist. No analysis was conducted to determine whether the reported
hydrogeologic setting data were correct. It would be expected that viruses would more likely be found in
sensitive hydrogeologic settings, as was the case with the Lieberman et al. 2002 data, because the ground
water flow within those aquifers is faster and more direct, and there are fewer opportunities for virus
concentrations to become attenuated due to interaction with the aquifer solid materials.
The Lieberman et al. 2002 study found higher virus concentrations and a greater range of
concentrations than those measured in the Abbaszadegan et al. 2003 study10. The Abbaszadegan et al.
2003 concentrations were uniformly low. Because of variations in well water matrix, source density,
proximity and the concentrations in each source, virus filtration and recovery and virus analyses, it is
impossible to assess the significance of the differing virus concentrations.
Overall, the magnitude and direction of the biases and uncertainties inherent to the Abbaszadegan
et al. 2003 study cannot be definitively quantified.
4.3.2.3 Pennsylvania Noncommunity Wells (Lindsey et al., 2002)
Study Objectives
The purpose of this study was to measure pathogen and indicator occurrence in a random
stratified sample of non-community water systems (NCWS) wells in primarily carbonate aquifers and
crystalline aquifers, which are hydrogeologically sensitive settings. The United States Geological Survey
(USGS) (Lindsey et al., 2002) analyzed samples from 60 NCWS wells from September to January 2001
to assess the occurrence and distribution of pathogens in ground water used for non-community water
supplies and indicator organisms (evaluated as surrogates for those pathogens).
10 In using the cell culture method for enterovirus detection, the Abbaszadegan et al. 2003 study identified
only poliovirus from wells (Abbaszadegan et al., 1999) as compared with the Lieberman et al. 2002 study which
identified no poliovirus.
Economic Analysis for the 4-31 October 2006
Final Ground Water Rule
-------
Well Selection
USGS personnel, in collaboration with the Pennsylvania Department of Environmental Protection
(PaDEP), selected random wells from a targeted population of primarily carbonate and crystalline
aquifers. Ten wells were chosen in areas underlain by either siliciclastic bedrock or unconsolidated
surficial aquifers. An unconsolidated aquifer is non-sensitive but the siliclastic aquifer can be either
sensitive or non-sensitive depending on whether it is considered to be a sandstone or a quartzite. Aquifer
sensitivity is best determined by the State, and EPA cannot make that determination based on the
available data.
The vast majority of the sites were TNC PWS wells. Only two wells were NTNC PWS wells.
Surrounding land use was included as a criterion for selection; a site was more likely to be selected if
potential fecal point sources were located nearby. However, water suppliers with known bacterial
contamination problems declined to participate while suppliers with no contamination history were much
more willing to participate.
Sample Results
Of 60 wells initially selected, 59 samples were analyzed for culturable viruses, Helicobacter
pylori (H. pylori), total coliform, Escherichia coll (E. coli), Clostridium perfringens (C. perfringens),
somatic coliphage, male-specific coliphage, and enterococcus.
Culturable viruses were detected in 5 wells, H. pylori in 4 wells, E. coli in 7 wells, total coliform
in 27 wells, C. perfringens in 9 wells, somatic coliphage in 5 wells, male-specific coliphage in 2 wells,
and enterococci in 8 wells.
Of the 5 wells with detectable culturable viruses, two were near 0.21 PFU per 100 L, while the
remaining three ranged from 18 to 56 PFU per 100 L.
Data Representativeness
This data set represents the only randomly sampled human pathogenic virus data from TNC wells
among the 24 studies considered. As such, it is an important data set for representing the large number of
untreated TNC wells in the United States.
4.3.2.4 Southeast Michigan (Francy et al., 2004)
Study Objectives
The purpose of this study of small (serving fewer than 3,000 people) public ground water supply
wells was to assess the presence of both viral contamination and microbiological indicators of fecal
contamination, relate the co-existence of indicators and enteric viruses, and consider the factors that affect
the presence of enteric viruses.
Economic Analysis for the 4-32 October 2006
Final Ground Water Rule
-------
Well Selection
Initially, 160 wells from a previously studied USGS National Water-Quality Assessment Program
site were proposed based on nominations from local State and county experts. Wells were nominated if
they produced from shallow sand and gravel aquifers, were undisinfected and did not have well
construction flaws. The 38 selected wells were randomly selected from the 160 nominated wells. Well
screens are typically shallow ranging from 50 to 150 feet below ground surface. In some places the
aquifer is unconfmed but more often the aquifer is semiconfmed or confined by glacial till. Where
semiconfined or confined aquifer conditions exist, these wells are protected from surficial fecal
contamination sources. From July 1999 through July 2001, researchers collected a total of 169 regular
samples and 32 replicate pairs in southeastern Michigan from 38 wells in discontinuous sand and gravel
aquifers. Not all 38 wells were sampled for all parameters. Only 34 wells (93 samples) were analyzed for
enteric virus by cell culture.
Sample Results
Two wells (two samples) were positive for enteric virus by cell culture. Four wells (four samples
were positive for E. coll. Six wells (7 samples) were positive for enterococci. Two wells (two samples)
were positive for male-specific coliphage and one well (one sample) was positive for somatic coliphage,
based on 1 liter samples. All wells sampled are undisinfected so the semi-confining or confining layers
are not sufficient protection against fecal contamination.
Data Representativeness
This study is unique among the 24 studies considered in that it sampled only undisinfected wells.
Other studies were typically not able to sample undisinfected wells because well operators did not allow
sampling. Thus, this study is representative of the large number of small, undisinfected PWS wells in the
United States. Despite the apparent random well selection process, seven wells were not further
considered for sampling at the request of the well owner or because they were found be unsuitable.
4.3.2.5 New Jersey (Atherholt et al., 2003)
Study Objectives
This study was designed to sample wells in New Jersey for fecal indicator organisms. No samples
were analysed for enteroviruses or other viruses pathogenic to humans. Thus, data from this study was
used only to determine the probability that a sample was fecally contaminated by E. coli.
Well Selection
Twenty-six public water supply wells were sampled for a variety of fecal indicator organisms.
Twelve wells were identified as GWUDI and so data from these wells are not used in this analysis.
Eighty-one samples were collected from the 13 ground water wells (128 from all wells) between June
1999 and February 2002. One well with one sample was not reported as ground water or GWUDI so this
value was not included. All of the wells were located in unconfmed aquifers. Although GWUDI wells
were selected to increase the likelihood that fecal indicator organisms were present, no information is
given for the selection of the other wells.
Economic Analysis for the 4-33 October 2006
Final Ground Water Rule
-------
Sample Results
All 13 wells (81 samples) were negative forE1. coli.
Data Representativeness
These data represent a subset of community ground water wells in New Jersey that produces
water from unconfmed aquifers.
4.3.2.6 Missouri Ozark Plateau #1 (Davis and Witt, 2000)
Study Objectives
The purpose of this study was to determine the water quality in recently constructed community
public water system wells in the Ozark Plateau region of Missouri. This largely rural region is
characterized by carbonate aquifers, both confined and unconfmed, with numerous karst features
throughout. A confining layer is defined in this study as a layer of material that is not very permeable to
ground water flow and that overlays an aquifer and acts to prevent water movement into the aquifer.
Well Selection
The US Geological Survey, working with the Missouri Department of Natural Resources,
selected a total of 109 wells, in both unconfmed and confined aquifers (Davis and Witt, 2000). In order to
eliminate poorly constructed wells from the study, wells that had been constructed within the last 15 years
were selected primarily. Wells were also selected to obtain good coverage of the aquifer and to reflect the
variability in land use. All wells were sampled twice, once in summer and once in winter.
Sample Results
One sample was reported as enteric-virus positive but this virus-positive well was not used in the
data analysis for Today's Rule because this sample (and others) had some quality assurance problems due
to cross contamination of samples with the poliovirus control. All wells from this study are counted as
negative for enteroviruses when evaluated for the probability of well (and sample) being positive for
entero virus.
Data Representativeness
These data are representative of wells in the Ozark Plateau aquifer of Missouri. These data
potentially underestimate the probability of wells and samples being positive for enteroviruses because
one positive well was not included in the data set for Today's Rule.
Economic Analysis for the 4-34 October 2006
Final Ground Water Rule
-------
4.3.2.7 Missouri Ozark Plateau #2 (Femmer, 2000)
Study Objectives
The purpose of this study is to determine the water quality in older (pre-1970) CWS wells in the
Ozark Plateau region of Missouri to supplement the Missouri Ozark Aquifer Study #1, by Davis and Witt
(1998, 1999). This largely rural region is characterized by carbonate aquifers, both confined and
unconfmed, with numerous karst features throughout.
Well Selection
The US Geological Survey, working with the Missouri Department of Natural Resources,
sampled a total of 106 wells (Femmer, 1999), in both unconfmed and confined aquifers. Wells (all of
which were constructed before 1970) were selected for monitoring to obtain good coverage of the aquifer,
and to reflect the variability in land use. Priority was given to wells that had completion records, well
operation and maintenance history, and wells currently being used. Each well was sampled once (during
the spring).
Sample Results
No wells were enterovirus-positive by cell culture.
Data Representativeness
These data are representative of PWS wells in the Ozark Plateau aquifer of Missouri.
4.3.2.8 Wisconsin Migrant Worker Camp (USEPA, 1998b)
Study Objectives
The purpose of this study was to determine the quality of drinking water in the 21 public ground
water systems serving migrant worker camps in Wisconsin (USEPA, 1998b). Each well was sampled
monthly for six months, from May through November, 1997. The study conducted sampling for
male-specific coliphage, total coliforms and E. coll. When detection of coliforms occurred, the specific
type of coliform was further identified (speciated). One total coliform positive sample was identified to
contain Klebsiellapneumoniae, which can be due to fecal or non-fecal origins. Along with the microbial
indicators, nitrate and pesticides were also measured.
Other factors were compared to the microbial and chemical sampling results of the study. Well
construction records were available for 14 of the wells. The mean casing depth was 109 feet (range 40 to
282 feet) and the mean total well depth was 155 feet (range 44 to 414 feet). Most of these 14 wells are
also reported to terminate in a sand or sandstone formation.
Well Selection
These transient, non-community water systems are located in three geographic locations across
the State.
Economic Analysis for the 4-35 October 2006
Final Ground Water Rule
-------
Sample Results
Investigators detected male-specific coliphage in 20 of 21 wells during the six-month sampling
period, but never detected E. coll. In addition, four wells had nitrate levels that exceeded the EPA MCL
for nitrate. No wells were analyzed for enteric virus by cell culture.
Data Representativeness
The data from this study are intended to be representative only of TNC wells in migrant labor
camps.
4.3.2.9 New England (Doherty et al., 1998)
Study Objectives
The purpose of this study was: (1) to determine the prevalence of enteric pathogens in New
England's public water supply wells, (2) to assess the vulnerability of different systems, and (3) to
evaluate various fecal indicators.
Well Selection
Wells were selected based on the following criteria: (1) must have constant withdrawal
throughout the year, (2) must be near septic systems, (3) should have, if possible, a history of violations
of the MCL for total coliforms or elevated nitrate levels, and (4) must not have direct infiltration by
surface water (Doherty, 1998).
Wells were nominated, characterized, selected, and sampled by regulatory staff of Connecticut,
Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont. The selection process considered
wells in different hydrogeologic settings. Of the 124 total wells, 69 (56%) were located in unconfined
aquifers, 31 (25%) were located in bedrock aquifers, 10 (8%) were located in confined aquifer
hydrogeologic settings, and 14(11%) were located in unknown aquifer settings. Each well was sampled
quarterly for one year. Enterococci were identified in 20 of 124 wells (16%) and in 6 of 31 (19%) bedrock
aquifer we 11s.
Sample Results
No wells were positive for enteric virus by cell culture. No wells were positive for E. coll.
Data Representativeness
These wells are intended to be representative of New England PWS wells. Two wells were
provisionally identified as cell culture positive (and reported as positive in EPA, 2000) but were found to
be laboratory contamination when the samples and lab controls were sequenced by CDC.
Economic Analysis for the 4-36 October 2006
Final Ground Water Rule
-------
4.3.2.10 Three-State Study: (Wisconsin-Battigelli, 1999)
Study Objectives
The Wisconsin study sampled 25 wells quarterly for 2 years
Well Selection
No explanation is available on the method used in selecting the wells.
Sample Results
One well in Wisconsin was positive for enteric viruses by cell culture.
Data Representativeness
No information is available to evaluate the representativeness of these data.
4.3.2.11 Three-State Study: (Maryland-Banks et al, 2001)
Study Objectives
The purpose of this study was to sample shallow wells in Worcester and Wicomico Counties on
Maryland's Eastern shore for enteric viruses. Twenty-seven wells were each sampled once. Three other
samples (two from two deep confined wells) were sampled for use as negative control samples. Each well
was sampled for enteric viruses by Buffalo Green Monkey (BGM) and RD (human embryonal
rhabdomyosarcoma) cell culture, Bacteroidesfragilis bacteriophage, somatic and male-specific coliphage,
E. coli, Clostridium perfringens, and TC. Serological testing on virus-positive RD cell cultures confirmed
the presence of rotavirus.
Well Selection
The 27 wells were located in two counties that are underlain entirely by sandy coastal plain
aquifers. The selected wells were chosen from 278 small PWS wells by a vulnerability score that
included factors such as historical fecal coliform occurrence, land use, well depth and age, and other
factors.
Sample Results
One well was positive for enteric viruses by cell culture. This sample was identified as rotavirus
by a non-standard test using an assay method normally used to detect rotavirus in stool. Because the assay
method works primarily for rotavirus at high concentrations, this well was considered to be enteric-virus
positive but not necessarily rotavrirus-positive. None of the wells was positive for E. coli.
Data Representativeness
These data are representative of deep wells in non-sensitive aquifers that are unlikely to be fecally
contaminated.
Economic Analysis for the 4-37 October 2006
Final Ground Water Rule
-------
4.3.2.12 Three-State Study: (Maryland-Banks and Battigelli, 2002)
Study Objectives
The purpose of this study was to sample shallow wells in the Maryland Piedmont physiographic
province. Each well was sampled for enteric viruses by BGM cell culture, Bacteroides fragilis
bacteriophage, somatic and male-specific coliphage, E. coll, Clostridium perfringens and TC. One-
hundred-one samples were collected from April 10, 2000 to November 13, 2000. This total included ten
replicate samples for QA purposes.
Well Selection
For this study, 91 small PWS wells were selected for sampling from 263 wells in the fractured
bedrock aquifer of two Maryland Piedmont Physiographic Province counties. Wells were selected to
distribute the sample sites evenly over the population and spatial extent of the study area. One well was
selected randomly.
Sample Results
None of the wells was positive for enteroviruses by cell culture. One well was positive for E. coll.
Data Representativeness
These data are representative of shallow wells in fractured bedrock (sensitive) aquifers with thick
soil and weathered rock (saprolite) cover that acts to protect against fecal contamination. This same
aquifer setting would be more likely to be contaminated if located further north where glacial advances
during the Ice Age removed much of the weathered soil and rock.
4.3.2.13 Three-State Study: (Minnesota-Banks and Battigelli, 2002)
Study Objectives
The purpose of the three-state study is to characterize the extent of viral contamination in PWS
wells by testing wells in differing hydrogeologic regions and considering contamination over time
(Battigelli, 1999). The Minnesota study (Minnesota Department of Health 2000) sampled 76 wells.
Seventy-four wells were sampled for at least four consecutive calendar quarters. The remaining two wells
were sampled for two consecutive quarters each. In addition to microbial indicator data, one sample from
each well was also analyzed for tritium and tritium/3helium.
Well Selection
Sampled wells were more likely to be selected if they were small, transient PWSs, and/or were
located in a aquifer that was perceived to be vulnerable. Of the 76 Minnesota wells sampled, six (8
percent) served community systems, 19 (25 percent) served non-community non-transient systems and 51
(67 percent) served transient systems. The aquifer types that are utilized by these wells include dolomite
(six wells), dolomite and sandstone (three wells), fractured crystalline bedrock (nine wells), sandstone (28
wells), sand and gravel (29 wells) and regolith (surficial materials) (one well).
Economic Analysis for the 4-38 October 2006
Final Ground Water Rule
-------
Sample Results
No wells in Minnesota were enteric virus-positive by cell culture.
Data Representativeness
These wells were selected to be representative of wells in Minnesota.
4.3.2.14 EPA Vulnerability Study (USEPA, 1998c)
Study Objectives
The purpose of this study was to conduct a pilot test of a new vulnerability assessment method by
determining whether it could predict microbial monitoring results (USEPA 1998c). The vulnerability
assessment assigned low or high vulnerability to wells according to their hydrogeologic settings, well
construction and age, and distances from contaminant sources.
Samples were taken and tested for enteroviruses (both by cell culture and PCR), hepatitis A virus
(HAV) (by PCR), rotavirus (by PCR), Norwalk virus (by PCR), and several indicators (total coliforms,
enterococci, male-specific coliphage, and somatic coliphage). The only positive result was one PCR
sample positive for HAV.
Well Selection
A total of 30 wells in eight States were selected to represent ten hydrogeologic settings. Selection
was based on the following criteria: (1) wells representing a variety of conditions relevant to the
vulnerability predictions, (2) wells with nearby sources of potential fecal contamination, and (3) wells
with sufficient well and hydrogeologic information available.
Sample Results
No wells were positive for enteric virus by cell culture. No E.coli data were collected.
Data Representativeness
Wells were selected to be representative of a variety of hydrogeologic settings in the United
States. However, the small number of wells in the study and the large number of hydrogeologic settings
makes such a comparison difficult.
4.3.2.15 Montana Study
Study Objectives
Miller and Meek (2006) sampled source water for E. coll, enterococci, male-specific and somatic
coliphage from wells representing primary aquifer types, bedrock and valley-fill aquifers.
Economic Analysis for the 4-39 October 2006
Final Ground Water Rule
-------
Well Selection
18 small PWS wells (and 20 residential wells) near Helena, Montana were sampled.
Sample Results
No E. coll, male-specific or somatic coliphage were detected in any well. Enterococci were
detected in two wells but it is not known if these wells were PWS or residential wells
Data Representativeness
Wells were selected for sampling to represent both primary aquifer types, bedrock and valley-fill
aquifers. Each well was sampled from one to three times.
4.3.2.16 Summary of New Data
Exhibit 4.13a identifies the new data available since the proposed GWR was published in 2000.
Although all new data were evaluated, not all new data were used in the economic analysis for the GWR.
Of the seven new studies described in Exhibit 4.13a, four studies were included in the occurrence
data compilation described in Section 4.3.2 and used to determine exposure in this EA. These four new
studies are the Pennsylvania study (Lindsey et al, 2002), the Michigan study (Francy et al, 2004), the
New Jersey study (Atheroholt et al, 2003), and the Montana study (Miller and Meek 2006). Other studies
not included were not used for the reasons described in the following.
Karim et al (2003, 2004) - This study selected 20 wells (15 because of enterovirus or indicator
occurrence) from the Abbaszadegan et al. 2003 study for additional (monthly) sampling. However, with
the available data provided by the researchers (raw spreadsheet data and summary reports), it is
impossible to combine the two data sets because the well site identifying characters differ in the two
studies. Thus, there is no alternative other than treating the two studies as if they are separate,
independent data sets. If treated as two data sets, significant bias is introduced. First, the same well is
counted twice. Second, well data are treated as if they are unbiased, independent data when they actually
were selected with based on prior sampling which identified either infectious enterovirus, enteric virus
RNA, one or more fecal indicator bacteria or no contamination, (five wells from each group).
USEPA (2006b) - This study was designed to field test new coliphage assay methods. Because
the objective was to better identify and count fecal indicators, where present, PWS wells with fecal
contamination were more likely to be selected. Most of the wells sampled in this study were not PWS
wells but rather were domestic water wells (not regulated by the GWR). The raw data were not proved by
the investigators so counting and analysis of the PWS wells is subject to error. Because the wells were
mostly not PWS wells and likely represented a biased set of PWS wells, these data are not included in the
exposure compilation in this EA.
Economic Analysis for the 4-40 October 2006
Final Ground Water Rule
-------
Borchardt et al (2004) - This study was designed to use detailed hydrogeologic data and
microbial assays to evaluate wells in one community that are not designated by the State as GWUDI
wells. Four wells were intensively sampled (two additional wells were sampled in one month as
substitutes). The study concluded that two of the four wells had substantial surface water contribution.
Because this study is small, all from the same community and half the wells are likely GWUDI, these data
are not included in the exposure compilation in this EA.
DeBorde et al (1995) - This study sampled two wells from the same community. Because this
study was small and both wells are located in the same community, these data are not included in the
exposure compilation in this EA.
Exhibit 4.13b presents a summary of the virus-positive results obtained from the 1,253 virus
assays performed in the studies described in this section. Exhibit 4.13c presents a similar summary of the
E.coli (indicator) positive results obtained from the 687 assays performed in these studies. Exhibit 4.13d
presents the number virus and E.coli assays used to evaluate wells in each of the 15 studies. These data
were used by EPA for the hit rate analysis discussed later in this chapter.
Economic Analysis for the 4-41 October 2006
Final Ground Water Rule
-------
Exhibit 4.13a New Data Available since publication of the Proposed GWR
Study
# PWS Wells
Sampled &
Location
Sampling frequency/volume
Indicators Monitored (#
Pos. Wells/# Wells Total,
Unless Otherwise Indicated
Pathogenic Viruses,
Legionella, (#Pos. Wells/ #
Wells Total, Unless
Otherwise Indicated)
Pennsylvania Noncommunity
Wells (Lindseyetal., 2002)
60 wells
59 samples.
Virus sample volume (200-1OOOL)
Bacterial sample volume (100ml_)
No detection limit
Measured values are 0.21, 0.21,
18.3, 33.4, and 52.0 MPN/100L
Male-specific coliphage (3/59)
Somatic coliphage (5/59)
Total coliform (27/59)
Łco//(7/59)
Enterococci (8/59)
C.perfringens (9/59)
H.pylori (by PCR) (4/59)
Cell culture: enteric virus (5/59)
Microbial Indicators for Assessing
the Vulnerability of ground water
to fecal contamination (Karim et
al., 2003, 2004)
20 wells (California-2,
lllinois-2, lndiana-3,
Massachusetts-2,
Missouri-2, New
Hampshire-2, New
Jersey-2, New
Mexico-1, Ohio-1,
Pennsylvania-3)
Each well was sampled monthly for
a year.
All indicators sampled using 100 ml_
and 1L samples (except coliphage
Method 1602, which used only
100mL samples)
Coliphage analyzed using Method
1601 and 1602
Method 1601: Male-specific
coliphage (1/20 for 100mL
sample, 4/20 for 1L sample)
Somatic coliphage (0/20)
Method 1602: Male-specific
coliphage (12/20)
Somatic coliphage (2/20)
Total coliform (13/20 for 100 ml_
sample, 16/20 for 1L sample)
Łco// (5/20 for 100 ml_ sample,
7/20 for 1L sample)
Enterococci (1/20 for 100 ml_
sample, 7/20 for 1L sample)
C.perfringens (1/20 for 100 ml_
sample, 3/20 for 1 L sample).
Cell culture:
enterovirus (2/20), Rotavirus
(7/20), RT-PCR: enterovirus
(5/20), rotavirus (9/20),
norovirus (8/20), adenovirus
(1/20)
Economic Analysis for the
Final Ground Water Rule
4-42
October 2006
-------
Environmental Factors and
Chemical and Microbiological
Water-Quality Constituents
Related to the Presence of
Enteric Viruses in Ground Water
(SE Michigan) (Francy et al.
2004)
38 wells
169 regular samples.
32 replicate pairs.
Mostly 5 samples per site.
Method 1601 & 1602.
Total coliforms (13/38) (34.2%)
(15/152 samples)
E.coli (4/38) (10.5%) (4/163
samples)
enterococci (6/38) (15.8%)
(7/158 samples)
Male-specific coliphage (2/34)
(5.9%) (2/117 samples) (1 L
Sample)
Somatic coliphage (1/34) (2.9%)
(1/118 samples) (1 L. sample)
Cell culture: enterovirus (2/34)
(2/93 samples)
RT-PRCR: enterovirus (4/38)
HAV (5/38)
Rotavirus (0/34)
Reovirus (0/34)
Norovirus (0/34)
Validation of methods to detect
coliphages in ground water
(USEPA, 2006b)
Phase II
Not Published
Note: This study included private
and public wells.
SE region (13 in NC
and 4 in FL)-27
wells
SW region (TX, NM) -
11 wells
Upper Midwest (MN) -
25 wells
NE region (12 in NH,
4 in ME, 3 in VT, 6 in
MA) - 25 wells
Two phases.
Somatic Coliphage SAL (19/116)
(16.4%)
F+ Coliphage SAL (13/1 16)
(11.2%)
Total Coliphage SAL
Somatic Coliphage enrichment
(8/1 16) (6.9%)
F+ Coliphage enrichment (4/1 16)
(3.4%)
Total Coliphage enrichment
(6/1 16) (5.2%)
Fecal coliform (11/80) (13.8%)
E.coli (5/1 16) (4.3%)
Enterococci ( 1 4/1 1 6) ( 1 2. 1 %)
Vulnerability of Drinking Water
Wells in La Crosse, Wisconsin to
Enteric-Virus Contamination from
Surface Water Contributions
(Borchardtetal.,2004)
6 PWS wells (not
GWUDI)
Sampled monthly for one year. Two
wells were shut down during one
sampling period; samples from
nearby wells were used for that
period.
2 wells have 12 samples; 2 wells
have 11 samples; and 2 wells have
1 sample.
TC (0/6)
E.coli (0/6)
Enterococci (0/6)
Somatic coliphage (0/6)
Male-specific coliphage (0/6)
RT-PCR:
Enterovirus (5/6)
Rotavirus (4/6)
Hepatitis A (3/6)
Norovirus G1 (3/6)
Norovirus G2 (0/6)
Cell culture:
Enterovirus (0/6)
Hepatitis A (3/6)
Economic Analysis for the
Final Ground Water Rule
4-43
October 2006
-------
Mountain Water Company in
Missoula, MT (DeBorde et al.
1995)
2 wells
Sampled monthly for one year.
F+Coliphage (1/2)(8%)
Somatic coliphage (0/2)(0%)
Enterovirus (0/2)
New Jersey (Atherholt et al.,
2003)
Note: This study included wells
that were ground water under the
direct influence of surface water)
26 wells
128 samples. Wells were sampled
from 1 to 10 times each. Bacteria
sample volumes were 100 ml.
Coliphage sample volumes were
100 ml_, but a few samples were
larger.
TC (8/26)
Łco//(3/26)
Enterococci (2/26)
Somatic coliphage (CN 13 host)
(5/26)
Male-specific coliphage (Famp
host) (5/26)
Montana Study (Miller and Meek,
2006)
18 wells (near
Helena, Montana)
Wells sampled 1-3 times
Eco//(0/18)
Enterococci (2/38) [uncertain if
the positive wells were PWS
wells]
Male specific coliphage (0/18)
Somatic coliphage (0/18)
Economic Analysis for the
Final Ground Water Rule
4-44
October 2006
-------
Exhibit 4.13b Virus Assays and Positives for 1,253 Wells Assayed for Viruses
Virus
Assays
1
2
3
4
5
6
8
12
14
Total
Number of Virus Positives
0
837
122
39
170
1
2
25
18
1
1215
1
28
1
1
2
3
35
2
1
1
3
1
1
4
1
1
Total
865
123
40
172
2
2
26
22
1
1253
Exhibit 4.13c E.coli Assays and Positives for 687 Wells Assayed for E.coli
E. Coll
Assays
1
2
3
4
5
6
8
12
11
14
Total
Number of E. Coli Positives
0
282
128
30
120
28
26
24
12
3
653
1
16
1
1
2
2
3
25
3
1
2
1
4
5
2
2
6
3
3
Total
298
129
30
121
30
27
26
22
3
1
687
Economic Analysis for the
Final Ground Water Rule
4-45
October 2006
-------
Exhibit 4.13d Number of Virus and E.coli Assays
Study ID
Abbas
Lieb
MD - 3 State, 2002
MD -3 State, 2001
SE Michigan
MN - 3 State
MO- 1
MO -2
New England
PA Noncommunity
Wl Migrant Worker Camp
Wl - 3 State
EPA Vuln
NJ Atherholt
Montana
Total
Number of Assays
Virus
539
298
91
30
95
299
218
109
458
60
0
200
30
0
0
2427
E.coli
0
298
90
30
167
92
218
109
462
60
126
200
0
71
38
1961
4.3.3 Well Vulnerability
4.3.3.1 Background
In the EA for the proposed rule EPA estimated that 17 percent of the wells in the United States
were improperly constructed and that 83 percent of the wells were properly constructed (ASDWA, 1997).
EPA used the Lieberman et al. 2002 data set to represent viral occurrence in improperly constructed wells
and the AwwaRF/AWWSC data set to represent properly constructed wells. It was implied that well
construction corresponded with vulnerability (i.e., poorly constructed wells would be vulnerable to
contamination). EPA received public comments that questioned the basis for using Lieberman et al. 2002
data to represent improperly constructed wells because the Lieberman et al. 2002 study sites were chosen
based on the presence of total coliforms and indicators of fecal contamination. To clarify this issue, EPA
is categorizing ground water systems into two groups: those that are more vulnerable and those that are
less vulnerable.11 Neither of the studies discussed above provide data regarding the percentage of ground
water sources that might be more or less vulnerable, and EPA needed to derive such estimates to support
EPA believes this terminology is more appropriate than that used in the proposal ("improperly
constructed" and "properly constructed") since the Lieberman et al (2002) study did not target poorly constructed
wells, but rather used criteria believed to favor the selection of vulnerable wells.
Economic Analysis for the
Final Ground Water Rule
4-46
October 2006
-------
this EA. EPA has used national TCR violation data from SDWIS to estimate the percent range of wells
that are more or less vulnerable. The proportion of wells in each of the well vulnerability categories is
necessary to properly apportion the virus concentration data. Viral concentration data from wells with a
history of TC contamination (i.e. the Lieberman et al. 2002 data) are used for the wells that are identified
as belonging within the more vulnerable group. Following is a description of these estimates and their
basis.
4.3.3.2 Estimating percent wells in vulnerability categories
EPA categorized systems into two groups: those that are more vulnerable and less vulnerable.
More vulnerable systems: These are systems that may be more vulnerable to source water
contamination, reflected by having MCL violations under the Total Coliform Rule (TCR) during
a calendar year (from SDWIS, USEPA 2003a).
Less vulnerable systems: These are systems that are expected to be less vulnerable to source
water contamination, reflected by having not had an MCL violation under the Total Coliform
Rule during the same year.
The percentage of systems in the "more vulnerable" category (and also the percentage in the "less
vulnerable" category) varies by system type (i.e., community, nontransient community, and transient
noncommunity) and system size, and ranges from zero to 6.83 percent. These proportions of wells in the
more vulnerable category are identified in Exhibit 4.14. For each element in the exhibit (system size and
type) the proportion of less vulnerable wells is 100% minus the value identified in the exhibit. Detail on
the derivation of these percentages is presented in Exhibit B. 18. MCL violations are of two types. Acute
violations indicate that the system tested positive for fecal coliform or E. coll in repeat samples following
samples that are total coliform positive. Non-Acute MCL violations for systems collecting at least 40
samples per month (i.e., those serving more than 33,001 customers) occur when more than 5 percent of
samples test positive for total coliforms during a sample period. For smaller systems, an MCL violation
occurs when more than one sample tests positive for total coliforms during a sample period (violations are
for a reporting period, which for most noncommunity systems is one sample per quarter; thus for these
systems, an MCL violation occurs when a repeat sample is positive for total coliforms). Either of these
two conditions indicates a potential problem with the integrity of the system and may indicate problems
with source water quality or other conditions that may make the system and its wells more vulnerable to
contamination.
Economic Analysis for the 4-47 October 2006
Final Ground Water Rule
-------
Exhibit 4.14 Mean Percent of Systems with Acute or Monthly MCL Violations
by System Type and System Size
System Size
(Population
Served)
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100K
100,001-1 Million
> 1 Million
CWS
2.87%
2.66%
1 .85%
2.23%
3.48%
3.41%
1 .82%
3.13%
0.00%
NTNCWS
2.84%
2.25%
1 .86%
2.34%
2.21%
0.00%
0.00%
0.00%
0.00%
TNCWS
2.33%
2.40%
2.30%
3.59%
2.82%
2.94%
0.00%
0.00%
0.00%
Source: Exhibit B.18, derived from SDWIS (2003a)
For ground water systems, the violation of MCLs under the TCR is an indicator of vulnerability,
especially when systems do not disinfect and distribution systems are small or do not exist. There is some
uncertainty associated with the data in Exhibit 4.14 because they include systems that disinfect as well as
those that do not disinfect. Exhibit 4.15 summarizes the available data on disinfecting systems. For
example, 64 percent of ground water systems provide no disinfection, and thus for such systems, TC
sampling under the TCR should reflect contamination in their source water. In transient noncommunity
systems, which essentially have no distribution systems, 82 percent of systems provide no additional
disinfection. In these systems, the influence of non-source water-related contamination is likely to be very
low relative to that of source water. In summary, it is assumed that disinfection has only a small
influence on the identification of more vulnerable wells using TCR violation data.
Exhibit 4.15 Number and Percent of Systems Disinfecting,
By Type of System
Number of Systems
Approximate Percent of
Systems Disinfecting
Total
147,330
36%
CWS
42,361
75%
NTNCWS
18,908
29%
TNCWS
86,061
18%
Source: Derived from Exhibit 4.2
Economic Analysis for the
Final Ground Water Rule
4-4
October 2006
-------
4.3.4 Occurrence Analyses
To assess potential costs and benefits of the GWR, it is necessary to estimate several occurrence
parameters. Estimates are made for both viral12 and indicator13 hit rates, viral concentrations, and co-
occurrence of viruses and indicators. Cost and benefit analyses performed for the EA accompanying the
GWR proposal drew data from two of the occurrence studies to inform the analyses - the Lieberman et al.
2002 and Abbaszadegan et al. 2003 studies. At the time, these two studies were considered to be the best
suited for representing viral and indicator hit rates as well as viral concentration. Specific co-occurrence
parameters derived from the studies were not used in the proposal EA analysis. Instead, a correlation
between indicator and viral occurrence was assumed. In response to comments received on the proposal
analyses, EPA performed further detailed review and analysis of all available occurrence data.
To improve the estimates of viral and indicator hit rates and concentrations using the data
available, EPA convened a 2-day statistical workshop in May 2005. The core workgroup included expert
participants form several government agencies and private consulting firms. A summary of the workgroup
proceedings, including a list of all participants, is included in the final docket for this rulemaking. The
charge to the workgroup was to consider how to obtain improved modeling of:
a) national viral occurrence in wells,
b) indicator efficiencies for identifying fecally contaminated wells,
c) indicator efficiencies for identifying virally contaminated wells, and
d) virus concentrations in virus-positive well water.
By the end of the workshop, approaches for modeling viral and indicator prevalence and viral
concentrations (items a, b, and d) were discussed, but methods for linking indicator occurrence and virus
occurrence (item c) were not. Following the workshop, EPA acted on the workgroup's recommendations,
provided feedback to participants, and generated model-based national estimates for both viral and
indicator occurrence. The results of this effort led naturally to a combined analysis, which also modeled
co-occurrence of viruses and indicators. This combined model serves as the basis of EPA's quantitative
occurrence estimates. The sections below describe in detail how these new data are used to model the
occurrence of virus and indicators in ground water sources.
The workgroup also considered the question of data selection with regard to the available
occurrence studies. Individually, the studies are not nationally representative, but represent select
portions of the ground water universe. Collectively, the studies describe a full range of geographic,
geologic, and other characteristics (e.g., variety of system sizes and system types). Workshop participants
recommended against discarding any study's data without cause but did not feel they had the expertise to
make any final calls regarding specific studies. Because the final selection of data was beyond the scope
of their expertise, the issue was remanded for further consideration by EPA's subject matter experts. The
discussion of each occurrence study is presented in Section 4.3.2, with a summary of the final use(s) of
each presented in Exhibit 4.12.
12 Although the GWR is aimed at preventing exposure to all viral pathogens, enterovirus data are used as a
proxy for all viral pathogens in both the viral hit rate and viral concentration analyses.
13 Although the GWR allows different indicators to be used for compliance purposes, E. coli is used as a
proxy for all indicators in the hit rate analyses.
Economic Analysis for the 4-49 October 2006
Final Ground Water Rule
-------
4.3.4.1 Viral and Fecal Indicator Hit Rates
This section discusses the estimation of the hit rates for viruses and fecal indicators in the source
waters of ground water wells. The rates are derived from pooled analyses of the data presented in Section
4.3.2, above. Hit rate information is a critical input for both the model of baseline risk of viral infections,
illnesses, and deaths and for modeling the reduction of viral risks from components of the GWR
dependent upon source water monitoring of fecal indicators. Indicator hit rate information is also a
critical input for cost modeling. The application of the hit rates for determining the baseline risk and risk
reductions from the rule options is described in detail in Chapter 5. The application of indicator hit rates
for determining the costs of some of the rule options is described in Chapter 6.
The term hit rate refers to the probability that a virus or fecal indicator, or both, will ever be
present in the source water of a well and, if so, how frequently each is expected to be present. Hit rates,
therefore, have two components which are referred to here as Pwell and Psampie.
Pwell refers to the probability that a randomly selected well will ever have a virus (or indicator)
present in its source water. Applying this probability to all wells provides an estimate of the number of
wells that ever have virus (or indicator) present in their source water.
Psampie refers to the probability that a random sample from a contaminated well will be positive
and will vary from well-to-well.
So, for example, a virus Pwell value of 0.10 implies that 1 out of every 10 wells will have
detectable virus present in its source water at some time. Conversely, it also implies that 9 out 10 wells
will not ever have detectable virus.
A well with a virus Psampie value of 0.25 would be expected to have detectable virus in 1 of every 4
samples assayed.
There are a number of factors that influence the estimation, as well as the interpretation, of Psample.
Microorganisms in water are dispersed spatially at low average concentrations relative to the volumes of
water typically collected in assays. As a result, a randomly taken sample of some volume V may not have
the microorganism present even when it is known to be in the source water. Often the recovery rate for
these pathogens is less than 100%, and viruses that are present in samples are not always detected. In
addition, the actual presence of microorganisms in the source water is recognized as being intermittent in
nature due to changes in the actual sources of the contamination as well as hydrogeological and other
physical factors affecting transport from the sources to the water used at that well.
It is important to recognize that while the same Pwell value applies to all wells, each individual
contaminated well is expected to have its own Psample value. That is, the underlying data suggest that
among those wells that have a virus or fecal indicator present at some time, the probability of observing it
in a given sample (that is, of it being present in that sample volume on that particular day) will vary from
well to well. Consequently, a distribution of Psample values was derived to reflect Psampk variability from
well to well. Specifically, a beta distribution of Psample was derived from the underlying occurrence data.
The beta distribution is often used for this purpose, that is, to describe distributions of probabilities or
other variables that range between zero and one. (The probit and logit distributions are sometimes used
for this purpose, and generally produce estimates similar to those produced using the beta distribution.)
Economic Analysis for the 4-50 October 2006
Final Ground Water Rule
-------
The beta distribution is a two-parameter distribution; the parameters are usually designated a and p. The
estimation of those parameters for the beta distribution of Psample is described further below.
The Venn diagram shown in Exhibit 4.16 describes the basic co-occurrence model. This diagram
shows that some fraction of wells (PI) has some virus contamination, but no indicator, while another
fraction of wells (P2) has both virus and indicator, and a third fraction of wells (P3) has indicator, but no
viral occurrence. A fourth fraction of wells (P4), having neither viral nor indicator occurrence is the
remainder: P4 = 1 - (P1+P2+P3).
Exhibit 4.16 Categories of Indicator and Viral Classification Among PWS Wells in
U.S.
PI
(virus, no indicator)
P2
(virus and indicator)
P3
(indicator, no virus)
P4
(no virus, no indicator)
P4= 1-(P1+P2+P3)
To fully characterize both Pwell and Psample for viral pathogens and fecal indicators, and to
characterize their co-occurrence, the model requires the estimation of seven parameters: PI, P2, P3, avims.
•rus? indicator?
well
and Vindicator from the available occurrence data. Pwell for virus is equal to P1+P2, and P
for indicators is equal to P2+P3. Psample for viruses (referred to hereafter as Pvsample) at different wells is
Economic Analysis for the
Final Ground Water Rule
4-51
October 2006
-------
described by a beta distribution with the parameters avirus and $virus. Similarly, Psample for indicators
(referred to hereafter as Pisampie) at different wells is described by a beta distribution with the parameters
amdicator and Vindicator-
It is important to note that the occurrence model developed by EPA relates virus and indicator
co-occurrence only in terms of Pwell (the fraction of wells having one, the other, or both). It does not
provide for different levels of Pvsample or Pisample in wells having both or only one of the two contaminants.
Wells having both virus and indicator presence may well have them greater fractions of the time than
wells having only virus or wells having only indicator. A model that would include this feature would
require four parameters to explain variable Pvsample and Pisample in wells having both virus and indicator.
However, the limited amount of occurrence data is not sufficient for a model with that degree of
complexity.
Another important point to note is that, while the preceding overview of the occurrence model
refers to the PI, P2 and P3 parameters for Pwell and the a and P parameters for Psample as though only single
"best values" are estimated, the occurrence model is actually designed to capture the uncertainty in those
values and produces a very large number (10,000) of sets of those seven parameters that are subsequently
sampled in the Monte Carlo simulations that are performed for both the risk/benefits model and the cost
model used to evaluate the impact of the ground water rule options.
Parameter Estimation Methods
Markov Chain Monte Carlo (MCMC) methods were used in a Bayesian framework to produce
samples from the joint posterior parameter distribution (the sample of 10,000 discussed in the paragraph
above). This posterior density function is a product of a prior density function and a likelihood function.
WinBUGS software (Gilks and Spiegelhalter, 1994) was used to produce the large MCMC sample,
which, in turn was used to inform the Rule's risk and cost analyses. This section describes the prior and
likelihood functions of the seven-parameter model.
Non-Informative Priors
Parameters PI, P2, and P3 [together with P4, where P4 = 1 - (PI + P2 + P3)] are the fractions of
all ground water wells falling into the four possible subsets as shown in Exhibit 4.11. A relatively
non-informative prior on these is Dirichlet with parameters (1, 1, 1, 1). This is the multivariate extension
of the Beta (1,1) distribution, which is often used for the one-parameter case. Beta (1, 1) is a uniform
distribution for one unknown over the range [0, 1] and likewise, Dirichlet(l, 1, 1, 1) is uniform over the
three-dimensional space where the sum of PI, P2, and P3 is in the range [0, 1].
Pvsample and Pisampie are both assumed to be beta-distributed across wells having virus presence and
wells having indicator presence, respectively. The Beta density function is usually expressed in terms of
its parameters a and P as:
/ n\
dbeta(p,a,p) =
a-l „ NB-1 r(
P
a
Economic Analysis for the 4-52 October 2006
Final Ground Water Rule
-------
where p in this expression is one of the Psampie variables. Assigning priors with this form of the density is
difficult. First, it is difficult because one cannot think about a (or P) without understanding p (or a).
Second, a non-informative prior leads to an improper posterior density, with increasing mass as the sum
(a + P) becomes large. In our case, a uniform prior on (a, P) suggests strong knowledge that the
variability of Psample is small about some mean value. Gelman, et al. (1995), discuss this problem in their
book "Bayesian Data Analysis" and suggest reparameterization in terms of the mean, a = a / (a + P) and
the inverse square root of the "sample size," b = l/(ot + P)°5. EPA adopted this parameterization and
utilized disperse uniform priors for the two new parameters (a and b). The conventional beta distribution
parameters can be derived from new parameters a and b as follows:
a = a / b2
P = (l-a)/b2
Therefore, for parameterization of the occurrence model, a,^, bvirus, a^^,, and bmdlcator are
estimated, and the corresponding a and p values for the beta distributions are computed from them as
shown above.
The Likelihood Function
The virus and indicator data for an individual well used as input to EPA's occurrence model can
be reduced to four integers. The four integers for a well are:
Nv = the total number of virus assays for the well
Kv = the number of virus positives for the well
Nj = the total number of indicator assays for the well
Kj = the total number of indicator positives for the well
Up to three of these may be zeros; at least one of the values Nv or N; must be >1 for it to be valid
input to the model.
The likelihood of a well's data, given parameter values (PI, P2, P3, • *irus,' *irus,' *ndicator, and
* Mcator, and the category of a well), is a function of the parameter values and the well's data (the well's
Nv, Kv, Np and K: values), where • ••„„ and • *lrus are parameters for beta-distributed Pvsample for viruses and
* Indicator and • fndicator are parameters for beta-distributed Pisample for indicators. The total likelihood (for the
entire data set) is simply the product of these individual well likelihoods.
In general, the likelihood for a well has three parts, the probability of what was observed for
virus, given the number of virus assays, the probability of what was observed for E. coli, given the
number of E. coli assays, and the probability of the well's membership in its category (PI, P2, P3, or P4).
Below, these two factors are defined for wells of the four different categories (virus only, virus and E.
coli, E. coli only, and no contamination):
Economic Analysis for the 4-53 October 2006
Final Ground Water Rule
-------
1. Well has some virus occurrence, but no E. coll occurrence (in area PI of Exhibit 4.16):
1
f
,N ) = PI-
v v/
•'o
where dbeta is the beta probability density function and dbinom is the binomial probability mass function.
A well of this type must have had no E. coll detections, so the probability of observing K: = 0 positives is
1 and its product with L1(KV, Nv) is simply L1(KV, Nv).
2. Well has both virus and E. coli occurrence (in area P2 of Exhibit 4.16):
fl
/Kv,NV dbeta(Psamplei,ai,pi)-dbiiiom(Ki,Ni,Psamplei)dPsamplei
J
= P2-LlK
Note that, to be in this category, it is not necessary that a well actually have observed virus and E. coli
positives. Having no positive, based on a small number of assays, is only weak evidence that a well
belongs to another category. At each uncertainty iteration, wells are assigned to categories according to
the likelihood, conditional on the well's data plus all other parameter values at that time. In this fashion,
parameters PI, P2, P3, and P4 also enter the likelihood. The only wells that are assigned to this category
with certainty are those which were observed to be positive for both viruses and E. coll. At every
iteration, they are placed in this category. All other wells are randomly assigned to different categories
from iteration-to-iteration, according to their likelihoods.
3. Well has E. colL but no virus occurrence:
= P3-
dbetalPsample^, aj, p J -dbinom/Kj, Nj, Psamplejj dPsamplej
4. Well has neither virus nor E. coli occurrence:
Wells having no observed contamination can belong to any category. Wells assigned to this category
must always have negative assays. The likelihood of observing no positives is certain, so the only
contribution to the likelihood is the probability of membership, P4.
Estimates for Combined Model
Estimates were produced by Markov Chain Monte Carlo methods using WinBUGS software
(Gilks and Spiegelhalter, 1994). An important feature of this MCMC analysis is that it produces a large,
well-mixed sample of outputs wherein each individual output contains a plausible value for each of the
seven parameters in combination with one another. In this modeling, EPA captured 10,000 sets of results
Economic Analysis for the 4-54 October 2006
Final Ground Water Rule
-------
for the seven parameters to characterize uncertainty about the parameter values. The MCMC modeling
captures the uncertainty in the parameter estimates through this large number of sets of results with
appropriate correlation structure.
The actual data used to estimate the seven parameters was the enteroviruses cell culture data for
viruses and the E. coll for the indicators data from the 15 occurrence studies described in Section 4.3.2.
Of those 15 studies, 12 have enterovirus cell culture data and 12 have E. coll data.
The following exhibits provide summaries of the Pwell and Psampie results obtained from the
modeling for viruses and indicators. Exhibit 4.17 shows the median values for PI, P2, P3 and P4. The
"error bars" included on the graphs reflect the 5th and 95th percentiles of the 10,000 values estimated. As
indicated earlier (refer to Exhibit 4.11), PI refers to the fraction of wells having virus at some time (but
no indicator), P2 refers to those wells having virus and an indicator at some time, and P3 refers to those
having an indicator at some time (but no virus). P4 are those wells having neither virus nor indicator
occurrence. Estimates were produced by Markov Chain Monte Carlo methods using WinBUGS software
(Gilks and Spiegelhalter, 1994).
Exhibit 4.17 Median of 10,000 Estimates of P1, P2, P3, and P4
(with error bars showing the 5th and 95th percentiles)
80.C
60.C
40.C
20.C
10.C
P1
P2
P3
P4
The median values obtained in the model for PI, P2 and P3 are 7.4%, 14.0% and 7.3%,
respectively. The median value for the sum P1+P2+P3 (wells with sometime presence of virus and/or
indicator) is 32.4%. The 5th and 95th percentiles on the sum of PI, P2 and P3 were found to be 18.2% and
70.6%. P4, the remaining wells that have neither virus nor indictor present at anytime is derived from the
model estimates for the other three as 1 minus (P1+P2+ P3). The median P4 value is 67.6%, with 5th and
Economic Analysis for the
Final Ground Water Rule
4-55
October 2006
-------
95th percentiles on P4 of 29.4% and 81.8%. Thus, approximately 90% of the 10,000 estimates of wells
having either virus or fecal indicator occurrence fall between about 20% and 70%, with a central estimate
of about 32%.
If E. coll was a perfect indicator of virus occurrence, there would be no wells with only virus or
only E. coll. PI and P3 would both be zero. Clearly, E. coll is not a perfect indicator of viral occurrence.
Exhibit 4.17 shows that most wells with virus occurrence tend to also have E. coll occurrence (P2 is
greater than PI) and that most wells with E. coll occurrence tend to also have virus occurrence (P2 is
greater than P3). Given that approximately 24% of wells have virus occurrence while 23% of wells have
E. coll occurrence, if viruses and E. coll were completely independent, then the fraction of wells having
both (P2) would equal the product 0.24 * 0.23, or 5.5%. The large median value of P2 (14.0%)
demonstrates that, though imperfect, E. coll is a positive indicator of viral occurrence.
As noted previously, two of the important hit rate values are Pwell for viruses and Pwell for
indicators. These are composed of P1+P2 for viruses and P2+P3 for indicators. Exhibit 4.18 provides the
median (and the 5th and 95th percentile values) for Pwell for viruses and for indicators.
Exhibit 4.18 Median of 10,000 Estimates of Pwel, for Virus and Indicator
(with error bars showing the 5th and 95th percentiles)
The median value of Pwell for viruses was found to be 23.6% with 5th and 95th percentiles of 9.8%
and 55.5%. The median of Pwell for indicators was 22.5% with 5th and 95th percentiles of 11.6% and
55.5%. As shown in Exhibit 4.18, median values and overall ranges of Pwell for viruses and indicators are
quite similar. However, the distribution of paired values for these covers a very wide range of
combinations. The scatter plot shown in Exhibit 4.19 shows the paired combinations of a sample of 1,000
of the 10,000 values. While most of the pairs tend to fall in the 10% to 20% range for both viruses and
Economic Analysis for the
Final Ground Water Rule
4-56
October 2006
-------
indicators, there are a substantial number that fall above this range including many where one value for
the pair is high and the other relatively low.
0.9 -
0.8 -
0.7 -
o 0.6 -
^ 0.5 H
o
S 0.4-
o.
0.3 -
0.2 -
0.1 -
Exhibit 4.19 Scatter Plot of PweM Pairs for Indicators and Viruses
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Pwell for Viruses
As described above, the Pvsample for viruses and Pisample for indicators are not single value estimates
but are, rather, distributions of values reflecting the variability in Psample from well to well. As a result, the
occurrence model generates 10,000 of these distributions for both Pvsample and Pisampie. It is difficult to
provide a summary of all 10,000 of those distributions, particularly because the beta distribution used in
this analysis can take on a wide range of shapes.
The beta distributions obtained for Psample have three different shapes: exponential, U-shaped, and
bell-shaped (right-skewed). Representative examples of these three shapes for Psample for viruses are
presented in Exhibit 4-20 as the density functions and in Exhibit 4-21 as the cumulative probability
distributions. (Note that these particular examples were selected because they present values that are
close to the central tendencies for the three distribution shapes of Psample for viruses.)
For Psample for viruses, about 73% of the distributions have the exponential shape, 23% have the
U-shape and 4% have the right-skewed bell shape. For Psample for indicators, about 79% of the
distributions have the exponential shape, 2% have the U-shape and 19% have the right-skewed bell shape.
Economic Analysis for the 4-57 October 2006
Final Ground Water Rule
-------
Exhibit 4.20 Density Function Shapes of Psampie for Viruses
Exponential (-73%
U-shaped (-23%)
Bell (-4%)
0.4 0.6
Psampie (Vims)
Exhibit 4.21 Cumulative Distributions of P
sample
for Viruses
o
Exponential (-73%
- U-shaped (-23%)
Bell (-4%)
0.4
0.6
Psampie (Vims)
One way to summarize the full set of 10,000 Psample distributions generated by the occurrence
model is in terms of the range and central tendency of their expected values. For Pvsample, the median of
the expected values is 9.4%, with 5th and 95th percentile values are 3.8% and 23.2%, respectively. For
Pisample, the median of the expected values is 12.7%, with 5th and 95th percentile values are 4.9% and
25.0%, respectively. These values are also shown graphically in Exhibit 4.22.
Economic Analysis for the
Final Ground Water Rule
4-58
October 2006
-------
Exhibit 4.22 Median of 10,000 Estimates of Psampie for Virus and Indicator
(with error bars showing the 5th and 95th percentiles)
0.300
0.250 -
0.200 -
0.150 -
0.100
0.050 -
0.000
Average Psampie Virus
Average Psampie Indicator
The range and central tendency of the expected values for Pvsample and Pisample are similar.
distribution of expected values for Pvsample and Pisampie
The
pairs produced by the model, shown in Exhibit 4.23
for a sample of 1,000 pairs, shows a substantial number of pairs where both have values in the 5% to 15%
range. However, there are a number where one of the pair is substantially higher (or lower) than the other
member of the pair.
Economic Analysis for the
Final Ground Water Rule
4-59
October 2006
-------
Exhibit 4.23 Scatter Plot of Means of Psampie Pairs for Indicators and Viruses
0.4
0.35-
0.3-
3*
in
Ł 0.25
(Q
U
'
f 0.2
Q.
8
Q.
"5 0.15
c
re
0.1 -
0.05-
: •:
*«* * * 4»*»» *
' * **' * * * * •* *
* * *
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Mean of Psampie for Viruses (av)
An important relationship that can be seen from the results of the occurrence modeling is between
paired Psample means and Pwell values. For both viruses and indicators, it was found that there is, generally,
an inverse relationship between them. The product of Pwell and the average of Psample is approximately
equal to the overall fraction of samples found to be virus-positive, therefore the inverse relationship is
expected. If either Pwell or the average Psample were increased without increasing the other, then
significantly more virus-positive results should have been observed across the survey data sets. Similarly,
a decrease in one, but not the other, would predict fewer positives than were observed. That is, a
characteristic of the uncertainty revealed by the 10,000 sets of results from the occurrence modeling is
that if the 'true' value of Pwell (the fraction of wells that have virus or indicator present at some time) is
high, the chance of finding the organism in a given sample at those wells tends to be low. Conversely, if
the 'true' value of Pwell is low, the chance of finding the organism in a given sample at those wells tends
to be higher. These relationships are shown in Exhibits 4.24 and 4.25 for viruses and indicators,
respectively, for a sample of 1,000 from the 10,000 sets of results produced by the occurrence model.
Economic Analysis for the 4-60 October 2006
Final Ground Water Rule
-------
Exhibit 4.24 Mean of Psampie Versus PweM for Viruses
(1,000 Pairs from Occurrence Model)
Exhibit 4.25 Mean of Psampie Versus PweM for Indicators
(1,000 Pairs from Occurrence Model)
Economic Analysis for the
Final Ground Water Rule
4-61
October 2006
-------
Use of Indicator Occurrence for Triggered Monitoring14
This section describes how the occurrence modeling described above, specifically that for the
indicator hit rates, is used to predict the number of wells "captured" by triggered source water monitoring.
The fraction of wells producing an indicator positive upon their first assay can be estimated as a function
of the following:
Piweii = fraction of wells with some indicator occurrence
oti = first parameter of beta-distributed Pisampie
pi = second parameter of beta-distributed Pisample
As discussed in the section above, in each uncertainty iteration of the occurrence model, a set of
parameter values describing indicator occurrence is selected from the MCMC sample (as well as
parameters for Pvwell and Pvsample that describe virus occurrence). The probability that an indicator positive
will be observed by the time of the ith assay can be obtained from Fni:
The probability that the i* assay will be the very first positive for the site is the difference Fn; -
F . = Piwell - Piwell •
r . -
dbeta(Ps,an,(3n) • | 1 - (1 - Ps)1+1 dPs
which simplifies to:
r(an+pn)-r(pn
• — AAVVWAA A A VVWAA / \ / \
n r(an+pn + i).r(pn)
F^J.J. This is, then, the fraction of all wells expected to return an indicator positive upon the 1th assay.
These probabilities (the Fn values) were derived for assays i = 1 through 200 for n = 10,000
uncertainty iterations. These probabilities are specifically associated with each set of the seven
occurrence parameters generated by the model as described previously.
Exhibit 4.26 shows the cumulative probability of having an indicator on or before the indicator
assay number. A sample of 1,000 sets was generated from the occurrence model, and three of the 1,000
curves are shown in the graph corresponding to the 5th percentile, median, and 95th percentile of all values
for that assay number. These data are used in the cost model simulation, discussed further in Chapter 6,
to determine whether, and if so when, a given well conducting triggered source water monitoring will
14 The discussion below can also be applied to estimate the impact of assessment monitoring as an optional
activity under the final rule, as well as under Alternative 3. The difference in the application of the analysis
described below is driven solely by the number of samples taken, which will be more under assessment monitoring
scenarios.
Economic Analysis for the 4-62 October 2006
Final Ground Water Rule
-------
have its first indicator positive and as a result initiate corrective action. These data suggest that of all
wells taking source water indicator samples, just under 20% would be expected to have a positive result
on or before the 200th assay, as a central tendency estimate, with an uncertainty range from
approximately 10% to over 25%.
Exhibit 4.26 Cumulative Probability of an Indicator Positive as a
Function of Assay Number - All Wells (used for cost analysis)
0.3
0.25
0.2
. 0.15
5 0.1
0.05
-95th Percentile of n = 1,000 Values
-Median of n = 1,000 Values
-5th Percenitle of n = 1,000 Values
50
100 150
Assay number
200
Similar data on the occurrence of the first indicator positive as a function of assay number are
used in the risk reduction model, as discussed further in Chapter 5, to determine the effectiveness of
indicator monitoring in source water to "capture" wells that are known to have virus present at some time
(that is, areas PI + P2 in the Venn diagram shown earlier in this chapter). For this part of the analysis, the
F^j values are adjusted to account for assays performed on those wells that are in the PI + P2 "space".
The adjustment made to the value of each assay probability result obtained as shown above is to multiply
it by:
P2
(Pl+P2)-(P2+P3)
Exhibit 4.27 shows the three corresponding distributions for these adjusted values used for the
risk reduction modeling. These data suggest that of those wells that have sometime presence of a virus
and take source water indicator samples, just under 50% would be expected to have a positive on or
Economic Analysis for the
Final Ground Water Rule
4-63
October 2006
-------
before the 200th assay, as a central tendency estimate, with an uncertainty range from approximately 20%
to 80%. The higher values shown here relative to the "all wells" data shown above reflect the outcome
that a much higher proportion of wells having some time virus presence also have some time indicator
presence [i.e., P2/(P1+P2) than do all wells [i.e., (P2+P3)/(P1+P2+P3+P4)].
Exhibit 4.27 Cumulative Probability of an Indicator Positive as a Function of
Assay Number -- Virus-positive Wells (used for risk reduction analysis)
95th Percentile of n = 1,000 Values
Median of n = 1,000 Values
5th Percentile of n = 1.000 values
100 150
Assay number
200
It is important to note also that both of these sets of results indicate that observing an indicator
positive in an early assay is more likely than on a later assay. This is because the structure of the model
accounts for the higher likelihood of observing positives among those wells where the frequency of
occurrence (that is, Psampie) is greatest.
4.3.4.2 Pathogen Concentration Analysis
The preceding section addressed hit rates, which comprise the first aspect of characterizing virus
occurrence in source water used by public ground water wells. This section addresses virus
concentrations which comprise the second aspect of occurrence.
Economic Analysis for the
Final Ground Water Rule
4-64
October 2006
-------
Hit rates primarily address the consideration of "presence / absence" of virus in the water. The
two components of hit rates are Pwell, which characterizes the fraction of wells where viruses are either
present at some time or are never present, and Psampie, which characterizes the fraction of samples or
duration of time that the organisms occur in those wells that have viruses at some time.
If a well is one in which the virus is never detected (which is expected to be the case the majority
of the time), the virus concentration is assumed to be zero. For those wells at which viruses are present at
detectable levels however, it is necessary to characterize the expected concentrations of viruses so that the
baseline risk and the risk reductions from regulatory alternatives can be estimated15.
The available information on virus concentrations in wells is limited. Although the analysis
performed here considers two categories of pathogenic viruses (Type A and Type B), useful information
on virus concentrations are only available from cell culture results for Type B viruses (enteroviruses).
For the purposes of this analysis, it has been assumed that the concentrations of Type A viruses are
similar to those for Type B.
Just as there is variability in virus occurrence with respect to prevalence (some wells have viruses
while others do not, and for those that do, the frequency of occurrence varies among contaminated wells),
so too, there is variability in expected concentrations of viruses from well to well among those wells
where viruses occur. As will be evident from the information presented, this variability encompasses
both large scale differences between those wells considered to be less vulnerable and those considered to
be more vulnerable, as well as differences from one location to another within each of these two
categories of wells.
As noted previously, participants in the May 2005 statistics workshop were asked to consider
how to model virus concentrations in virus-positive well water. Several options were considered both for
stratifying the wells into different categories to reflect different ranges of expected concentrations and for
fitting the concentration data to specific distributional forms to use in the baseline risk and risk reduction
modeling. However, no specific recommendations were made.
Following the workshop, EPA decided to stratify wells into two categories according to overall
vulnerability characteristics (more and less vulnerable wells). Unlike the hit rate analyses which draw on
data from 15 different studies, EPA relied upon only three key studies for viral concentration data. The
data from the Lieberman et al. 2002 study are used to represent virus concentrations in more vulnerable
wells and the combined data from the Abbaszadegan et al. 2003 study and the Pennsylvania study are
used to represent concentrations from less vulnerable wells. The Lieberman et al. 2002 concentration
data comes from wells that were included in the study because they had a history of TC contamination or
other evidence of vulnerability. As such, they are most like wells with TCR violations and therefore are
assumed to be representative of this group of more vulnerable wells. The Abbaszadegan et al. 2003
study and the Pennsylvania study include wells selected for reasons other than a TC occurrence history.
As such, they are assumed to represent the less vulnerable wells group. Furthermore, the Pennsylvania
wells are exclusively noncommunity wells and therefore the measured concentrations in these wells
represent the group of less vulnerable noncommunity wells.
Although hit rates were developed for both viruses and indicators, virus concentration modeling is necessary for the
risk and benefits analysis, \mtE.coli concentration modeling is not. For the indicators, the hit rate information is needed to
estimate risk reduction for the regulatory alternatives.
Economic Analysis for the 4-65 October 2006
Final Ground Water Rule
-------
Virus Concentration Data Used
Concentrations for More Vulnerable Wells
EPA identified the Lieberman et al. 2002 study as providing the most complete set of virus
concentration information for wells considered to be more vulnerable. These data are from cell culture
assays for Type B viruses. As described in section 4.3.2, seven of the 30 wells in this study were found to
have virus present by the cell culture method. A total of 20 positive values were observed. The
concentrations of the positive values are presented in Exhibit 4.28 below.
Exhibit 4.28 Summary of Virus Concentrations Observed
in the Lieberman et al. 2002 Study
Study Well Number
29
29
29
29
29
29
29
31
31
31
31
47
47
47
47
61
61
91
97
99
Concentration (PFU or MPN per 100 L)
6.55
12.32
27.01
0.86
3.72
2.01
10.59
19.63
15.37
10.76
9.61
45.33
3.17
43.99
47.72
53.37
25.17
12.78
9.52
212.51
Note: Shaded rows indicate State determined GWUDI wells.
Economic Analysis for the
Final Ground Water Rule
4-66
October 2006
-------
Concentrations for Less Vulnerable Wells
EPA identified the Abbaszadegan et al. 2003 and the Pennsylvania Noncommunity Well studies
as providing the most complete set of virus concentration information for wells considered to be less
vulnerable. These data are from cell culture assays for Type B viruses.
In the Abbaszadegan et al. 2003 data, there were a total of 22 samples taken from 21 different
wells with cell culture concentration data, as summarized in Exhibit 4.29 below.
Exhibit 4.29 Summary of Virus Concentrations Observed
in the Abbaszadegan et al. 2003 Study
Study Well Number
AZ-0001 / 3
AZ-0001 / 3
ID-0002
MO-0001
NH-1
IL-5
CA-1
PA-7
PA-21
NJ-13
CA-1 2
NJ-12
IL-10
IN-32
O-NY-15
O-WI-10
O-CA-22
O-CA-21
O-OH-6
OH-1
OH-3
IN-31
Concentration (viruses per 100 L)
1.89
0.18
0.09
0.36
0.19
1.56
0.45
0.15
0.17
0.17
0.45
0.18
0.18
0.64
0.18
0.46
0.92
0.18
0.19
0.92
0.15
0.18
Economic Analysis for the
Final Ground Water Rule
4-67
October 2006
-------
In the Pennsylvania Noncommunity Well study, there were a total of 5 samples taken from 5
different wells with cell culture concentration data, as summarized in Exhibit 4.30 below.
Exhibit 4.30 Summary of Virus Concentrations Observed
in the Pennsylvania Noncommunity Study
Study Well Number
HU425
JU372
CE396
CH 5994
BR852
Concentration (viruses
perl 00 L)
0.21
51.99
18.30
0.21
33.4
Application of Virus Concentration Data for Baseline Risk and Risk Reduction Models
As noted above, the participants in the May 2005 statistics workshop discussed alternative
distributional forms to fit to the concentration data for use in the risk and risk reduction models.
Following the workshop, EPA explored several options for fitting the data but determined that because of
the limited number of data points and the considerable variability in the data even within the two
vulnerability strata, that rather than fitting the data to a specific distributional form it was preferable to
use the data directly and draw from them randomly, with replacement, in the simulation model.
Therefore, for the baseline risk and risk reduction simulation models as described in Chapter 5,
each well that is identified as having virus present at some time has a concentration value drawn from one
of the 20 values shown above from the Lieberman et al. 2002 study if that well is in the more vulnerable
stratum, and from one of the 27 values shown above from the Abbaszadegan et al. 2003 and
Pennsylvania studies if that well is in the less vulnerable stratum.
The concentration thus selected is assumed to be the average concentration in those samples or on
those days when the virus is present. The use of these concentrations along with the Psample value for the
wells identified as having virus present is described in more detail in Chapter 5.
4.4 Outbreak Baseline and Causes of Contamination
CDC, EPA, and the Council of State and Territorial Epidemiologists have maintained a
collaborative surveillance program for collection and periodic reporting of data on waterborne disease
outbreaks since 1971. The CDC database and biennial CDC-EPA surveillance summaries include data
reported voluntarily by the States on the incidence and prevalence of waterborne illnesses. According to
the CDC-EPA database for ground water systems, between 1991 and 2000, a total of 68 outbreaks and
10,926 cases of illnesses were reported for GWSs (see Exhibit 4.32). Although CDC has data dating back
to 1971, the 1991 to 2000 data represent the best available data since the implementation of the current
Economic Analysis for the 4-68 October 2006
Final Ground Water Rule
-------
drinking water regulations (e.g., Total Coliform Rule). The reported outbreaks resulted from virus
contamination, bacterial contamination, and unknown factors. Exhibit 4.31 shows a summary of
waterborne disease outbreaks for ground water systems. Causes reported as miscellaneous are outbreaks
where insufficient data exist to accurately categorize the source of the contamination.
Exhibit 4.31 Summary of Waterborne Disease Outbreaks Attributable to PWSs
Served by Wells using Ground Water: 1991-2000*
Cause of Contamination
Community Water Systems
Untreated Ground Water
Treatment Deficiency
Distribution System Deficiency
Miscellaneous/Unknown
Total
Number of
Outbreaks
5
7
5
2
19
Percent
Outbreaks
26%
37%
26%
11%
100%
Cases of
Illness
167
1,624
803
183
2,777
Percent
Illnesses
6%
58%
29%
7%
100%
Cases
per
Outbreak
33
232
161
92
146
Noncommunity Water Systems
Untreated Ground Water
Treatment Deficiency
Distribution System Deficiency
Miscellaneous/Unknown
Total
Combined
Untreated Ground Water
Treatment Deficiency
Distribution System Deficiency
Miscellaneous/Unknown
Total
23
19
6
1
49
28
26
11
3
68
47%
39%
12%
2%
100%
41%
38%
16%
4%
100%
4,057
3,264
442
386
8,149
4,224
4,888
1,245
569
10,926
50%
40%
5%
5%
100%
39%
45%
11%
5%
100%
176
172
74
386
166
151
188
113
190
161
*Excludes disease caused by pathogenic protozoa in PWS since such systems are deemed
as ground water under the direct influence (and are subject to surface water treatment rule requiements).
Sources: CDC, 1993; Kramer et al., 1996; Levy et al., 1998; Berwick et al., 2000; Lee et al., 2002
The number of outbreaks reported to the CDC are believed to be an underestimate of the total
number of waterborne outbreaks that actually occur (National Research Council 1997a, Frost et al.,
1996). Some of the reasons for the lack of recognition and reporting of outbreaks include the following.
•• Some States do not have active disease surveillance systems. Thus, States that report the
most outbreaks may not be those in which the most outbreaks occur.
•• Health officials may not recognize the occurrence of small outbreaks, even in States with
effective disease surveillance systems. In cities, large outbreaks are more likely to be
recognized than sporadic cases or small outbreaks in which ill persons may consult
different physicians.
Economic Analysis for the
Final Ground Water Rule
4-69
October 2006
-------
•• Some States do not always report identified waterborne disease outbreaks to the CDC.
Reporting outbreaks is voluntary.
•• Most cases of waterborne disease are characterized by general symptoms (diarrhea,
vomiting, etc.) that cannot be distinguished from other illnesses.
•• Only a small fraction of people who develop diarrheal illness seek medical assistance.
•• Many public health care providers may not have sufficient information to request the
appropriate clinical test.
•• Even if a clinical test is ordered, the patient must comply, a laboratory must be available
and be proficient, and a positive result must be reported in a timely manner to the health
agency.
•• Not all outbreaks are effectively investigated. Outbreaks are included in the CDC
database only if water quality and/or epidemiological data are collected to document that
drinking water was the route of disease transmission. Monitoring after the recognition of
an outbreak may be too late in detecting intermittent or one-time contamination events.
•• The vast majority of ground water systems are NCWSs. Outbreaks associated with many
types of NCWSs may be less likely to be recognized than those in CWSs because
NCWSs generally serve nonresidential areas and transient populations.
Although they may be under-reported, documented outbreaks demonstrate that ground water
sources are not free of pathogenic contaminants and thus support the need for the GWR. The true
incidence of waterborne outbreaks and associated illness is unknown. In addition, persistent low to
moderate levels of endemic waterborne illness often go undetected by routine disease surveillance
programs. This lack of knowledge stems from inadequate surveillance of disease outbreaks, insufficient
outbreak detection methods, lack of epidemiologic investigation, and lack of microbial monitoring.
4.5 Summary of Uncertainties in Development of GWR Baselines
Uncertainty in this baseline analysis is due to limitations of the available information. These
uncertainties contribute to uncertainties in the cost and/or benefit estimates presented in Chapters 5 and 6.
Exhibit 4.32 presents a summary of these uncertainties and indicates their potential impacts on the cost
and benefits estimates.
Economic Analysis for the 4-70 October 2006
Final Ground Water Rule
-------
Exhibit 4.32 Summary of Uncertainties Affecting GWR Baseline Estimates
Uncertainty
Uncertainty in baseline
data inputs (SDWIS and
1 995 CWSS data)
Mixed systems in
baseline
CWS flow equations
used for NCWSs
Percent of wells in
vulnerable category
Percent of sensitive
wells
Viral hit rates
Viral concentrations
Indicator hit rates
Wells with viruses have
them at same levels
regardless of indicator
presence
Wells with viruses have
them the same fractions
of time regardless of
indicator presence
Wells with indicator have
it same fractions of time
regardless of virus
presence
Indicator and virus
occurrence not related
to well's hydrogeologic
sensitivity
Section
with Full
Discussion
of
Uncertainty
4.2.3
423
424
4.3.3
below
4.3.2,
4. 5.4.1, and
below
4.4.2 and
below
4.3.2
below
below
below
below
Effect on Benefit Estimate
Under-
estimate
X
x
X
X
X
Over-
estimate
X
Unknown
Impact
X
x
X
X
X
X
Effect on Cost Estimates
Under-
estimate
X
X
X
Over-
estimate
X
Unknown
Impact
X
x
X
X
X
X
Economic Analysis for the
Final Ground Water Rule
4-71
October 2006
-------
About half of the impacts are "unknown," meaning that they could lead to either over- or
under-estimates. Of those whose effect is under- or over-estimation, the two that seem most significant
are those for hit rate and concentration. Viral hit rates tend to be understated because virus recovery is
highly dependent upon the matrix and virus types and because the measurement methods fail to detect
many pathogenic viruses. Assuming zero viruses when none is detected leads to systematic
underestimation of average concentrations. Perhaps more important than these is the fact that the
occurrence model assumed no relationship between indicator occurrence and either a) the fraction of time
that virus is present in virally-contaminanted wells or b) the virus concentration levels that are realized
whenever virus is present. The only relationship modeled between indicator and virus is the fraction of
wells having both (P2) as compared to wells having only virus (PI) or only indicator (P3). It is likely that
wells containing E. coll and virus will have virus more often and at higher concentrations than in wells
containing only virus. Not modeling this relationship between E. coli and virus results in a significant
underestimate of benefits derived from the Rule's indicator-based corrective action.
The occurrence models do not distinguish between sensitive and nonsensitive wells. The benefits
analysis assumes no difference between these with respect to virus and indicator hit rates. The effects of
this assumption on the cost and benefits estimates are unknown. If occurrence (hit rates and/or
concentration) in sensitive wells is greater, or if the virus-indicator relationship is stronger, then the
benefits of monitoring are probably underestimated.
Economic Analysis for the 4-72 October 2006
Final Ground Water Rule
-------
5. Benefits Analysis
5.1 Introduction
The Ground Water Rule (GWR) reduces the public health risks of illness and death (morbidity
and mortality), by reducing public exposure to viral and bacterial pathogens present in ground water wells
due to intrusion of fecal matter. The health-related benefits of the GWR are due to reductions in both
endemic and outbreak risks and reductions in both acute and chronic illnesses. These health benefits and
non-health benefits are presented in Exhibit 5.1.
Exhibit 5.1 also presents a comparison of the subset of quantified benefits quantified in this EA
with the total (both quantified and nonqualified) GWR benefits. The quantified benefits are a small
subset of the total benefits because 1) of the many ground waterborne viruses avoided, only illnesses and
deaths from subsets of two types of viruses are quantified, 2) no quantified benefits accrue from avoided
acute and chronic bacterial illnesses and deaths, 3) only endemic illnesses and deaths avoided are
included in the analysis (epidemic illnesses and deaths are specifically excluded), 4) of these endemic
illnesses, only avoided acute illnesses and resulting deaths-from two types of viral illnesses* are
quantified, and 5) non-health related benefits are excluded.
Exhibit 5.1 Overview of Quantified and Nonquantified GWR Benefits
(continued on next page)
Benefit
Category
Total Benefits
GWR EA Quantified Benefits
Health Benefits
Reduction in
endemic illness
incidence
viral exposure risk reduction
(morbidity and mortality)
bacterial exposure risk reduction
(morbidity and mortality)
chronic sequelae reduction
reduction in secondary
transmission of viral or bacterial
illness from symptomatic and
asymptomatic individuals
acute rotavirus (Type A) illnesses
and deaths avoided
acute enterovirus (Type B)
illnesses and deaths avoided
subsets of viruses within Type A
and Type B categories and
bacterial illness and death not
quantified
reduction in secondary
transmission of viral illness from
symptomatic individuals
Reduction in epi-
demic (outbreak)
illness incidence
viral exposure risk reduction
(morbidity and mortality)
bacterial exposure risk reduction
(morbidity and mortality)
chronic sequelae reduction
reduction in secondary
transmission of viral or bacterial
illness from symptomatic and
asymptomatic individuals
Not quantified
Economic Analysis for the
Final Ground Water Rule
5-1
October 2006
-------
Benefit
Category
Total Benefits
GWR EA Quantified Benefits
Reduction in
treatment failures
Decreased illness through
minimizing treatment failures or
fewer episodes with inadequate
treatment
Not quantified
Non-health Benefits
Outbreak
responses
avoided
Avoided costs to affected water
systems, local governments
(provision of alternate water,
issuing warnings and alerts), and
community (decreased tourism due
to bad press).
Not quantified
Avoided costs of
averting behavior
reduced need or perceived need to
use bottled water, point-of-use
devices, etc. (includes time and
material costs)
less time spent on averting
behavior: hauling/boiling water, etc.
Not quantified
Increased
confidence
Perceived reduction in risk
associated with perceived
improvement in drinking water
quality
Not quantified
This chapter presents estimates of the GWR health-based benefits, including a discussion of
nonqualified benefits and provides estimates of the monetized value of the quantified avoided illnesses
and deaths. It describes the methodology of the risk assessment and benefits valuation that is outlined in
Exhibit 5.2. In addition, this chapter discusses health and nonhealth benefits of the GWR uncertainties and
sensitivities and presents a comparison of other regulatory alternatives considered. More detail on the risk
assessment for the GWR is presented in Appendices F and G.
5.1.1 Quantified Benefits
The quantified benefits include the avoided endemic acute illnesses (morbidity) and associated
deaths (mortality) each year from a subset of pathogenic viruses. The risk assessment is used to quantify
the number of baseline viral illnesses and deaths (i.e. those under current exposure conditions) and those
remaining after implementation of the GWR for the various regulatory alternatives. The differences
between the baseline estimates and those following implementation of the GWR are the avoided cases of
morbidity or mortality. EPA uses the estimates of avoided illness and deaths to establish the quantified
benefits of the GWR. Section 5.4 of this EA discusses the nonquantified benefits (both endemic and
epidemic), including chronic sequelae, for viruses and bacteria. The GWR total benefits include both the
quantified and nonquantified benefits, which are illustrated in Exhibit 5.1 above.
The risk assessment modeling requires a number of assumptions to be made regarding exposures
to viral pathogens in drinking water. An overview of the risk assessment methodology is presented in
Economic Analysis for the
Final Ground Water Rule
5-2
October 2006
-------
section 5.2.1. The hazard identification process is presented in section 5.2.2. The assumptions that are
required to assess pathogen exposures are presented in section 5.2.3. The dose response assessment is
presented in section 5.2.4. Baseline risk and the results of the risk reduction analysis (i.e., in terms of
illnesses and deaths avoided) are presented in section 5.2.5.
The benefits valuation process applies monetary values to the results of the risk reduction analysis
to estimate the value of the illnesses and deaths avoided in exposed populations. The assumptions and
inputs used to determine the unit values of avoided illness and death are discussed in section 5.3.1 and
Appendix A of this EA. The results of this analysis, the monetized benefits of avoided morbidity and
mortality for the GWR, are presented in section 5.3.2 and the results for all regulatory alternatives
considered are compared in section 5.7. The monetized benefits are compared to the rule costs for the
GWR and all regulatory alternatives in Chapter 8 of this EA.
Exhibit 5.2 Overview of Viral Pathogen Risk Assessment and Benefits Valuation
Procedure for Quantified Benefits (Main Analysis)1
Risk Assessment
Endpoint
Reduced Morbidity
Effects
Considered
Lost Work
Time
Reduced Mortality
Lost Work
Productivity
Valuation
Approach
Lost Leisure
Time
Medical
Costs
Pain and
Discomfort
Cost of Illness Analysis
Risk of
Premature
Death
Not Quantified:
Qualitatively Discussed
Benefits Transfer of Value
of Statistical Life from
Studies of Individuals'
WTP to Avoid Mortality
Risks
'This schematic presents an overview of only the quantifiable benefits of the GWR. The nonqualified benefits are expected to
comprise a significant portion of the overall benefits of the Rule and are presented in Section 5.4.
Economic Analysis for the
Final Ground Water Rule
5-3
October 2006
-------
5.1.2 Nonqualified Benefits
There are substantial benefits attributable to the GWR that are not quantified within this EA as part
of the main analyses because of data limitations. Nonqualified benefits due to etiological agents, disease
endpoints and exposure scenarios not considered in the main analysis are discussed in the nonqualified
benefits section (Section 5.4).
The monetized benefits from avoided acute illness due to many viral pathogens and all waterborne
bacterial pathogens are not quantified as a part of the main analysis. Chronic health effects associated with
viral and bacterial contamination are also not quantified. As discussed in section 5.4.2, the benefit of
avoiding these chronic cases may be significant, as affected individuals incur significant costs in medical
care and losses in productivity and quality of life in such instances. EPA considers these nonqualified
factors to be significant and a qualitative discussion of these nonquantified benefits is included in Section
5.4.
The natural history of infectious disease suggests that symptomatic and asymptomatic carriers can
infect other individuals by secondary transmission, either directly by physical contact and respiratory
droplets or indirectly via aerosols and contaminated surfaces. EPA estimates the number of secondary
cases arising from each primary case in the main analysis. Additionally, the nonquantified benefits section
(5.4) and Appendix E address this issue in a qualitative discussion supported by simulations of the
infectious disease process in large, susceptible (capable of being infected) populations.
In addition to the health-based benefits identified above, there are a number of nonhealth benefits
that also arise from promulgation of the rule. Other nonquantified benefits may result from overall system
improvements (e.g., upgrades to distribution systems, increased water treatment plant operational
efficiency and reliability, increased frequency/intensity of system surveillance), from improved risk
perception of drinking water quality, or from avoided outbreak response costs. While the value of these
nonhealth benefits are not quantified for this EA, these potential benefits are discussed qualitatively in
Section 5.4.
5.2 Quantified Health Benefits from Reduction in Exposure to Viruses
This section describes the risk assessment methods and assumptions used to quantify the baseline
health risks and the expected health benefits of the GWR. Quantified health benefits for a subset of
reduced pathogen exposure from the final GWR are derived from the risk assessment estimates of the pre-
and post-GWR annual endemic acute illnesses and deaths attributable to ground water source
contamination. Annual endemic acute illnesses are those that occur as a result of drinking water
contaminated with waterborne pathogens occurring under normal operating conditions. The main analysis
addresses only viral acute illnesses and, specifically, those caused by rotavirus and enteroviruses. Other
illnesses resulting from other etiologic agents, disease endpoints, and exposure scenarios are addressed in
the nonquantified benefits section (Section 5.4).
The two viruses quantified in the main analysis (Type A virus represented by rotavirus data and
Type B virus represented by enterovirus data) were selected because human challenge study data are
available for both viruses and because they are suitable representatives of two broad classes of pathogenic
viruses (Ward et al., 1986 and Schiff et al., 1984).
Economic Analysis for the 5-4 October 2006
Final Ground Water Rule
-------
5.2.1 Overview of Risk Assessment Methodology1
Risk assessment is an analytical tool that can be used to characterize the expected incidence of
adverse health effects associated with exposure to an environmental hazard. For the GWR, the EPA has
developed a risk assessment model to estimate the baseline number of endemic acute illnesses and deaths
associated with ingesting a subset of pathogenic viruses present in public ground water systems (GWSs).
The risk assessment uses a standard framework that is organized in accordance with EPA Policy for Risk
Characterization (USEPA 1995a), EPA's Guidance for Risk Characterization (USEPA 1995b), and EPA's
Policy for Use of Probabilistic Analysis in Risk Assessment (USEPA 1997b).
This standard framework requires the use of scientific data (or reasonable assumptions if data are
not available) to produce estimates of the nature, extent, and degree of a risk. Where there is uncertainty in
the data and assumptions used, that uncertainty is described and its impact on the risk estimates is
characterized. The risk assessment used in the GWR incorporates information on variability and
uncertainty associated with the data that characterize both the distribution of risk levels within the affected
population (variability) and the confidence bounds on key parameters of the risk assessment model
(uncertainty). Variability arises from true heterogenicity across people, places and time, and uncertainty
represents the lack of knowledge of the true value of the factor being considered (EPA 1997b).
According to the 1995 EPA Policy for Risk Characterization (USEPA 1995a), health risk
assessments for environmental contaminants generally involve four components:
Hazard Identification addresses the nature of the potential adverse health effects associated
with exposure to the contaminant.
• Exposure Assessment addresses both the number of people in the population exposed to
the contaminant and the distribution of levels of exposure within that population.
Dose Response Assessment addresses information concerning the relationships,
quantitatively where possible, between the magnitude of exposure to the contaminant and
the extent and severity of the adverse health effects that may occur.
Risk Characterization combines the hazard identification, dose response, and exposure
assessment information to describe overall risk to the exposed population, both in terms of
the distribution of individual risk levels in the population and the total number of cases of
adverse effects anticipated.
Exhibit 5.3 depicts these major elements of the risk assessment for characterizing the risk of illness
(morbidity) and death (mortality) from exposure to viral pathogens in drinking water systems covered by
the GWR. Each of these four components is addressed specifically for the GWR risk assessment in
sections 5.2.2 through 5.2.5.
1 For a more detailed description of the risk assessment process, see Appendix F.
Economic Analysis for the 5-5 October 2006
Final Ground Water Rule
-------
Exhibit 5.3 Health Risk Assessment Framework
Hazard Identification
•Viral pathogen health endpoints:
Morbidity and Mortality
Exposure Assessment
•Number of people exposed to viral
pathogens in drinking water
•Distribution of average daily ingestion
levels across the exposed population
Risk Characterization
•Estimated cases of viral illness and
death in the affected population
•Distribution of individual risks
Dose-Response Assessment
Relationships for the probability of:
•Infection given exposure
•Illness given infection
•Death given illness
The GWR is expected to reduce the current incidence (baseline) of acute and chronic illness
caused by a wide variety of viral and bacterial pathogens associated with fecal contamination of ground
water. Although the quantified benefits risk assessment accounts for only some viral pathogens and only
some acute illnesses and deaths associated with these viral pathogens, different pathogens can cause
different types of illnesses, and pathogens differ in their virulence, the degree or ability to cause illness or
death. Medical research has not isolated all waterborne pathogens, nor has it thoroughly characterized the
virulence of all of the waterborne pathogens that have been identified. As detailed in Chapter 4, there are
two types of viruses found in water with data judged to be adequate for a risk assessment: Type A
(rotavirus) and Type B (enteroviruses). EPA has chosen to quantify these two types of viruses to represent
the range of waterborne viral pathogens for the GWR risk assessment. A more comprehensive range of
bacterial and viral pathogens addressed by the GWR is discussed in Section 5.4.
The GWR risk assessment is based on routine exposure to two representative virus types (Type A
represented by data on rotavirus and Type B represented by data on echovirus or enterovirus). As
discussed in Chapter 4, Section 4.3.2, the virus assay method used in all viral occurrence studies was
developed and optimized for recovery of poliovirus in water. Poliovirus is a member of the enterovirus
group and so the virus assay is also relatively efficient at recovering the enteroviruses other than poliovirus
with the prominent exception of some coxsackie A viruses, which are not recoverable. Other viruses,
including Type A and similar viruses such as rotavirus and hepatitis A virus are substantially less
Economic Analysis for the
Final Ground Water Rule
5-6
October 2006
-------
efficiently recovered. Thus, the viral occurrence data used in the risk assessment are based primarily on
enterovirus recovery. Based on EPA's analyses of the occurrence data, enteric viruses such as Type B
viruses are intermittently present in a virus positive well each year for only short durations (generally from
a few days to a few weeks). EPA assumes that equal concentrations of Type A viruses are also present
when enteric viruses such as Type B are measured in wells during those short intervals of virus
contamination.
The first step of the risk assessment framework is to estimate the baseline (pre-GWR) number of
acute illnesses and deaths occurring from rotavirus (Type A viruses) and enteroviruses (Type B viruses).
An analysis is then conducted for conditions reflecting changes resulting from the final GWR regulatory
requirements (as well as several other regulatory alternatives) to produce an estimate of the number of
illnesses and deaths remaining after implementation. The quantified health benefits of the GWR are then
obtained from the difference between the baseline and the results under the regulation (the estimated cases
of illness and deaths avoided as a result of the regulation).
5.2.2 Hazard Identification
This section presents summary information on the adverse health effects associated with ingesting
waterborne viral pathogens, including a discussion of the effects on sensitive subpopulations. The acute
and chronic health effects associated with bacterial pathogens are included in the nonqualified benefits
discussion (Section 5.4).
5.2.2.1 Health Effects of Viral Infections
A review of the medical and epidemiological literature reveals that the extent of acute viral health
effects varies by severity and by population subgroup for each virus. As described previously in Chapter
4, the viruses of concern were categorized as Type A (represented by rotavirus) and Type B (represented
by enterovirus). The subsections below summarize the typical health effects of each virus type.
Additional descriptions of the health effects of viral illnesses are presented in Chapter 2 and Section 5.4 of
the GWR EA.
Type A Viruses (rotavirus)
Generally, Type A viruses are highly infectious viruses that cause acute gastroenteritis, resulting in
symptoms that include watery diarrhea, fever, abdominal pain, and vomiting. Although they are highly
infectious, Type A viruses generally lead to mild, non-life-threatening acute illnesses. Common strain
rotaviruses typically affect young children, particularly those less than 3 years old, but other strains (Gl,
G2 and G9) have been found to be common in adults and the elderly in nursing homes, and found to be
responsible for more severe illness in children (Griffin et al., 2000). As discussed further in Section 5.4,
most ground water-borne gasteroenteritis associated with viral exposure in the U.S. is believed to be the
result of norovirus, although rotavirus outbreaks are known (e.g. Gallay et al, 2006). Norovirus causes
gastroenteritis with vomiting in both adults and children. Due to data limitations, this EA only addresses
the disease burden related to a few strains of rotavirus, which rarely causes illness with debilitating
symptoms in adults and is not often accompanied by vomiting in either adults or children. Therefore, the
EA quantifies only one part of the total Type A virus risk.
Economic Analysis for the 5-7 October 2006
Final Ground Water Rule
-------
Type B Viruses (enteroviruses)
People of all ages may experience various adverse health conditions as a result of ingesting Type
B viruses. Type B viruses are less infectious than Type A but result in more severe illnesses than Type A
viruses. For this type of virus, acute illnesses are classified into three levels of severity:
• Illness that does not require medical attention (mild symptoms)
Illness that requires a doctor visit
• Illness that requires hospitalization
Mild enteroviral illness include nonspecific febrile illness, respiratory illness, photophobia or
sensitivity, stiff neck, and gastrointestinal illness. Aseptic meningitis may or may not require a doctor's
visit, but more severe illnesses such as viral encephalitis, myocarditis, and non-polio flaccid paralysis are
likely to require hospitalization. Most likely to be hospitalized are young infants (<3 months old) with
non-specific febrile illnesses that require treatment to rule out and expectantly treat serious bacterial
illness. Chronic illnesses such as diabetes and dilated cardiomyopathy more likely result from Type B
infection as compared with Type A (chronic illnesses are not quantified in this EA but are discussed in
Section 5.4).
5.2.2.2 Sensitive Subgroups
Although it is generally believed that most people are vulnerable to repeated infection by viruses
and other microorganisms during their lifetime, factors such as being very young, elderly, or
immunocompromised can affect the probability of illness given exposure, the severity of illness, and the
likelihood of a fatal outcome. Virus morbidity and mortality (see section 5.2.4 for further description of
these factors) incorporated in the GWR model include higher rates for some age groups, typically for
neonates (Type B viruses) and children (both Type A and Type B viruses). Some of these sensitive
populations are presented with their percentages of the United States population in Exhibit 5.4.
Economic Analysis for the 5-8 October 2006
Final Ground Water Rule
-------
Exhibit 5.4 Sensitive Populations in the United States
Sensitive Population
Individuals
% of US
population1
Citation/Notes
Pregnant women and neonates
Pregnant Women
Neonates ( < 1 month) 2
6,240,000
317,137
2.2
0.1
Vital and Health Statistics, CDC (Ventura, 2000)
US Bureau of the Census (2000)
Age-based sensitive populations
Children ( < 5 years)
Elderly ( > 64 years)
19,175,798
34,991,753
6.8
12.4
US Bureau of the Census (2000)
US Bureau of the Census (2000)
Compromised Immune Status
Bone marrow transplant recipients
AIDS patients
Organ Transplant Recipients
20,000
816,149
23,143
0.01
0.3
0.01
National Marrow Donor Program
http://www.marrow.org/MEDIA/facts_figures.pdf
HIV/AIDS Surveillance Report, cases through
2001 (CDC, 2002)
US Bureau of the Census, Stat Absof US, based
on 1998 data 2001 b)
1U.S. Census estimtate July 2000.
2Reflects the estimated number of newborns age <1 month (1/12 of the 2000 census population for age < 1 year).
The Very Young
The very young [e.g., infants (neonates) less than one month old] are considered to be more likely
to develop severe illness and death from gastroenteritis and other waterborne viral and bacterial infections
than the general population. Most vertical transmission between mother and infant occurs during and after
childbirth. Viral gastroenteritis, caused mainly by Type A viruses, is prevalent among U.S. children. The
Agency's document, "Health Risks of Enteric Viral Infections in Children" provides a comprehensive
review of viruses that pose adverse health implications for children (USEPA, 2000g). Primary food or
waterborne exposure followed by secondary transmission via the fecal-oral route contributes to high rates
of illness in group settings where care is provided for children that wear diapers such as day-care centers.
The Centers for Disease Control and Prevention (CDC) has determined that the incidence of rotavirus
diarrhea can reach 0.30 episodes/child/year by age two, with a cumulative incidence approaching 0.80
episodes/child by age five (Glass et al., 1996). Hospitalizations for rotavirus diarrhea are most common in
children 6 months to 3 years of age (Parashar et al., 1998), while self-limiting norovirus infections are
prevalent in school-age children (LeBaron et al., 1990). Although deaths from infectious diarrhea have
generally declined among U.S. children since 1965 because of re-hydration therapy, newborn children,
especially infants born prematurely, remain at risk of death from severe diarrheal illness (Kilgore et al.,
1995).
In addition, some viral pathogens such as coxsackie B virus (a Type B enterovirus) can be
transmitted transplacentally from an infected mother to her child in utero, during birth, or shortly
thereafter. This type of transmission places the infected newborn infant at risk of severe symptomatic
illness from meningitis or myocarditis, for which the case fatality rates are high (Gerba et al., 1996a).
Economic Analysis for the
Final Ground Water Rule
5-9
October 2006
-------
The Elderly
The elderly (individuals over 65 years of age) are also at greater risk than the general population of
experiencing severe health effects from rotavirus diarrhea, hepatitis, and other viral infections. Sensitivity
among individuals in this age group is due to declining immunity and poorer general health (Gerba et al.,
1996a and Lew et al., 1991). Conditions such as cardiovascular disease make the elderly more susceptible
to complications of diarrhea such as electrolyte imbalance, dehydration, and shock (Maasdam and Anuras
1981). More than half of the diarrheal deaths that occur in the United States are among persons older than
74 years of age, and the risk of death from diarrhea is generally higher among elderly persons confined to
nursing homes and other care facilities (Lew et al., 1991; Gerba et al., 1996b).
In this EA, due to limited data the morbidity and mortality factors assigned to the elderly for Type
A and B virus infections are the same as those used for the general population. These factors
underestimate the risks for elderly individuals. The EPA believes there are insufficient data available to
assign higher morbidity or mortality rates from waterborne viral infections.
The Immunocompromised
Immunocompromised and immunosuppressed persons comprise a population subgroup who are
more susceptible to viral and bacterial infections and are more sensitive to serious health effects from
them. More specifically, Acquired Immunodeficiency Syndrome (AIDS) patients (e.g. Morpeth and
Thielman, 2006), organ transplant patients, and persons undergoing bone marrow transplantation are
considered sensitive to viruses based on compromised immune status. These immunocompromised groups
constitute approximately 0.3 percent of the general population. This estimate is based on best available
data and assumptions as summarized in Exhibit 5.4. The estimate does not include other groups of
immunocompromised people, including those with chronic illness or those cancer patients who are
undergoing chemotherapy and radiation (Morris and Potter, 1997). The population estimate also does not
account for the estimated 25,000 to 50,000 persons in the U.S. with primary immunodeficiency diseases
that primarily affect B-lymphocyte cell function. The most common and serious primary immune diseases
are X-linked Agammaglobulinemia (NICHD, 1999; Gerwurz et al., 1985; McKinney et al., 1987; Hertal et
al., 1989). In addition, the population estimate does not include persons with potentially autoimmune
diseases (e.g. type 1 diabetes, myocarditis, multiple sclerosis and rheumatoid arthritis) who may also
experience more serious adverse health effects brought on by viral and other infections (Fujinami, 2006).
A limited number of available studies suggest that viral infections can contribute to deaths in
immunocompromised persons. Chronic diarrhea is a serious complication of AIDS, and rotavirus and
adenovirus are commonly isolated from stool samples from AIDS patients with diarrhea (Gerba et al.,
1996a). Enteric rotavirus and coxsackievirus infections were reported as the cause of death among bone-
marrow transplant patients (Yolken et al., 1982).
In this EA, the morbidity and mortality factors assigned to immunocompromised subgroups for
Type A and B virus infections are the same as those used for the general population. These factors
underestimate risks for these subgroups, but the EPA believes that there are insufficient data available to
assign higher morbidity or mortality rates from waterborne infections based on immune status. However,
because there is a higher cost-of-illness among severely immunocompromised persons having lengthy viral
illnesses, the reductions in numbers of illnesses and deaths attributable to the GWR in this subgroup are
calculated separately by the risk model and valued separately in the monetary benefits calculations.
Economic Analysis for the 5-10 October 2006
Final Ground Water Rule
-------
5.2.3 Exposure Assessment
This section discusses the key elements for characterizing human exposure to viral pathogens in
drinking water. The primary exposure pathway is ingestion of drinking water from public ground water
supplies that are contaminated with viruses from fecal pollution.2 EPA believes that there are three key
exposure scenarios or events whereby virus-contaminated drinking water is delivered to consumers in
public GWSs.
1) Source water contamination under normal operating conditions
2) Source water contamination due to treatment failures
3) Contamination due to distribution system deficiencies
The remainder of this discussion of the main exposure assessment for the GWR addresses only the
first contamination scenario due to data availability. EPA's modeling of this first scenario includes
exposure from both systems that serve untreated ground water and systems with low levels of viral
inactivation.
With regard to the other scenarios, there is insufficient information on the frequency and severity
of drinking water treatment failures and distribution system deficiencies in ground water public water
systems (PWSs) to directly model these events. However, the proportions of reported outbreaks caused by
the three contamination scenarios listed above serve as an indicator of the relative frequency of GWS
contamination events that cause disease in exposed populations (see Section 5.4.7 and 5.4.8 for further
discussion).
The assessment of exposures to pathogens from ground water sources requires that the following
factors be quantified:
• Occurrence (presence/absence) of pathogens in source water (including duration of time
that it is present)
Concentration of pathogens in source water when it is contaminated
• Level of pathogen inactivation in the system and resulting pathogen concentration in tap
water
Size of the exposed population, including sensitive subgroups
Volume of water ingested daily and how many days per year it is ingested
EPA evaluated available occurrence and exposure data and developed assumptions where needed
regarding these exposure factors, each of which is discussed briefly below.
The risk assessment also considers exposure via secondary spread from those who become ill through
drinking water. This is addressed in section 5.2.4.
Economic Analysis for the 5-11 October 2006
Final Ground Water Rule
-------
5.2.3.1 Source Water Viral Occurrence and Concentration
Virus occurrence values for this exposure assessment are based on occurrence data from 12 studies
of U.S. ground water source quality. Presence/absence of enteroviruses were directly measured in 12
studies using methods optimal for recovering enteroviruses but not other viruses.3 Viral concentrations are
based on three of the 12 studies that included concentration data (Lieberman et al, 2002; Abbaszadegan et
al, 2003; Lindsey et al., 2003). EPA's risk model applies a randomly chosen concentration from the data
in these three studies. However, concentration data from the Lieberman data set is only applied to a small
subset (2.5% on average) of wells with predicted viral presence, whereas the concentration data from the
Abbaszadegan and Lindsey data sets are applied to all other wells (97.5% on average) with predicted viral
presence (see section 4.3.4.2 for further discussion). More detailed descriptions of each study are
presented in section 4.3.2 and in the Ground Water Rule Occurrence and Monitoring Document (USEPA,
2006b).
Modeling of the available virus occurrence data predicts that enteric viruses such as Type B
viruses are intermittently present in a virus positive well each year for only short durations (generally from
a few days to a few weeks). EPA assumes that equal concentrations of Type A viruses are also present
when enteric viruses such as Type B are measured in wells during those short intervals of virus
contamination. The probability that a well or sample is virus contaminated is unchanged by this
assumption.
EPA recognizes that Type A virus concentration assumed present in wells could be either an
underestimate or an overestimate. The assumption could be an underestimate if Type A viruses are present
at higher concentrations or if they contaminate wells for longer intervals than assumed because they may
be more prevalent in the human population and may be shed at much higher concentrations than Type B
viruses. The assumption could be an overestimate if the prevalence of Type A viruses in human
populations is less than the Type B viruses.
EPA considered using other assumptions about Type A occurrence but identified problems with
each. EPA considered randomly assigning Type A or Type B to the randomly chosen viruses
concentrations or assigning a fixed percentage to the number wells that are assumed to exhibit Type A
virus concentrations. However, these assumptions would contradict available data by assigning Type A
virus character to samples that are measured using methods that are favorable for recovering Type B
viruses.
5.2.3.2 Finished Water Concentrations in Disinfecting Ground Water Systems
For the purposes of this exposure assessment, EPA assumed that the pathogen concentration in tap
water from GWSs that serve untreated ground water is the same as the pathogen concentration in source
3 The infectious viruses counted in the occurrence studies represent a group of enteric viruses that are
favored for recovery because they are most efficient at infecting the host cell line used in the concentration
measurements. Poliovirus is most favored but other closely-related Type B enteroviruses (including echovirus) are
also likely to be recovered. Rotaviruses are also likely to be present because they are shed in fecal material at
concentrations several orders of magnitude greater than the enteroviruses. However, rotavirus is not efficiently
recovered in that commonly used host cell line and other Type A viruses, such as norovirus, are not recoverable.
Economic Analysis for the 5-12 October 2006
Final Ground Water Rule
-------
water. In contrast, properly operating disinfecting systems are assumed to inactivate 99 percent (2-log), or
99.99 percent (4-log) of viral pathogens depending on the disinfection practices employed. Therefore, the
concentration of pathogens in tap water from properly operating disinfecting (4-log inactivation) systems
is assumed to be 0.01 percent of the concentration in source water. For 2-log inactivation, the
concentration is assumed to be 1.0 percent of the concentration in source water. These finished water viral
concentration estimates do not take into account the effects of elevated concentrations from any upsets or
treatment failures that might occur.
5.2.3.3 Size of Exposed Population
As presented in section 4.2.3 and in Exhibit 4.4, ground water CWSs serve over 100 million
people, while ground water NCWSs serve about 14 million people. This population has the potential to be
exposed to pathogens via drinking water from GWSs.
5.2.3.4 Drinking Water Consumption Factors
The amount of drinking water consumed daily by individuals is a key input to the exposure
assessment component of the risk analyses. The higher the average daily consumption of water by an
individual, the higher the risk of infection for a given level of pathogen occurrence. EPA bases its
estimates of per-capita water ingestion on data collected by the U.S. Department of Agriculture's (USD A)
1994-96 Continuing Survey of Food Intakes by Individuals (CSFII). Data derived from this survey are
presented in the report, "Estimated Per Capita Water Ingestion in the United States" (USEPA 2000c). For
noncommunity water systems, which represent a significant number of the systems potentially impacted by
the GWR and where individuals consume water for shorter periods, EPA further adjusted the USDA
estimates to better reflect drinking water patterns at those systems. These adjustments were based on the
classifications of noncommunity water systems from the EPA report, Geometries and Characteristics of
Public Water Systems (USEPA, 2000a) and EPA estimates of daily ingestion in these systems.
The EPA water ingestion study used to estimate water consumption reports information for two
different aggregations of the population: all respondents (which is used in this exposure assessment) and
the subset who report consuming water directly ("consumers"). The category of all respondents is more
appropriate to this exposure assessment as EPA assumes that all people consume or are exposed to tap
water, even if they reported no tap water consumption during the three day diary period of the CSFII
survey. This is because even people who report no direct consumption may consume some public water at
other times during the year. Furthermore, they still do ingest water indirectly (for example, through
washing vegetables and other foods, and consuming foods prepared in restaurants) or are otherwise
exposed to potential waterborne viruses in tap water (during showers and brushing teeth, for example).
Exhibit 5.5 summarizes the CSFII data for consumption of water from all sources.
Economic Analysis for the 5-13 October 2006
Final Ground Water Rule
-------
Exhibit 5.5 Distribution of Individual Daily Drinking Water Consumptionby Age
Group (L/person/day)
Age
(years)
<0.5
0.5-0.9
1-3
4-6
7-10
11-14
15-19
20-24
25-54
55-64
65+
All ages
Mean
0.409
0.569
0.417
0.544
0.604
0.811
0.990
1.271
1.480
1.529
1.451
1.232
Percent! les
D1
-
-
0.001
0.004
0.006
0.01
-
0.001
0.041
0.118
0.245
0.009
D5
-
0.03
0.046
0.087
0.115
0.119
0.108
0.117
0.301
0.473
0.531
0.163
D10
-
0.86
0.09
0.147
0.174
0.209
0.231
0.237
0.473
0.652
0.651
0.283
D25
0.002
0.248
0.196
0.276
0.305
0.382
0.407
0.554
0.798
0.946
0.935
0.573
D50
0.394
0.548
0.346
0.462
0.512
0.643
0.768
1.000
1.272
1.378
1.344
1.037
D75
0.696
0.771
0.580
0.719
0.808
1.066
1.276
1.577
1.893
1.952
1.832
1.633
D90
0.903
1.126
0.805
1.017
1.130
1.623
1.891
2.506
2.631
2.557
2.323
2.341
D95
0.969
1.272
0.993
1.267
1.422
1.960
2.387
3.608
3.333
2.997
2.708
2.908
D99
1.307
1.671
1.393
2.026
2.170
3.025
4.020
5.796
5.244
4.393
3.747
4.805
Source: "All Sources" from USEPA 2000c. Includes bottled water and tap water; see text describing the
adjustment made to account for bottled water consumption.
The survey also reports information by type of source water (community water, bottled water,
other sources, and non-reported source). The survey questions categorized respondents based on their
reported "main" source of direct water and indirect water. Thus, many respondents who reported that their
main source of drinking water was bottled water or another source, may still consume water from
community sources at least some of the time. Likewise, respondents who categorize their drinking water
as being mainly from a community source may also consume bottled water and water from other sources.
More importantly, those who are not now served by community water systems would report "other
sources" or bottled water as their main source of water. These groups are presented as components of the
national average and cannot be subtracted from the total without affecting the value of the national
average. Thus, the consumption of those who reported no source or "other sources" as their main source
of drinking water are included in the national average, because subtracting these categories would lead to
an underestimate of average consumption levels. Their consumption patterns are assumed to be similar to
those served only by community systems (or at least, cannot be adjusted using the available data).
Economic Analysis for the
Final Ground Water Rule
5-14
October 2006
-------
Therefore, there were no adjustments for those who reported no source or other sources as their main
source of water.
The consumption of bottled water, however, is thought to reflect a replacement of tap water, and
thus an adjustment for this is more appropriate. Further, there is evidence that consumption of bottled
water is significant. In the CSFII study, 13.4 percent of all water was consumed by those who categorized
bottled water as their "main" source of water for direct or indirect ingestion. The implicit assumption is
that the distribution of consumption for bottled water drinkers is proportionally about the same as for all
individuals. EPA believes the closest approximation of the distribution of consumption for the GWR is the
distribution for "All Sources Less Bottled Water" (13.4 percent). Because the survey did not attempt to
determine for each individual the proportion of water from each source, the approach used in the exposure
assessment (i.e., applying the 13.4 percent reduction to all mean values) may understate or overstate actual
consumption from public water systems, depending on the extent and direction of overlap in drinking
water sources. EPA believes however, that making this adjustment produces an estimate of drinking water
consumption closer to actual practices. In the EA model, EPA used the mean values for each size category
in Exhibit 5.5, adjusted downward by 13.4 percent. EPA used this approach rather than considering the
percentile distribution because this gives similar estimates and reduces modeling complexity.
To better estimate the daily drinking water consumption patterns for the populations served by
noncommunity GWSs, which reflect a significant number of systems potentially impacted by the GWR,
EPA further adjusted its estimates of daily consumption. Adjustments were made based on the system
classification (non-transient or transient) and the area served (i.e. campgrounds, restaurants, etc).
Nontransient noncommunity water systems, by definition, serve the same population each day and
many of these system are businesses where employees spend their work day. Based on this definition,
EPA assumes that populations are served by these systems for 8 eight hours of each day and then return
home to be served by a CWS. For the purposes of estimating exposure, EPA assumes that populations
served by NTNCWs consume 50 percent of the mean daily consumption levels (first reduced by 13.4
percent) from Exhibit 5.5 for each population category.
For transient noncommunity GWSs, EPA estimated daily consumption for the 25 types of transient
noncommunity GWSs, that serve the largest transient populations. EPA used Best Professional Judgement
to estimate daily consumption in each of these 25 system types and weighted these consumption levels by
the populations served by these systems. This weighting captures the influence of the significant
differences in populations served by different types of systems. To estimate exposure in all transient non
community GWSs, EPA used the average of all 25 weighted daily consumption estimates. As with the
NTNCWS adjustment described above, the average TNCWS consumption factor of 0.4 is applied to the
mean daily consumption levels (first reduced by 13.4 percent) from Exhibit 5.5. Details of this analysis
are presented in Exhibit 5.6 below.
Economic Analysis for the 5-15 October 2006
Final Ground Water Rule
-------
Exhibit 5.6 Derivation of Transient Noncommunity Water System Consumption
Factor
TNCWS Type
Restaurants
Churches
State Parks
Wholesalers
Summer Camps
Campgrounds
Hotels/Motels
Highway Rest Areas
Misc. Recreational Services
Service Stations
Golf & Country Clubs
Mixed Service Areas
Medical Facilities
Office Parks
Retailers - Non-Food Related
Manufacturing: Food
Schools
Retailers - Food Related
Federal Parks
Amusement Parks
Mobile Home Parks
Day Care Centers
Manufacturing: Misc.
Non-water utilities
Total
% Daily
Consumption
Relative to CWS
A
0.3
0.25
0.5
0.5
1
0.5
0.5
0.25
0.5
0.25
1
0.5
0.25
0.5
0.25
0.5
0.5
0.25
0.5
1
1
0.5
0.5
0.5
N/A
Total Population
Among All
TNCWSs Within
Type
B
2,255,959
1,301,552
842,518
791,429
765,742
639,160
558,443
516,369
337,152
326,644
254,016
214,345
208,623
197,600
184,128
158,301
150,365
142,988
93,665
88,038
66,797
10,213
8,991
6,025
10,119,063
Daily
Consumption *
Total Poulation
C = A*B
676,788
325,388
421,259
395,715
765,742
319,580
279,222
129,092
168,576
81,661
254,016
107,173
52,156
98,800
46,032
79,151
75,183
35,747
46,833
88,038
66,797
5,107
4,496
3,013
4,525,560
Weighted Daily
Consumption
C / Total Population
0.0669
0.0322
0.0416
0.0391
0.0757
0.0316
0.0276
0.0128
0.0167
0.0081
0.0251
0.0106
0.0052
0.0098
0.0045
0.0078
0.0074
0.0035
0.0046
0.0087
0.0066
0.0005
0.0004
0.0003
0.4
Sources: [A] Geometries and Characteristics of Public Water Systems (USEPA, 2000a).
[B] EPA Estimate.
In addition to daily consumption, EPA estimated the number of days per year tap water is
consumed by users of different types of water systems. EPA's estimates for exposure days are presented
in Exhibit 5.7. For the TNCWSs, it is assumed that the average number of days of exposure per year for
each individual consuming water at one of these systems is 10. To account for all of the individuals who
are expected to be consume water at a TNCWS over the course of a year, an adjustment was made to the
population reported in SDWIS for these systems. The population for TNCWS reported in SDWIS is
intended to reflect the population served during the peak month of operation. Assuming 30 days in the
peak month, and an average of 10 days of exposure for each individual at a TNCWS, the total number of
different individuals exposed per month at each TNCWS is assumed to be 3 times the reported SDWIS
population number (that is, the peak month serves 3 cohorts of 10 days each). Further, it is assumed that
TNCWSs are open on average of 6 months per year, with an uncertainty range of 3 months to 9 months.
(This number of months of operation for TNCWS is included in the risk model as a triangular distribution
with minimum = 3, mode = 6 and maximum = 9 months.) Therefore, for TNCWSs, the risk model
assumes 10 days of exposure for each individual and a total population exposed ranging from 9 to 27 times
Economic Analysis for the
Final Ground Water Rule
5-16
October 2006
-------
the peak month population served reported in SDWIS (i.e., 3 cohorts / month * 3 months (=9) to 3 cohorts/
month * 9 months (=27). Since the SDWIS population values reflect the peak month of operation, the
total population served by TNCWS for their duration of operation as determined here should be considered
a likely overestimate.
Exhibit 5.7 Estimated Exposure Days by System Type1
Type of System
cws
NTNCWS
TNCWS
Exposure Days Per Year
350
250
10
1 Number of days in which tap water is consumed per individual in each system
type. See text for additional discussion pertaining to the TNCWS exposure.
5.2.4 Probability of Infection, Illness, and Mortality
This section presents information on the relationship between ingestion of viruses and the
probability of infection, illness, and mortality. Specific elements of the dose response relationships
addressed in this section include the following:
• Infectivity (the ability of a microorganism to colonize in the body of the host, expressed as
probability of infection, a function of dose)
• Morbidity (the probability of illness given infection)
Secondary Spread (the expected additional illnesses due to contact with those affected
directly by ground water consumption)
• Mortality (the probability of death given illness)
5.2.4.1 Infectivity from Dose response Modeling of Human Challenge Study Data
All of the elements listed above depend on high quality infectivity data and an appropriate
dose-response model. Two key publications report high quality human challenge study data suitable for
dose-response modeling. One, by Ward et al (1986), reports dose-response data for rotavirus, which
represents Type A viruses. The other, by Schiff et al (1984), reports dose-response data for echovirus type
12, which represents Type B viruses. The summary data from the two studies, shown in Exhibits 5.8 and
5.9, are simple, but their use in dose-response modeling involves a number of key decisions or
assumptions that are discussed below.
Microbial risk assessment requires an estimate of the dose required to cause infection in humans.
These data can be prospectively determined directly by dosing volunteers, indirectly by dosing animals, or
can be retrospectively estimated using outbreak data. The highest quality data are acquired from dosing
human volunteers. However, human volunteer data are necessarily biased for ethical reasons. All studies
Economic Analysis for the 5-17 October 2006
Final Ground Water Rule
-------
with human participants must be approved by an ethical review board. Limitations in pathogen human
feeding studies include using only low virulence pathogens (not necessarily representative of the virulence
encountered in nature) and studying only healthy populations (whose response to the experiment is not
indicative of that of sensitive subpopulations). For example, the echovirus type 12 strain used in the
human challenge study by Schiff (1984) was a benign strain not associated with serious human illness. In
fact the strain was originally recovered from a child with a clinical diagnosis of erythema infectiosum
(fifth disease), a mild facial rash.
Human challenge study data are also available for poliovirus (and for aerosol exposure to
coxsackievirus [Couch, 1970]), which together with echovirus, comprises the enteroviruses which are one
of the target viruses for determining GWR benefits. Three poliovirus challenges studies were conducted,
all using infants. Two studies used poliovirus I (Minor et al., 1981; Lepow et al, 1962) and one study
used poliovirus III (Katz and Plotkin, 1967).
EPA considered using data from all four enterovirus human challenge studies to represent the
Type B enterovirus dose response. (EPA did not think it appropriate to use the coxsackievirus dose
response data because of the route of exposure.) All the poliovirus challenge study data are from infants
who, unlike adults exposed to Echovirus type 12, exhibit lifetime immunity after infection. Even among
identical strains, there are potential virulence differences between laboratory-maintained isolates and
isolates recovered in human or environmental samples. Because poliovirus is no longer a health issue in
the U.S. and because of differences among the challenge studies, EPA decided to use only the echovirus 12
data. However, EPA believes that a meta-analysis of all the enterovirus human challenge study data would
document that the enteroviruses are more infectious than is predicted by the echovirus 12 human challenge
study data. EPA believes that the echovirus 12 data substantially underestimates the infectivity of the
enteroviruses. The GWR EA would predict more enteroviral infections, illnesses and deaths if other
enterovirus human challenge study data were considered.
The rotavirus human challenge study by Ward et al (1986) is the only published Type A human
challenge study data. The rotavirus challenge study was conducted using the CJN strain which was
recovered from an infected infant. The CJN strain is now classified as a Group A rotavirus and is among
the most common strains currently identified in the United States. The proportion of adult volunteers
infected by the CJN rotavirus strain at a dose of 1000 FFU in Ward et al (1986) was confirmed in a later
challenge study (13 of 14 adults infected in Ward et al, 1991). Thus, unlike the echovirus data, EPA
believes that the rotavirus human challenge study data are representative of the rotaviruses. However, the
infectivities of other Type A viruses, such as norovirus, are not necessarily represented by the rotavirus
human challenge study data. Funded under an EPA cooperative agreement, Moe et al. (2001) reported on
human challenge studies using Norwalk virus (now known as norovirus). They conclude that norovirus "is
one of the most infectious agents that has ever been described." Infection was observed at doses below
one PCR dectectable unit (which represents as few as one physical virus particle) (Moe et al, 2001).
Although the norovirus human challenge study data are not yet published, EPA believes that the
noroviruses are at least as infectious as the rotavirus, and likely more infectious. If the norovirus data were
made available to EPA, EPA would use the norovirus rather than rotavirus human challenge study data to
characterize Type A viruses, because norovirus disease affects all populations equally. Epidemiological
data suggests that other Type A viruses such as hepatitis A virus, adenovirus and astrovirus are also highly
infectious and could be more infectious than the rotaviruses. Thus, EPA believes that the rotavirus human
challenge study data, while representative of the rotaviruses, may underestimate the infectivity of the Type
A viruses. If so, then the GWR EA under-predicts Type A infections, illnesses and deaths.
Economic Analysis for the 5-18 October 2006
Final Ground Water Rule
-------
Exhibit 5.8 Rotavirus Dose-Response Data
Dose (Poisson),
focus-forming units
(ffu)
0.009
0.09
0.9
9
90
900
9,000
90,000
Subjects
receiving
Dose
7
7
7
11
7
8
7
3
Subjects
Infected
0
0
1
8
6
7
5
3
Exhibit 5.9 Echovirus Dose-Response Data
Dose (Poisson),
plaque-forming
units (pfu)
330
1000
3300
10000
33000
330000
Subjects
receiving
Dose
50
20
26
12
4
3
Subjects
Infected
15
9
19
12
2
2
Rotavirus Dose-Response Modeling
The human feeding study data shown in Exhibit 5.8 cover a wide dose range (0.009 to 90,000
units). Ground water concentrations are expected to be similar to those found for echovirus, that is,
generally less than 2 per 100 liters. If contamination is intermittent, such that it is present about 1% of the
time, then the average concentration would be 0.02 per 100 liters or 0.0002 per liter. A person ingesting
one liter per day would receive a dose of 0.0002 infectious units, which is considerably smaller than the
smallest dose used in the human feeding study. Predicting risks for such low doses requires extrapolation,
using a biologically-plausible dose-response model.
The World Health Organization (WHO, 2003) has considered the issue of low-dose modeling /
extrapolation, and recommends using models that have the following characteristics: absence of a
threshold and independent action among infectious units, i.e., absence of synergistic action. Based on
these factors, WHO (2003) identified a family of acceptable models and these included the Beta-Poisson
and exponential models, while excluding the Probit model.
Economic Analysis for the
Final Ground Water Rule
5-19
October 2006
-------
Ward, et al. used data collected at the dose range 0.009 to 90,000 ffu to estimate parameters of a
probit model. Other authors, notably Teunis and Havelaar, use these same data, but with a beta-Poisson
model (Teunis et al. 1996). These two models (probit and beta-Poisson) predict similar infection
probabilities across the range of doses employed in the study. As discussed above, EPA does not believe
that the probit model's predictions outside of the dose range of the study are biologically plausible. For
this reason, EPA has selected the beta-Poisson model for rotavirus dose-response modeling. In the main
analysis (reported in this section), all of the data shown in Exhibit 5.8 are used to inform the beta-Poisson
model.
Arguably, the most significant dose in terms of environmental exposure is that nearest 1 ffu: 0.9.
One of seven subjects was infected at that dose. Given the beta-Poisson model, and data at all of the
higher doses, this outcome at dose 0.9 ffu is not surprising, i.e. at order of magnitude or greater doses most
subjects were infected and at order of magnitude or lower doses no subjects were infected. However,
since most exposures to populations receiving contaminated ground water occur near one viral infectious
unit, there is a question as to the usefulness of other dose response data being used well outside this range
for estimating infection probability at this dose level. The smallest dose, 0.009 ffu, is smaller than 0.9 ffu
by a factor of 100. The greatest dose, 90,000 ffu, is greater by a factor of 100,000. To address this
concern about the influence of extreme dose data, EPA has estimated the parameter of an exponential
model using only the fact that one of seven subjects was infected at dose 0.9. This alternative or
sensitivity analysis is described in Appendix F.
EPA recognizes that there is significant uncertainty associated with measuring viruses in
environmental samples and has summarized that uncertainty in Chapter 4, Section 4.3.2. Virus
aggregation in routine environmental exposure is especially difficult to quantify, so there is considerable
uncertainty associated with the assumption above that most ground water consumers are getting a virus
dose of about one ffu. Virus occurrence and concentration is measured in the laboratory after the sample is
manipulated to disaggregate virus clusters by changing the ground water ionic strength. After
disaggregation, the highest measured virus concentration in any ground water sample (Exhibit 4.28) is 2.12
PFU (or MPN) per liter. Clusters may exist in routine ground water exposure (in aquifers where ionic
strength is relatively stable) and ground water consumers may get doses higher than one FFU on occasion
as a result.
Echovirus Dose-Response Modeling
The echo virus data shown in Exhibit 5.9 cover a narrower relative range than rotavirus (from 330
to 330,000 pfu), but all doses are significantly greater than the environmentally-relevant dose of 1 pfu.
Using any dose-response models with these data to predict environmental risks involves extrapolation and
the potential to over- or underestimate the low-dose risk.
Before discussing the modeling, it is important to understand how the data in Exhibit 5.9 were
derived. Schiff et al. described how their data were generated in two phases. In the first, a "range-finding"
phase, small numbers of subjects were dosed at each level over a wide range. Three or four subjects were
dosed at each of some number of levels between 330 and 330,000 pfu. Exhibit 5.9 only shows that four
subjects received dose 33,000 and three subjects received dose 330,000. In the second phase, numerous
additional subjects were challenged with doses in the 330 to 10,000 pfu range (for these doses, Exhibit 5.9
includes subjects from both phases).
Economic Analysis for the 5-20 October 2006
Final Ground Water Rule
-------
EPA considered two published echovirus dose-response analyses: that of the original researchers,
Schiff et al. (1984), and that of Teunis and Havelaar (Teunis et al., 1996). In their dose response analysis,
Schiff et al. discarded the data from doses 33,000 and 330,000 and used the remaining data to estimate
parameters of a probit model. Teunis and Havelaar utilized all of the echovirus dose-response data
(including 33,000 and 330,000) and fit a beta-Poisson model. After assessing the published analyses, EPA
decided to use the full data set with the beta-Poisson model in the main analysis and the partial data set
with the exponential model as a sensitivity analysis.
EPA based its analysis of the full data set on the analysis of Teunis and Havelaar (Teunis et al.,
1996). An important feature of this model is its ability to explain variable susceptibility; the notion that
human subjects may differ. Technically, this model acts as though each subject has his/her own
exponential dose-response parameter (probability of infection, given exactly one infectious unit) and that
these parameters are beta-distributed. EPA decided to employ the beta-Poisson model in its analysis of the
full data set.
For the analysis of the low-dose echovirus data, EPA examined the Schiff analysis (Schiff et al.,
1984). Although Schiff used a probit model, EPA discovered it behaves implausibly when extrapolating to
estimate the risk of environmental exposures. Theoretically, risk at low doses should be approximately
proportional to probability of ingesting an infectious unit, but in this application, the probit model predicts
extreme sublinear behavior. In contrast, the simple exponential and beta-Poisson models (which is a
derivation of the simple exponential) both feature approximate low dose linearity. When applied to the
data below dose 33,000, the exponential model fits very well. The extra parameter provided in the
beta-Poisson model doesn't seem to add significantly to the goodness-of-fit (or likelihood), so EPA
decided to employ the exponential model in its analysis of the low-dose echovirus data.
Two reasonable options are therefore available for echovirus modeling: utilizing the partial data
set with an exponential model and utilizing the full data set with the beta-Poisson model. As detailed in
Appendix F, these two analyses produce considerably different predictions of low-dose risk for echovirus.
Taken together, the results demonstrate considerable model uncertainty.
5.2.4.2 Morbidity and Mortality Data Sources and Uncertainty
EPA based its analysis of morbidity and mortality on epidemiologic data published in the scientific
literature. The literature includes epidemic (outbreak) and endemic disease studies conducted using a
variety of epidemiological methods, both prospective and retrospective. Some study cohorts represented
the entire U.S. population, other cohorts represented selected sub-populations. The studies used in the EA
include data based on person-to-person spread, foodborne exposures, and waterborne exposures, as well as
those for which no source of viral exposure or transmission mode was identified. Due to the limited
number of published studies on water as a primary route of exposure, viral morbidity and mortality rates
incorporate information derived from all routes of exposure and transmission. EPA used to the extent
possible, prospective, endemic data of study cohorts that represent the entire U.S. population. Prospective
studies are assumed to have less reporting bias than retrospective studies (only those who got ill are likely
to have a reason to report their exposure and the likelihood of reporting increases with illness severity).
Where such data were not available, EPA used retrospective endemic disease studies of the entire
population. At times, general population study data were not available and so EPA used prospective (and
retrospective, as necessary) studies of sub-populations and applied these values to the general population,
but noted the associated uncertainty or bias. This EA describes the data source, the type of study, the
Economic Analysis for the 5-21 October 2006
Final Ground Water Rule
-------
population cohort studied and the assumptions and limitations associated with epidemiological data that, at
times, were not optimal.
Morbidity rates are based, to the extent possible, on prospective epidemiological studies of
endemic illness in the general population. Morbidity data were specifically excluded if they were based
on; 1) retrospective outbreak studies, 2) retrospective or prospective sensitive sub-population studies
(limited prospective age related sub-population studies were included), 3) retrospective passive disease
reporting studies or 4) predictions from infectious disease simulation of illness acquisition and spread. The
majority of studies cited in the EA were conducted in the U.S. or Canada. Where U.S. (and Canadian) data
were identified as limited, incomplete or biased, data from other developed countries were used to inform
data ranges for sub-populations.
EPA excluded the above four categories of morbidity data because of an assumption that including
these data would bias the illness predictions upward. However, there are some uncertainties associated
with that EPA assumption. The following text addresses these uncertainties about the estimated morbidity
rates for each of the four exclusions listed.
Retrospective outbreak studies have an inherent apparent disease burden bias. That is, common
source outbreaks come to the attention of public health authorities and are more likely to be reported to
CDC if the outbreak is large or causes illness that is relatively severe, compared with milder disease
endpoints such as rash or muscle stiffness. (Other factors, not related to disease burden, such as
occurrence in a locality with more effective public health surveillance and sufficient laboratory capacity
also affect outbreak reporting.) EPA excluded retrospective outbreak studies because of a concern that
outbreak data may not be reflective of endemic illness rates. However, in doing so EPA may be biasing
the estimated viral morbidity rates downward because the more virulent etiologic agents more likely
associated with outbreaks may be excluded from the analysis.
For the more highly infectious viruses, such as rotavirus and norovirus (Type A viruses), there
may be little difference between small epidemics and endemic illness. Unlike bacteria, waterborne viruses
do not have known animal reservoirs so the agent is continually circulating within the human population
within the U.S. (albeit supplemented by travelers coming or returning to the U.S.). A small outbreak
(CDC defines an outbreak as two or more contemporaneous illnesses), such as one occurring within a
family, may be counted as either endemic or epidemic disease. For especially infectious agents, epidemic
rather than endemic cases might be observed. Thus, there is uncertainty as to how to separate endemic
versus epidemic disease for the Type A viruses because they are highly infectious, continually circulating
within the human population, cause only mild illness and may be counted as either endemic or epidemic
disease, depending on the definitions used in the compilation.
EPA excluded sub-population morbidity data (with the exception of some age related data
discussed later) because of a concern that it may not reflect endemic illness in the general U.S. population.
Studies of sub-populations in the U.S. report differing likelihood of becoming ill or severely ill. As
discussed in Section 5.2.2. sensitive subpopulations comprise a substantial proportion of U.S. population,
as much as 20% in one estimate (Gerba, 1996). EPA recognizes that the EA underestimates the number of
illnesses and deaths because greater morbidity and mortality rates among some sensitive sub-populations,
such as pregnant women (e.g. in developing countries, twenty percent of pregnant women have a fatal
outcome due to Hepatitis E viral illness) are not recognized in the main analysis. In addition, some
sensitive sub-populations have greater morbidity rates than the general population. Because the EA does
Economic Analysis for the 5-22 October 2006
Final Ground Water Rule
-------
not consider data on a large portion of the population that generally is more highly affected by rotavirus
and enterovirus, it underestimates the morbidity and mortality rates for rotavirus and enterovirus.
With respect to age related sub-populations, EPA captured as much of the age related difference as
possible but could not do so comprehensively. Age is important because sub-populations divided by age
have significantly differing rates of morbidity and mortality. All waterborne viruses, and especially
rotavirus, are more likely to affect children under five than other age groups. As discussed by Ramig
(2004) the reason that children are more likely to become ill from rotavirus is due to several possibilites.
Ramig lists three reasons that apply to children in the U.S. (a fourth reason is malnutrition); one reason is
that children are less likely to be previously exposed and therefore have no pre-existing partial immunity.
The other two reasons have to do with role of key enzymes with prevalence that changes with age. The
GWR accounted for some age heterogeneity among the general population by, for example, using differing
rotavirus morbidity rates for children under two and determining echovirus case-fatality ratios among
neonates. However, EPA recognizes that some age related sub-populations, such as the elderly, that may
have differing morbidity and/or mortality rates are not explicitly characterized.in the EA and that absence
contributes to uncertainty.
EPA explicitly excluded enteroviral disease data reported to CDC for three reasons. First, the data
are not current because passive enteroviral etiology reporting was discontinued in 1994 (Currently,
hepatitis A virus is the only potentially waterborne virus reported to CDC.). Second, as recognized by
CDC (Mead et al., 1999), disease reporting is designed and administered so that reporting is done
voluntarily by State and local public health agencies and laboratories. Third, an enteroviral etiology comes
to the attention of State and local public health authorities primarily when significant illness and/or
significant frequency occurs. Because many enteroviral illnesses do not result in symptoms sufficiently
grave so as to require a physician's care, the passively reported illnesses associated with an enterovirus
etiology are assumed to be biased toward the more severe outcomes.
EPA explicitly excluded some rotavirus and enteroviral illnesses due to disease secondary
transmission within a community. EPA identifies primary illnesses as those that are acquired by
consuming rotavirus and enterovirus from a contaminated PWS well. Secondary illnesses are those cases
that are acquired by individuals who do not consume the contaminated drinking water, but instead acquire
illness through person-to-person exposure from a primary case. The EA does include a secondary disease
transmission factor to account for the number of secondary illnesses that are estimated to arise from each
primary illness. However, as is discussed in Section 5.2.4.3, this secondary disease transmission factor is
not a secondary morbidity rate because it assumes that only ill individuals can make others ill by
secondary transmission. As is well known from the case of Typhoid Mary, asymptomatic individuals can
also serve as carriers of disease, and thus infected individuals can infect others who can also become ill or
who can become carriers also. As a result of neglecting secondary cases that arise from asymptomatic
carriers as opposed to symptomatic carriers, the EA underestimates the total number of secondary cases.
In addition to estimating morbidity rates, EPA has estimated illness severity. Severity is important
because monetization of the rotavirus and enterovirus disease burden bestows greater benefits (avoided
costs), such as hospitalization costs and physicians's office visit costs, if the disease is more severe. In this
EA, fatal outcomes occur for the general population (non-neonates) ill from enterovirus only if disease is
severe. Thus, to estimate mortality rates, the number of severe cases must be determined.
EPA used total disease burden estimates from CDC to account for the relative proportion of severe
enterovirus (Type B) outcomes. EPA defines severe outcome as any illness resulting in hospitalization.
Economic Analysis for the 5-23 October 2006
Final Ground Water Rule
-------
CDC estimates that about 10 million cases of enteroviral disease occur each year in the U.S.. EPA
estimated that about 100,000 severe outcomes result from illness (myocarditis, encephalitis, and viral
meningitis) potentially caused by enteroviruses. The ratio of 100,000 severe outcomes divided by 10
million cases was used in the EA to estimate that 1% of enteroviral cases are severe. There are
uncertainties in using this 1% severity factor for the enteroviruses. In determining enteroviral severity,
EPA believes that a large proportion of the 10 million enteroviral cases in the U.S. could be due to hand,
foot and mouth disease (e.g. Pichichero et al. (1998) found in a four month study that 244 of 372
physician's office visits were due to symptomatic (stomatitis) or diagnosed hand, foot and mouth disease)
which is a benign childhood disease (in the United States but not necessarily elsewhere such as Taiwan)
caused by enterovirus 70 and 71.4 Enterovirus 70 or 71 has never been found in drinking water or ground
water and is not believed to be waterborne in the U.S.. Rather, it is spread almost exclusively by person to
person transmission. Because enterovirus 70 and 71 illness is included in the total cases of enteroviral
disease, this EA could significantly underestimate the proportion of severe outcomes due to enterovirus in
the general population. However, in years or seasons for which no data are available, depending on which
enterovirus strains predominate, clinical symptoms not characteristic of enterovirus 70 or 71 could be more
likely observed and could account for more or most of the enteroviral illness.
As compared with the expansive definition of Type B (enteroviral) disease severity described
above, Type A disease severity is restricted to consideration only of the rotaviruses. For the Type A
viruses, only rotavirus is included in the estimate of the total disease burden despite the fact that norovirus
disease is widespread and norovirus makes both adults and children ill at significant rates. Other Type A
viruses, such as adenovirus and hepatitis A virus also disproportionately cause illness in younger adults.
As a result of restricting the Type A disease burden to rotavirus only, the EA significantly underestimates
the proportion of severe Type A disease in the general (adult) population.
Mortality from ground waterborne agents, especially viruses, is sufficiently infrequent that a
mortality rate for the general population is difficult to determine from the scientific literature. This
difficulty is compounded for Type A virus because the available data are limited to rotavirus which is
highly infectious but has relatively mild severity. Mortality data are available only for sub-populations at
greatest risk of a fatal outcome. For enterovirus, data are available from a study of neonates with
enteroviral infection supplemented with fatal outcome case data from several other studies and applied to
the neonate population. For rotavirus, a fatal outcome is based on deaths in children under five identified
by CDC as part of a study to assess the costs and benefits of a rotavirus vaccine (subsequently removed
4 Enteroviral type 71 disease was particularly virulent in Taiwan in 1998 where there was over 129,000
cases and 78 deaths, 90% among children less than 5 years old (Lin et al, 2003; Chang et al, 2004). In severe cases,
hand, foot and mouth disease proceeds to encephalomyelitis and cardiopulmonary collapse. The widespread
transmission is enhanced by the long period of viral shedding (as much as 5 weeks) in infected individuals.
Infection rates for parents (41%) were higher than other adults (26%). These data suggest that enterovirus 71 is
efficiently transmitted by asymptomatic or mildly symptomatic adults (Chang et al, 2004). It is likely that most
hand, foot and mouth disease in the United States is also due to contact with other individuals. However, enterovirus
71 is copiously shed via the gut and can, along with other fecal/oral agents, be transmitted via ground water,
although the available data seem to indicate that this pathway is not likely. Three possible reasons that enterovirus
71 is not likely to be transmitted via ground water are that it may not survive long at ground water temperatures, it
may not be mobile in the subsurface and/or the infectious dose required to cause illness is so large that dilute
environmental samples cannot transmit infection. However, as the data from Taiwan indicate, the disease can cause
fatal outcomes at rates greater than the rates used in this EA for the enteroviruses, albeit perhaps not in developed
countries.
Economic Analysis for the 5-24 October 2006
Final Ground Water Rule
-------
from the market due to unintended intestinal obstruction effects). Thus, the available mortality data are
limited to data from children. No U.S. endemic mortality data are available for fatal outcomes in the
sensitive subpopulations, such as the elderly, a population also greatly susceptible to dehydration-related
mortality due to gastroenteritis.
As discussed in Section 5.2.4.3, only rotavirus mortality rates in children are applied to the entire
population. The neonate mortality rate for enterovirus is restricted to neonates and not used to inform
mortality rates for other population sub-groups. EPA believes that, for the enteroviruses, neonate mortality
rates are higher than for the general population and thus the neonate data cannot be applied more broadly.
On the other hand, EPA has no data on rotavirus mortality in population subgroups other than children
under 5 years in age, so EPA cannot determine if the rotavirus mortality rate in children is high or low
compared to other subgroups. EPA applies the rotavirus mortality to all age groups, recognizing that this
is a source of uncertainty (it could be an over or underestimate). The enterovirus mortality rates for the
non-neonates is based on a simple calculation that estimates the number of severe (hospitalized)
enterovirus cases together with an estimate of the number of fatal outcomes in severe cases. This mortality
rate is sensitive to the proportion of severe cases as discussed above and could be an underestimate if the
number of severe cases is underestimated.
Another source of uncertainty in the EA estimates of morbidity and mortality rates for rotavirus
(Type A virus) is whether or to what degree these rates could change due to a rotavirus vaccine. In 2006,
the U.S. Food and Drug Administration (FDA) approved a new pentavalent vaccine for rotavirus (Vesikari
et al, 2006) in children. The vaccine protects against the most common rotaviruses now found in the U.S.
(serotypes Gl, G2, G3, G4, G9) but does not protect again all rotaviruses found in the U.S. or all
rotaviruses. Most current rotavirus infections in the U.S. are Gl serotype so Vesikari et al (2006) were
able to demonstrate that the vaccine was 74.9% (95% CI: 67.3-80.9) efficient in protecting against Gl
gastroenteritis. Because few rotavirus cases occurred in the U.S. for the other serotypes during the study,
Vesikari et al (2006) report protection efficiency with very large uncertainty [G2 - 63.4% (95% CI: 2.6-
88.2); G3 - 82.7% (95% CI: <0-99.6%); G4 - 48.1 (95% CI: <0-91.6%); G9 - 65.4 (95% CI: (<0-99.3%)].
These data suggest that after a prolonged vaccination program, substantial rotavirus gastroenteritis will
continue to occur in the U.S., most from Gl serotype, but some from G2, G3, G4 and G9 serotype in
vaccinated children as well as from other rotavirus serotypes and additionally, in unvaccinated children
and adults. The most common rotavirus strains found in the U.S. could change in the future, either
naturally or as an unintended result of vaccination. EPA believes that norovirus, rather than rotavirus is
the more important health concern in the U.S. as a result of Type A viral exposure from ground water.
EPA believes that, successfully implemented, the rotavirus vaccine has the potential to affect the future
number of health effects in children. Within the first 25 years of GWR implementation, all children that
are vaccinated would have reduced likelihood of incurring illness from rotavirus. Similarly, rotavirus
secondary spread, which in this EA is assume to originate only from children less than five years in age,
would be significantly reduced.
Most importantly, a rotavirus vaccine will have no impact on norovirus illnesses or other Type A
viral illnesses resulting from ground water exposure. Type A virus morbidity and mortality rates are
estimated with data on rotavirus because only rotavirus human challenge study data are available to
determine the likelihood of infection given a virus dose. No such infectivity data are currently available
for other Type A viruses (see Section 5.4.1 for a discussion of other Type A data including unpublished
norovirus infectivity data). EPA believes that the total Type A disease burden, especially in adults, is not
captured by monetizing rotavirus illness, with or without consideration of a rotavirus vaccine.
Economic Analysis for the 5-25 October 2006
Final Ground Water Rule
-------
5.2.4.3 Quantified GWR Benefits (Predictions of Illnesses and Deaths)-Model Input Values
The quantified GWR benefits are determined for rotavirus and enterovirus using dose response
models fitted to human challenge study data, morbidity, mortality and secondary illness factor when
ingested from ground water sources. These data are summarized in Exhibit 5.10 and described in more
detail below.
Exhibit 5.10 Dose Response Assumptions of Viral Pathogens for the
GWR Risk Assessment (continued on next page)
Pathogen
Hazards
Definition
Infectivity1
Infectivity measures the
probability of infection for
exposure to a specified
dose. "Annual infectivity
risk" estimates risk for a
year for exposure to a
daily dose, N, for a given
number of days, D, during
the year (see text for
details.) Dose response
function for annual risk is:
Annual \ Daily /
(see below for estimating
the probability of daily
infection PDailv)
Morbidity
MP
Primary morbidity is
the probability of
illness given
infection; it can vary
from person to
person and can be
greater in sensitive
subgroups.
Ms
Secondary
morbidity is the
probability of
illness given
exposure to an
asymptomatic or
symptomatic ill
person.
Mortality
(MJ
Mortality is the
probability of
death as a
result of
illness.
Economic Analysis for the
Final Ground Water Rule
5-26
October 2006
-------
Pathogen
Hazards
Type A
viruses
(Highly
infective
virus)
Represente
d by data on
Rotavirus
Type B
viruses
(Moderately
infective
virus)
Represente
d by data on
Echovirus or
Enterovirus
Infectivity1
Dose response function
for daily risk of infection is
given by the expected
value of the Beta Poisson
dose response function:
p _ A^f a I
lDa'!y "'\a + p)
PDaNy for a single exposure
to one Type A infectious
unit is estimated to be
22.4% with 90%
confidence bounds of
8.8% and 41 .4% reflecting
uncertainty in the • *and • •
parameters.
Dose response function
for daily risk of infection is
given by the Pareto
approximation to the Beta
Poisson dose response
function:
[x -a
' + -]
PI
PDaNy for a single exposure
to one Type B infectious
unit is estimated to be
0.44% with 90%
confidence bounds of
0.06% and 1 .65%
reflecting uncertainty in
the • «and • "parameters.
Morbidity
MD
< 3 yrs2 = 0.10-
0.88, uniform
distribution
• »3yrs3 = 0.10-
0.50, uniform
distribution
< 5 yrs 5 = 0.5 -
0.78, uniform
distribution
••5-19 years 5 =
0.12- 0.57, uniform
distribution
> 19 years5 = 0.12 -
0.33, uniform
distribution
Ms
< 3 yrs2a = 0.55
(applied to the
entire exposed
population)
• »3 yrs = 0
(assumed)
Triangular
distribution (all
age groups)6,
from 0.11 to 0.55;
mode = 0.35
Mortality
(Mm)
All ages 4 =
5.7 x10-6-
7.3 x10-6,
uniform
distribution
< 1 month 7=
9.2 x10'3
• »1 month 8 =
.02
(applied only
to illnesses
that require
hospitalization
, which are
1% of all
illnesses)
1 See Appendix F for details of infectivity dose response functions.
2 Based on Rodriguez et al. (1987) for the upper range value and Perez-Schael (1984) for the lower range value.
2a Based on Kim et al. (1977)
3The range for older children and adults based on Ward et al. (1986) (upper range value), Kim et al. (1977) (lower
range value), and Wenman et al. (1979)
4 From Tucker et al. (1998) and Kapikian (2001) for children less than 5 years old and assumed to apply to all
ages.
5 Hall et al. (1970); Kogon et al. (1969)
6Morens et al., 1978 [citing Karzon et al (1961), Winkelstein et al (1957) and Lehan et al (1957)]
7From Jenista et al. (1984), Modlin (1986) and Kaplan and Klein (1983)
8Based on data from Melnick (1996), Modlin (1995) and Bennet et al. (1987).
Notes: Mortality rates are rates of mortality given illness.
Economic Analysis for the
Final Ground Water Rule
5-27
October 2006
-------
Morbidity
In the EA, Type A viruses are represented by data on rotavirus because published data from
human challenge studies to determine the probability of infection given a virus dose are available only for
rotavirus (see Section 5.4.1 for a discussion on other Type A viruses, including unpublished human
challenge study data for norovirus which are not used in this analysis). For rotavirus, separate morbidity
factors were used for children less than 3 years old and for individuals greater than or equal to 3 years old.
For children less than 3 years old, a morbidity factor was taken as a uniform distribution from 0.10 to 0.88
(Perez-Schael et al, 1984; Rodriguez et al, 1987). The lower range value was reported by Perez-Schael et
al (1984) and the upper range value was reported by Rodriquez et al (1987). There are significant
differences in the two study designs that may account for the great variability in the reported morbidity
rate. Rodriguez studied children under 3 while Perez-Schael studied newborns. Newborns are protected
against illness by maternal antibodies which fade by about 6 months in age. Older children can report a
wide range of clinical symptoms (e.g. pharyngitis and otitis media) in contrast to newborns who cannot
report symptoms. Thus Rodriquez used multiple disease endpoints while Perez-Schael was able to use
only diarrhea as a clinical symptom of disease. The Rodrizquez study was conducted in a private pediatric
medical practice in Virginia USA while the Perez-Schael study was conducted in a large maternity hospital
in Caracas, Venzuela. Because of the differences in the study designs used to determine rotavirus
morbidities, there is significant uncertainty in the rotavirus morbidity range.
For people aged 3 years or more, this EA uses a uniform distribution from 0.10 to 0.50. The lower
bound of 0.10 is from Kim et al (1977). Kim et al. (1977) found that 3 parents of infant children
hospitalized with rotavirus illness were ill (with gastroenteritis) among the 26 parents exposed to those ill
children and infected with rotavirus. The upper bound of 0.50 is from Ward et al. (1986). Ward et al.
(1986) found a 50% diarrhea (other ill individuals exhibited differing disease endpoints) morbidity rate in
adults challenged with a rotavirus dose under experimental conditions. Wenman et al. (1979) reported that
17 parents became ill of 43 parents infected by rotavirus (40%) in a prospective study of diarrhea in
households with newborn children in Canada. As discussed in Section 5.2.4.2, the Gl serotype is most
common in the U.S. and most adults have had one or more exposures to this serotype. Rotaviral immunity
to reinfection and illness is short (as opposed to lifetime immunity characteristic of poliovirus and hepatitis
A virus) and multiple infections are possible (Anderson and Weber, 2004; Koopman and Monto, 1989).
Griffin et al. (2002) analyzed rotavirus outbreaks and identified one rotavirus serotype (G2) that is
associated with outbreaks in adults. Outbreak data from nursing homes in Australia (Marshall et al, 2003)
identified Gl, G4 and G9 rotavirus strains as causing outbreaks among adults. Because multiple serotypes
can produce rotavirus disease in adults, these data suggest that the morbidity value for adults can differ
from the morbidity identified from a single strain. If the human subjects in Ward et al (1986) were
challenged by G2, G4 and G9 strains, in addition to Gl, then it is likely that a greater percentage would
become ill. There is uncertainty in the EA estimate of rotavirus morbidity for adults because it is based
primarily on infections produced by only one strain of rotavirus (Gl). It is possible that other rotavirus
strains, such as G2, and consideration of disease endpoints other than diarrhea, would provide higher
morbidity estimates.
Type B viruses are represented by data on echovirus or enterovirus. The morbidity data for Type
B viruses are taken directly from two virus watch studies in Seattle (Hall et al., 1970) and New York
(Kogon et al, 1969) which are prospective epidemiological investigations undertaken in a general
population cohort. As such, they are high quality (and costly) data.and represent the best prospective
enterovirus morbidity data available for the general population. To the extent that current molecular
Economic Analysis for the 5-28 October 2006
Final Ground Water Rule
-------
methods are more sensitive in determining infection by identifying viruses shed in stool, the cell culture
methods used in the virus watch studies may underestimate the number of infected individuals.
Exhibit 5.11 shows echovirus morbidity data from both the Seattle and New York Virus Watch
studies. Coxsackievirus (another subgroup of the Type B enteroviruses) are also reported from the New
York Virus Watch study and shown in Exhibit 5.11. However, only echovirus data were used in the EA to
represent the Type B viruses.
Three age-based morbidity rates ranges for echovirus (one group of the enteroviruses) from the
Seattle (Hall et al., 1970) and New York virus watch studies (Kogon et al., 1969) are used in the EA, each
in the form of a uniform distribution. The morbidity rate used for children less than 5 years of age is 0.5 -
0.78; for children 5 years to 19 years is 0.12 - 0.57; and for adults greater than 19 years old is 0.12 - 0.33.
For children less than five years old, the lower bound comes from the Seattle study and the upper bound
comes from the New York study. For children between ages 5 and 19, the lower bound comes from the
New York study and the upper bound comes from the Seattle study. For adults, the lower bound comes
from the New York study and the upper bound comes from the Seattle study. The coxsackievirus
morbidity data from the New York study were not used.
As shown in Exhibit 5.11, the Seattle virus watch reported data for children aged 5 to 19 years and
for individuals over 19 years; the New York virus watch data reported data only for individuals under and
over four years in age. Because the New York virus watch only reported data for the over four age group,
these data (0.12) were used to populate both the 5 to 19 year group and the greater than 19 year group. For
the Seattle virus watch study, a total of 43 individuals were infected by echovirus and 20 people became
ill. For the New York virus watch study, 53 individuals were infected by echovirus and 24 people became
ill. Thus, despite the high quality of the data, the total number of people used to determine morbidity is
small.
There is large uncertainty associated with the Type B morbidity values currently used in the EA
because they are based only on echovirus data. For other parameters in the GWR EA (e.g., disease
severity) data from all enteroviruses were used. A significant uncertainty is the exclusion of the
coxsackievirus morbidity data. If coxsackievirus (also an enterovirus) morbidity data were used from the
New York viral watch study (Kogon et al, 1969) then the mean morbidity rates would be significantly
higher because coxsackievirus is more likely to cause illness than echovirus, upon infection. Kogon et al
(1969) report 10 illnesses resulting from 15 infections. Thus, the moribidity range for the 5-19 year olds
would change from 0.12 - 0.57 to 0.12 - 0.67 and the morbidity range for adults would change from 0.12-
0.12 - 0.67. For all age groups in the New York Viral Watch study, Kogon reported 53 illnesses resulting
from 103 infections.
In this EA, only echovirus data were used to determine Type B virus morbidity rates (despite the
availability of coxsackievirus morbidity data) to ensure consistency with the infectivity data for Type B
viruses. The human challenge study data representative of Type B viruses was conducted using echovirus
(type 12). By using only echovirus morbidity data, both infection and illness rates are determined using
only echovirus data, thereby ensuring that the predicted number of illnesses is more closely represented by
the challenge doses.
Exhibit 5.11 presents a summary of the morbidity data from the Seattle and New York Virus
Watch studies discussed above.
Economic Analysis for the 5-29 October 2006
Final Ground Water Rule
-------
Exhibit 5.11 Summary of Virus Watch Morbidity Data
Seattle Virus Watch Study
(Hall etal, 1970)
Age
Bins
0-5 years
5-19 years
1 9+ years
Echovirus
Morbidity
12 ill/
24 infected
(50%)
4 ill/
7 infected
(57%)
4 ill/
12 infected
(33%)
Coxsackievirus
Morbidity
no data
no data
no data
New York Virus Watch Study
(Kogon etal, 1969)
Age
Bins
0-4 years
5-9 years
(Coxsackie-
virus)
5+ years
(Echovirus)
1 0+ years
(Coxsackie-
virus)
Echovirus
Morbidity
21 ill/
27 infected
(78%)
Sill/
26 infected
(12%)
see above
Coxsackievirus
Morbidity
28 ill/
54 infected
(52%)
15 ill/33 infected
(45%)
10 ill/1 5 infected
(67%)
The mean morbidity rate is based on the observed range and assumed distribution for both Type A
(rotavirus) and Type B (echovirus) illnesses. This value is likely underestimated because viral pathogens
other than the surrogate types will also be avoided by rule implementation and may include agents with
greater capability to cause illness, as is shown in the above example for Coxsackievirus, thereby increasing
the upper part of the range and the corresponding mean value. The potential benefits of reducing illness
from pathogens other than enterovirus are discussed in section 5.4 of this chapter. In particular, section 5.4
identifies norovirus as a common waterborne etiologic agent that, unlike rotavirus, causes illness in adults
as well as in children. Thus, in choosing a differing Type A surrogate virus, the mean morbidity rate in the
general population (adults and children) could increase, as compared with rotavirus that causes illness
primarily in children.
Secondary Spread
Secondary spread can occur from contact with an infected or ill individual. It is likely that most
secondary spread occurs from contact with asymptomatic carriers (e.g., Typhoid Mary) because, for most
age groups, there are no clinical symptoms expressed to keep others away. Because data on secondary
spread from asymptomatic carriers are not available, in this analysis, it is assumed that only ill persons are
capable of causing secondary transmission. Thus, the secondary spread value used is the EA is not a
secondary morbidity rate (which includes contact with both symptomatic and asymptomatic carriers) but
rather a secondary spread factor. This assumption underestimates the secondary spread morbidity rates
which ideally should be applied to the total number of infected rather than ill individuals.
Secondary spread of waterborne illnesses is a reasonable assumption because the pathogens of
concern for the GWR are also commonly transmitted by respiratory or direct contact (fecal-oral) pathways.
Aerosols containing fecal-oral viruses can be transmitted either by the ingestion or the inhalation route.
Economic Analysis for the
Final Ground Water Rule
5-30
October 2006
-------
Some waterborne-viruses, such as coxsackievirus and adenovirus cause eye infections (conjunctivitis) and
so they may be transmitted by the dermal pathway via primary or secondary exposure.
Because the enteroviruses and rotavirus do not have animal reservoirs, they circulate only within
the human population. For the viral strain to avoid dying out, each individual with a viral infection must,
on average, transmit at least one additional infection to maintain that virus strain within the human
population. Thus, high rates of secondary transmission are necessary for virus viability.
Most available secondary morbidity data are obtained from studies of households. In these
households, both parents and children become ill and secondary transmission values are based on data
and/or assumptions about whether parents or children were the primary case. Few data are available to
evaluate secondary transmission outside the household (within the community) or within other higher risk
settings such as nursing homes. For rotavirus, a virus watch study was used to collect data from
Tecumseh, Michigan (Koopman and Monto, 1989; Koopman et al, 1989) and an infectious disease
simulation was conducted to determine the secondary morbidity rates (Longini and Koopman, 1982). As
discussed in section 5.2.4.2, data derived from infectious disease simulations are specifically excluded
from consideration in the GWR EA.
For rotavirus (Type A), the secondary spread factor was determined separately for children under
3 and all others. For all individuals over age 3, it is assumed that there is no secondary spread of rotavirus.
This assumption is based on the apparent exposure of most individuals to multiple strains of rotavirus early
in life. As a result, it is assumed that most individuals have acquired some temporary immunity to
rotavirus by age 3 and so, while there is a likelihood of developing clinical illness from primary exposure,
it is expected that there would be a lower likelihood of developing clinical illness from secondary
exposure. Primary exposure morbidity data from Ward et al. (1986) and secondary morbidity data from
Wenman et al. (1979) (assuming adults are secondarily exposed in Wenman) show that the higher
rotavirus morbidity rates comes from primary exposure while the lower (but not zero) rates come from
exposures that might be secondary. The EA underestimates the number of illnesses and deaths in adults
from rotavirus by assuming no clinical illness in individuals over age three by secondary transmission.
For children under age 3, the rotaviral secondary illnesses factor is 0.55 (i.e. for every two primary
rotavirus illnesses among children under age 3, there is about one secondary illness in individuals of any
age that is included in the GWR EA benefits analysis). This value is based on Kim et al. (1977). Kim et
al. (1977) is a study of the adult contacts of pediatric patients with gastroenteritis. It was not designed to
determine secondary transmission values, either secondary morbidity rates or secondary illness factors.
For example, there are no data that definitively identify whether the adults or the children were the primary
case and either assumption is appropriate. EPA has identified at least two differing analyses for
determining secondary illness using the Kim paper. Fortunately, the two differing analyses (either may be
correct) give approximately similar results which are presented here (Exhibit 5.10).
Echovirus (Type B) secondary illness factor is a triangular distribution for individuals of all ages.
The range is 0.11 - 0.55 with a mode of 0.35. These data are derived from echovirus outbreak data
reported by Morens et al (1991) (Table 17-3) in a reference work on the enteroviruses. The lower value in
the range comes from a study of a New York aseptic meningitis outbreak in 1956 (Karzon et al, 1961).
The upper value in the range comes from a New York aseptic meningitis outbreak in 1955 (Winkelstein et
al, 1957). The mode is informed by the other aseptic meningitis outbreak (in Iowa) listed in the table
(Lehanetal, 1957).
Economic Analysis for the 5-31 October 2006
Final Ground Water Rule
-------
In general, primary sources are used in the GWR EA to determine input values for the risk
assessment. However, secondary transmission data are very difficult to acquire and evaluate because key
data describing which household members are primary and which are secondary cases are typically not
available. There are no data on secondary transmission of echovirus or other enteroviruses (Type B)
outside the household. For these reasons, the echovirus secondary illness factor used in the EA is taken
directly from a reference work where, it is presumed that the authors had adequate expertise to evaluate the
original paper and make an informed judgement as to whether to use these data to inform secondary spread
values.
The enterovirus secondary transmission factor is defined only by echovirus data. Because of low
infectivity rates, secondary transmission data are rare for the other enteroviruses. It is likely that, if other
enteroviruses, such as coxsackievirus, were considered, the secondary transmission values could differ.
For example, most enterovirus 70 or 71 is acquired by secondary transmission via person to person
contact. Thus, if secondary transmission data were available and used in this EA, the secondary spread
factor would likely be higher.
The use of observational epidemiologic data in the quantitative benefits analysis is an appropriate
and acceptable method to determine the number of secondary cases. However, it is not the only
appropriate and acceptable method. In Appendix E, Potential Implications of Population Dynamics and
Secondary Transmission of Infection on the Benefits of the Ground water Rule, another method is
described and results are presented for the number of secondary cases that might accrue, after infection or
illness from a primary ground water exposure. One difference between these two methods is that in the
quantified benefits section of the GWR EA, EPA assumes that only ill individuals can make others ill. As
is well known from the case of Typhoid Mary, asymptomatic individuals can also serve as carriers of
disease, and thus infected individuals can infect others who can also become ill or who can become
carriers also. The number of secondary cases is limited by the number of susceptible individuals in the
population under consideration and the length of acquired immunity, if appropriate, for the agent.
Appendix E presents a different modeling approach which accommodates dynamic phenomena
typical of real infectious disease transmission systems. A dynamic model is described and results are
presented for the number of secondary cases that might accrue in a large community after Type A virus
(e.g. rotavirus) infection and illness is introduced into a subset of travelers who acquire the infection from
drinking contaminated ground water at a TNCWS well outside their home community. The travelers
return to the home community and secondary cases accrue from normal interactions among infected and ill
individuals who transmit the disease to others in their household or their community. The analysis
considers only infection and illness from two Type A viruses, rotavirus and norovirus.
One important difference between the simple secondary spread factor multiplier used in the GWR
EA and the partial differential equation methodology of Appendix E is in the assumptions about the
characteristics of secondary spread of Type A virus. The GWR EA assumes that 1) only rotavirus causes
Type A illness and 2) all children younger than 3 years of age can infect other individuals with rotavirus,
while older (than 3 years) children and adults can not transmit rotavirus to others. Norovirus is not
considered in the quantified benefits analysis of the EA. The alternative analysis in Appendix E assumes
1) Type A viruses other than rotavirus, such as norovirus are causing illness in the people that consume
water from the hypothetical TNCWS well and 2) symptomatic and asymptomatic individuals of all ages
infected with a Type A virus can infect others who can also become ill or can become asymptomatic
carriers.
Economic Analysis for the 5-32 October 2006
Final Ground Water Rule
-------
The results of the analysis, presented in Appendix E, suggest that population dynamics could
substantially impact the potential benefits calculated quantitatively in this EA, depending on the suite of
population dynamic elements considered. The results demonstrated that important assumptions had a
strong influence on the estimated number of secondary cases. Model parameters that strongly influence
the estimate include the duration of clinical disease, duration of shedding (etiological agent carrier),
etiologic agent infectivity, duration of protective immunity, and person-to-person transmission
characteristics of the infectious agent.
The population dynamic simulation results illustrate that the number of additional illnesses due to
secondary exposure could either increase or decrease relative to the method used in the GWR EA,
depending on the choice of elements. For example, using median values for the elements, it is determined
that approximately 0.18 additional secondary illnesses, on average, would result from each primary
infection and illness. The median value analysis also showed that the number of secondary illnesses due to
asymptomatic carriers is roughly equal to the number of secondary illnesses due to contact with ill
individuals. However, simulations using different model parameters (sensitivity analysis) also demonstrate
that the predicted number of additional illnesses due to secondary transmission could either increase by
approximately an order of magnitude or be reduced to effectively zero, depending on the assumptions
about infection transmission parameters.
The model results presented in Appendix E serve to demonstrate the concept and method of
dynamic transmission system analysis to predict the number of secondary cases that might accrue.
Mortality
For rotavirus (Type A), the mortality factor for all ages used in this analysis is a uniform
uncertainty distribution from 5.7 x 10~6 to 7.3 x 10~6. This range is based on data from a cost-effectiveness
analysis of a rotavirus immunization program (Tucker et al. 1998). The cumulative total for rotavirus
diarrhea was 2,730,000 cases (by year five in a birth cohort), and the estimate of rotaviral deaths for that
same birth cohort was 20. To determine the upper value of the mortality rate range, the deaths are divided
by the diarrhea events (20/2,730,000) resulting in 0.00073%. The lower value for that range is based on a
differing estimate for the forthe total number of rotavirus cases. Kapikian (2001) estimated 3.5 million
rotavirus cases as compared with the CDC estimate of 2.73 million. Applying the CDC estimate of 20
deaths to the Kapikian value for the number of cases yields a lower bound factor of 5.7 x 10"6. These rates
are assumed to apply applied to all age categories.
For Type B viruses, the mortality factor for infants less than 1 month old is based on case fatality
rates from three sources (Jenista 1984, Modlin 1986, Kaplan and Klein 1983). Jenista (1984) reported a
case-fatality rate of 3% for neonates with culture-proven enterovirus infection in the Strong Memorial
Hospital study. Modlin (1986) reported 7 of 206 infants in newborn nurseries who became ill during
echovirus outbreaks died (3.4%). Kaplan and Klein (1983) reported that 6 of 77 patients younger than
three months of age hospitalized with cultures positive for Coxsackievirus died (7.8%). The mean from
these studies is 4.9%. Since this is a hospital case fatality rate, it is necessary to multiply by the proportion
of infants hospitalized to arrive at the mortality rate. Jenista (1984) reported the percent of enteroviral-
infected infants (re-)admitted to the hospital with suspected sepsis was 18.7%. The final mortality rate for
neonates of 0.0092 was calculated as 0.187 * 4.9%. This analysis assumes that neonate enteroviral illness
severity is represented by the Jenista (1984) cohort. Modlin (1986) reports differing severity depending on
whether the neonate is infected by echovirus in utero as compared with during and after birth. The EA
Economic Analysis for the 5-33 October 2006
Final Ground Water Rule
-------
mortality rate is based on hospitalization data from all neonates in a prospective study (Jenista, 1984)
rather than data only from sensitive subgroups such as neonates with infection acquired in utero.
For Type B viruses, the mortality factor used for persons of all ages (other than neonates) is
0.02%. It is assumed that 1% of cases are severe enough to require hospitalization. Modlin (1995) finds
that 0-4% (mode of 2%) of cases that require hospitalization result in death. Currently a general mortality
value of 0.001% for cases that do not require hospitalization is used when specific rates are not available
(Bennet et al., 1987). Therefore, the calculated mortality rate is (0.01 x 0.02) + (.00001* 0.99) = .02%.
The proportion of severe cases is determined to be 1% based on the following calculation. CDC
estimates about 10 million enteroviral illnesses each year (e.g. Strikas et al, 1986). By compiling data on
the annual number of severe enterovirus cases, assumed to be encephalitis (19,000 cases; Khetsuriani et al,
2002), myocarditis (60,000 cases; Kim et al, 2001), and viral meningitis and meningioencephalitis (34,000
cases; Khetsuriani et al, 2003), it is estimated that about 100,000 severe acute illnesses occur each year.
The number of severe cases divided by the total number of cases yields a factor of 1% of cases. As
discussed in Section 5.4, the severe chronic illnesses due to echovirus or other enterovirus, such as dilated
cardiomyopathy and Type I diabetes are not included in the tally of severe cases. In addition, as discussed
in Section 5.2.4.2, the total number of cases (ten million) includes a substantial number of mild hand, foot
and mouth disease cases caused by enterovirus 70 and 71. Thus, the percentage of cases that are severe
would likely be higher if only waterborne enterovirus were considered.
5.2.5 Risk Characterization
Risk characterization combines the hazard identification, exposure assessment, and dose-response
information to describe the overall risk to the exposed population. The following sections will describe the
risk assessment methodology for the baseline risk calculations, provide results of the risk calculations, and
describe the methods used in estimating the reduction in risk from the regulatory alternatives.
5.2.5.1 Risk Assessment Methodology for Baseline (Pre-GWR) Risk Calculations
This section summarizes the risk assessment modeling approach used by EPA to estimate the
baseline number of annual endemic infections, illnesses and deaths due to viruses in GWSs. For a more
complete and detailed discussion of the risk assessment model, the reader is referred to Appendix G.
A Monte Carlo simulation model was used for this risk assessment so that the effects of those
model inputs for which variability and/or uncertainty could be described quantitatively could be reflected
in the estimated cases of illness and deaths.
The risk model operates on 2,376 distinct population categories. The 2,376 population categories
considered in the risk model reflect differences in virus occurrence and individual exposure factors that
influence individual risks of infection and resulting illnesses and deaths in both the pre-GWR baseline and
following implementation of the final GWR (or the other regulatory alternatives considered). The 2,376
strata result from the 3 system types, 9 systems size categories, 8 well types, and 11 age groups that are
considered (3 * 9 * 8 * 11 = 2,376). The 3 system types are CWS, NTNCWS, and TNCWS. The 9 system
size categories are those described in Chapter 4 and provide information on the number of people served
per entry point (well) in each size group of each system type. The 8 well types reflect the various
Economic Analysis for the 5-34 October 2006
Final Ground Water Rule
-------
combinations of being more or less vulnerable to contamination, disinfecting or nondisinfecting, and
construction to standards (2*2*2 = 8). The 11 age groups, as described in the preceding section, reflect
differences in average daily water consumption. The risk assessment model has three major parts that are
summarized below.
The first major part of the risk model generates information on the annual individual risks of
infection for 2,376 population categories. There are two types of outputs from this first part of the model
for each of the 2,376 population categories: (1) uncertainty distributions of average risk of infection and
(2) variability distributions of individual risk of infection.
The second major part of the risk model uses the uncertainty distributions of average individual
risk of infection in each population category as its inputs to calculate the expected number of annual
endemic illnesses and deaths in each of the 2,376 strata, which are then aggregated to arrive at the national
totals.
The third major part of the risk model uses the variability distributions of individual infectivity
risks generated in the first part of the model. In this third part of the model, these 2,376 individual risk
distributions are drawn from to generate an overall individual infectivity risk curves for the baseline and
post-regulatory conditions.
One of the key factors considered in constructing the risk model in parts was to efficiently
accommodate addressing the variability and the uncertainty in the model inputs. The computation of the
range of individual risks of infection in the population groups served by GWSs involves inputs that reflect
both variability and uncertainty. However, the computation of cases of illnesses and deaths, based on the
individual infectivity risks, involves inputs that reflect uncertainty only. The third part of the model
addresses only variability in individual risks of infection.
The first part of the risk model was therefore constructed as a two-dimensional (2D) Monte Carlo
simulation to properly manage both the variability and uncertainty factors involved. For each of the 2,376
population categories, the modeling involved 250 uncertainty loops with 1,000 variability loops within
each of the uncertainty loops. (The specific inputs considered as variability and uncertainty items are
described later in this section.) The first output of this first part of the model (for each of the 2,376
categories) is a set of 250 average individual infection values (that is, the average of each of the 1,000
estimates of individual risk in each of the 250 uncertainty loops). The second output of this first part of the
model (again for each of the 2,376 categories) is a distribution of 1,000 individual infectivity risk values
averaged across the 250 uncertainty loops (where each is first sorted from lowest to highest risk values).
The second part of the risk model is a one dimensional Monte Carlo simulation model that uses the
first output of part one of the model, which is an uncertainty distribution, along with other inputs that are
either constants or are distributions reflecting uncertainty. Again, this second part of the risk model
produces estimates of cases of illness and death, presented as uncertainty distributions from which the best
estimate (taken as the mean) and the lower and upper 90% confidence bounds on the cases (taken from the
5th and 95th percentiles, respectively) are obtained.
The third part of the risk model is also a one dimensional Monte Carlo simulation model in which
values are drawn from the 2,376 distributions of individual risk variability in proportion to the population
that they each represent to construct overall risk distribution curves for both the baseline conditions and for
post-regulatory conditions.
Economic Analysis for the 5-35 October 2006
Final Ground Water Rule
-------
The following provides some further discussion of each of the three parts of the risk assessment
model. Again, the reader is referred to Appendix G for details of the modeling performed. Exhibit 5.12
provides a summary description of the various factors used in the risk calculation and indicates whether
they were inputs as variability distributions, uncertainty distributions, or constants.
Economic Analysis for the 5-36 October 2006
Final Ground Water Rule
-------
Exhibit 5.12 Summary Table of Risk Calculation Factors, Distribution Category
(Variability, Uncertainty, Constant) and Distribution Type) (continued on next page)
Risk Calculation
Factor
% Vulnerable Wells
Occurrence Hit Rate
(Pwell)
Occurrence Hit
Rate (Psample)
Occurrence
Concentration
Fraction of
Disinfecting and
Non-disinfecting
Systems
Log Removal for
Disinfecting
Systems
Description and
Use in Calculations
Used in Step 1 to
characterize whether
the well is less or more
vulnerable
Used in Step 1 to
characterize the
fraction of systems (and
therefore of population)
having viruses present
in the source water.
Used in Step 1 to
determine the
number of days of
exposure to water
with virus present
Used in Step 1 to
characterize the
concentrations of
viruses in source
water.
Used in Step 1 to
separate those
systems currently
practicing disinfection
from those that do
not.
Used in Step 1 , an
assumed 2 or4-log
removal of virus
concentration in
source water for
those systems
practicing
disinfection.
Variability
X
X
X
X
Uncertainty
X
X
Constant
X
Type
Different point
estimates for
different system
types and sizes
Multiple input
sets from
MCMC model
output
Multiple input
sets of
distributions
from MCMC
model output
(uncertainty);
selection of a
specific value
from the
distributions
for each well
(variability)
sample with
replacement
Different point
estimates for
different
system types
and sizes
point estimate
Economic Analysis for the
Final Ground Water Rule
5-37
October 2006
-------
Risk Calculation
Factor
Viability
Drinking Water
Consumption
Type A and Type
B Dose response
Equation
Parameters
Days of
Consumption
Population
Exposed
Average Annual
Individual Risk of
Infection
Description and
Use in Calculations
Used in Step 1 to
indicate the fraction
of viruses in water
considered to be
infectious (assumed
hereto be 1.0).
Used in Step 1 to
characterize the daily
water consumption
by various age
groups in the
exposed population.
Used in Step 1 ,
empirically derived
parameters in
equations used to
calculate daily and
annual risk of
infection.
Used in Step 1 to
indicate the number
of days per year an
exposed individual
consumes water from
ground water
sources.
Used in Step 1 to
indicate the
population
consuming drinking
water from ground
water sources.
Product of the Step 1
calculation used in
Step 2 to calculate
cases of illness and
death.
Variability
X
Uncertainty
X
X
(for
TNCWSs)
X
Constant
X
X
(for CWSs
and
NTCWSs)
X
Type
point estimate
point estimate
for each age
group derived
from
distributions
multiple pairs
of parameters
obtained by
simulation
triangular for
TNCWS,
constants for
CWSs and
NTCWSs
point
estimates
calculated in
risk model
Economic Analysis for the
Final Ground Water Rule
5-38
October 2006
-------
Risk Calculation
Factor
Population Served
Primary Morbidity
Factors
Secondary
Morbidity Factors
Mortality Factors
Description and
Use in Calculations
Used in Step 2 to
scale up the annual
individual infection
risks to total cases of
infection, and
ultimately to total
cases of illness and
death in the exposed
population.
Used in Step 2 to
estimate the number
of illnesses per
infection in the
exposed population.
Used in Step 2 to
estimate the number
of additional illnesses
resulting from contact
with individuals
becoming ill through
primary consumption
of drinking water.
Used in Step 2 to
estimate the number
of deaths in the
exposed population.
Variability
Uncertainty
X
(for Type B
virus)
Constant
X
X
X
(for Type
A virus)
X
Type
point
estimates
point estimate
point estimate
(Type A)
triangular
(Type B)
point estimate
Part 1. Estimation of the Annual Individual Risk of Infection
The key value calculated in Part 1 of the risk assessment model is the annual individual risk of
infection. The algorithm used for this calculation is:
P
Annual
D
In this algorithm, it is implied that an individual's risk of being infected at some time during the
year is the result of his or her experiencing an average daily risk of infection given by PDaiiy for D days of
exposure during the year. The value for D in this algorithm is generated from the value selected for a
given well from the Psampie distribution. The Psampie value, which falls between 0 and 1 and represents the
duration of fraction of time that a contaminated well has virus present, is multiplied by the days of
exposure for each type of well as shown in Exhibit 5.7 to give the days of exposure to water with virus
present.
Economic Analysis for the
Final Ground Water Rule
5-39
October 2006
-------
For reasons discussed in detail in Appendix F, there were different, though related, forms of the
dose-response model for determining the average daily risk of infection for the Type A and Type B virus.
Both model forms are rooted in the Beta-Poisson dose-response model described by Haas et al. (1999).
For Type A viruses, the model form used was:
Pn, =Nx
Daily
where N is the expected number of viruses ingested per day, and • "and • ^re the parameters of a beta
distribution that characterizes the variability in the survival probability of the Type A viruses that are
ingested.
For Type B viruses, the model form used was:
PDa* = l~(l+~/3J
where N is again the expected number of viruses ingested per day, and • "and • ^re the parameters of a
Pareto Distribution used here as an approximation to the exact Beta Poisson distribution.
The expected number of viruses ingested per day, N, in this model is the product of the average
concentration of the virus in the water consumed and the average daily amount of water consumption of
the individual consuming that water. Also, for those categories of wells that practice disinfection, the
concentration of virus was further reduced to reflect either 2-log or 4-log inactivation in proportion to the
fraction of disinfecting systems in that category achieving those levels of disinfection.
As indicated earlier, this part of the risk model was structured as a 2D Monte Carlo simulation to
accommodate both variability and uncertainty in various inputs.
The inputs to this part of the model that were treated as uncertainty items were:
The hit rate, used to characterize the probability that a well would have virus present in the
source water;
• The • ^nd • "parameters used in the daily risk dose-response functions; and
• The fraction of wells that are either more or less vulnerable.
The inputs to this part of the model that were treated as variability items were:
• The concentration of virus in source water when present, and
• The daily water consumption amount (varied by age group).
Economic Analysis for the 5-40 October 2006
Final Ground Water Rule
-------
More details on the characterization of uncertainty in the hit rates and of variability in virus
concentration for the various well types is presented in Chapter 4. More information on water
consumption variability distributions is presented in section 5.2.3 of this chapter. Also, a more thorough
discussion of the derivation of the • "and • •parameters of the infectivity dose-response functions to reflect
uncertainty in those values is provided in Appendix F.
Part 2. Estimation of the Annual Cases of Illness and Death
In the second part of the risk assessment model, the uncertainty distributions of average individual
risk of infection for each of the 2,376 population categories are used together with information on the
number of individuals in each of these categories, risk of illness given infection (morbidity factor), and
factors describing secondary spread of illness to compute the number of cases of annual endemic illnesses.
The basic algorithm for the calculation of illnesses across all 2,376 categories is:5
792
/=!
These calculations are carried out as a Monte Carlo simulation using 1,000 iterations. In addition
to the uncertainty distributions for individual risk of infection that are used as inputs (PAmuaioX uncertainty
is also reflected in the secondary spread factor used for Type B viruses. The resulting output of this step is
an uncertainty distribution of estimated annual endemic cases of illnesses for each virus type from which
the mean is used as the best estimate of the annual cases and the 5th and 95th percentile values are used as
the lower and upper 90% confidence bounds.
The basic algorithm for the calculation of annual endemic deaths is a simple extension of the
algorithm for total illnesses shown above:
792
/=!
which includes the mortality factor (MM(i)) for each category.6 In the risk assessment model, deaths are
calculated in tandem with illnesses in the simulation as noted above. Similarly, the output for this step is
an uncertainty distribution of estimated annual endemic deaths for each virus type from which the mean is
5 It is necessary to modify some of the age categories within these 2,376 groups to apply the morbidity and
secondary spread factors to reflect the specific age ranges as shown in Exhibit 5.10 for which they have been
developed.
6 As with the morbidity and secondary spread factor noted above, some adjustment of the age groups is
made to apply the morbidity factors to reflect the specific age ranges as shown in Exhibit 5.10.
Economic Analysis for the 5-41 October 2006
Final Ground Water Rule
-------
used as the best estimate of the annual cases and the 5th and 95th percentile values are used as the lower and
upper 90% confidence bounds.
Part 3. Characterization of the Distribution of Individual Risks
The outputs of the second part of the model focus on estimates of the number of cases of illnesses
and deaths in the population. Those estimates are built up from consideration of the average or expected
risk of infection from the consumption of virally contaminated ground water applied to the population
groups to which those averages apply. It is noteworthy that these average annual risks of infection span
the full range from 0 (for the vast majority of the population consuming water where there is no "hit" and
therefore no virus present) to a risk of 1 (for those individuals found to be consuming water where viruses
are present at levels sufficiently high to ensure an infection occurring at some point during the course of a
year of exposure).
This third part of the risk assessment model is aimed at providing some insights into how the
annual risks of infection are distributed in the population within this range of 0 to 1. In part 1 of the risk
model, for each of the 2,376 population groups, the simulation produces 1,000 estimates of individual risk
for 250 alternative sets of uncertainty inputs. In Part 3 of the model, the 1,000 individual risk estimates in
each of these 250 distributions of the 2,376 population groups is sorted from lowest to highest risk value.
An overall "expected individual risk" distribution for each of the 2,376 population groups is created by
computing the average of the sorted 1,000 values across the 250 sorted distributions.
Then, in a separate simulation, a probability is assigned to each of these 2,376 "expected
individual risk" distributions reflecting the fraction of the total population served by ground water that
each represents. Then 25,000 individual values are drawn from these 2,376 distributions in proportion to
these probabilities to determine which distribution to draw from on a given iteration, and then randomly
from that "expected individual risk" distribution that was selected.
The result of this process is an overall distribution of individual risk that reflects the entire exposed
population. The shape of this distribution for the baseline conditions can then be compared to similar
distributions constructed for the post-GWR conditions to provide insights into how the rule affects not
only the number of cases of illness and death, but also how it differentially affects those individuals
experiencing different levels of risk prior to the rule.
5.2.5.2 Results of the Baseline Risk Calculations
Estimated annual numbers of endemic illnesses from ingestion of Type A and Type B viruses in
ground water PWSs are summarized in Exhibit 5.13. These are the average of the avoided illnesses and
deaths calculated for each of the 25 years following rule promulgation. This table presents the calculated
mean, as well as the 5th and 95th percentile estimates of annual illness and deaths for Type A and Type B
viruses from the Monte-Carlo simulation. The mean from the simulation is the expected number and the
5th and 95th percentile reflect uncertainty around that reflection.
Economic Analysis for the 5-42 October 2006
Final Ground Water Rule
-------
Exhibit 5.13 Estimates of Annual Baseline Viral Illness and Death1
Virus Type
Type A
TypeB
Total
Illnesses per Year
mean
175,168
10,018
185,186
5th
32,652
501
33,153
95th
435,381
40,718
476,099
Deaths per Year
mean
1.16
2.01
3.18
5th
0.22
0.04
0.26
95th
2.92
8.10
11.02
1 Illnesses are rounded to nearest whole number and deaths to the nearest tenth. Detail may not sum
due to independent statistical analyses.
It is important to recognize that the two-step procedure for calculating the number of cases of
illness and death in the population from exposure to viruses in ground water was carried out separately for
the Type A and Type B virus categories (reflecting different occurrence distributions and dose-response
relationships), different age groups (reflecting different morbidity and mortality factors), different water
system size groups (reflecting different numbers of people served), and different water system types
(reflecting different exposure days of consumption per year for CWSs and NCWSs). The results of these
many separate estimates of risk and cases of illness and death are then summed to obtain the overall
estimates presented in Exhibit 5.13 (for Type A and Type B viruses).
In presenting the results of this two-step procedure for computing the baseline illnesses and deaths
as shown in Exhibit 5.13, the best estimate is the mean of the iterations run in the second step. The
uncertainty in that estimate is characterized by the 5th percentile and 95th percentile values obtained from
those iterations (90 percent confidence bounds). These imply that, given the variability and uncertainty
factors explicitly included in the analysis, there is a 5 percent chance that the actual number of cases falls
below the 5th percentile value, and a 5 percent chance that it falls above the 95th percentile value, and
therefore a 90 percent chance of falling within the specified bounds.
Summing the estimates of illness for both types of viruses gives a combined estimate of more than
180,000 illnesses each year, the majority of which are attributable to the highly infective, but less lethal,
Type A viruses. The estimated combined number of deaths per year is approximately three, with slightly
more of those being due to the more lethal, but less infectious, Type B viruses.
5.2.5.3 Baseline Illnesses and Deaths in Sensitive Subgroups
Exhibit 5.13 above summarizes the total estimated numbers of illnesses and deaths each year from
ingestion of virally contaminated ground water under baseline exposure conditions. Of these illnesses and
deaths, a portion will occur in sensitive subgroups (see Exhibit 5.4). The sensitive subgroups included in
this analysis include the following:
Immunocompromised persons in all age groups: bone marrow transplant recipients, AIDS
patients, and organ transplant patients.
• Children less than 5 years old
• Elderly adults greater 64 years old
Economic Analysis for the
Final Ground Water Rule
5-43
October 2006
-------
These sensitive subgroups comprise approximately 33 percent of the total exposed population
(37.7 million exposed sensitive / 114.3 million total exposed = 33%) (from Exhibits 4.4 and 5.24). The
baseline illnesses and deaths for these sensitive subgroups are shown in Exhibit 5.14. For Type B
illnesses, the high mortality rate of enteroviruses among neonates (infants less than one month old)
contributes to a higher proportion of deaths in sensitive subgroups, which account for approximately 75
percent of the echovirus deaths (1.5 / 2.0 = 75%) (Exhibits 5.13 and 5.14). For this dose response
assessment, there is no specific morbidity and mortality information for the elderly or for the
immunocompromised. Also, persons with autoimmune disorders were not included as part of the sensitive
population. Although these persons may experience severe consequences of infection, their increased
severity has not been modeled or included as part of the economic analysis.
Exhibit 5.14 Baseline Illnesses and Deaths in Sensitive Subgroups
Virus Type
lype M
lype a
Health
Effect
Illness
Death
Illness
Death
Immunocom-
promised1
(all ages)
526
0.0
30
0.0
Infants and
Young
Children
<5 years old2
11,413
0.1
761
0.2
Elderly
Adults
>65 years
old2
23,370
0.2
6,542
1.3
Total
Sensitive
Subgroups
35,308
0.2
7,333
1.5
Footnotes: 1) The immunocompromised population is estimated to account for 0.3% of the total number
of illnesses and deaths. 2) The Immunocomprised portion has been excluded from "Infants and Young
Children" and "Elderly" subgroups to avoid counting them twice.
Sources: U.S. population data is from the 2000 U.S. census.
5.2.5.4 Baseline Risk to a Highly Exposed Individual
The annual risk of illness was also estimated for a highly exposed individual. For this calculation,
it is assumed that a typical, highly-exposed individual would ingest drinking water from an untreated,
more vulnerable well contaminated by virus. The source water from such a system is assumed to be
contaminated with viral pathogens at a concentration of 40 viruses/100 L, which approximates the mean
value (41.5) of the 7 virus concentration values from the Lieberman data used for the More Vulnerable
wells. Because the source water is not disinfected, there is no inactivation of viral pathogens in the
system. This person would be a member of the age category with the highest water ingestion rate at the
75th percentile, which is 55-64 years. The 75th percentile of daily intake of drinking water for this age
group is 1.925 L/day. Annual exposure under this scenario is 49 days, based on 350 days/year
consumption from a CWS and a Psampie value of 0.14, which is the 75th percentile of the expected values
from the Psampie distributions. The analysis also assumes dose response parameter inputs and morbidity
factor inputs corresponding to mean values of their respective distributions.
Using these assumptions for drinking water exposure, the calculated annual probability of Type A
viral illness for a person with these characteristics is 0.30. For Type B illness, the annual probability of
illness is 0.035. These results reflect the higher infectivity of Type A viruses in comparison to the
moderately infective Type B viruses.
Economic Analysis for the
Final Ground Water Rule
5-44
October 2006
-------
5.2.5.5 Sensitivity of Baseline Estimates to Quantified Uncertainty Inputs
The Monte Carlo simulation model includes several inputs for which EPA has explicitly
considered uncertainty. In some cases, these values are selected from a specific distribution describing the
range and probability of particular values. In other cases, these values are selected from a set of plausible
values that have been generated using Bayesian methods as described previously. In some cases, a
particular uncertain input value is specifically related to one or more other uncertain input values.
This section describes an analysis performed by EPA to assess the relative influence of these
inputs on the estimated annual cases of illness. The inputs fall into two general categories: occurrence
inputs and dose-response morbidity inputs.
Uncertainty in Occurrence Inputs
The key occurrence inputs to the are: Pwell, Psampie distribution parameters, and Percent More
Vulnerable Wells.
The Pwell, and Psample distribution parameters used in the model are selected from a set of 10,000
values generated by a Bayesian analysis as described in Section 4.3.4.1 of Chapter 4 of the GWR EA. It is
important to note that Pwell, and Psample are selected from the dataset for input to the model as paired values.
As described previously, there is an inverse relationship between Pwell, and the expected value of the Psample
distribution. As Pwell, increases, the expected value for Psample decreases and vice versa. Therefore, it is
necessary to select an appropriately paired set of Pwell, and Psample values as inputs to each iteration of the
model to ensure that this relationship is maintained.
The Percent More Vulnerable parameter is sampled from uniform distributions that are determined
for each water system type and size based on acute and non-acute MCL violation data (see Section 4.3.4.2
in Chapter 4). In the risk model, the More Vulnerable wells typically have higher virus concentrations,
which are expected to result in higher numbers of infections and illnesses. Note, however, that the range
of the uniform uncertainty distributions of Percent More Vulnerable wells is relatively small (typically
0.5% to 5%), so the overall influence of this uncertainty input on the number of cases of illnesses is small
(Exhibits 5.15a-b).
Uncertainty in Dose Response Inputs
The key dose-response inputs are the parameters for the infectivity dose-response equations and
the morbidity factors.
The infectivity dose-response relationships for Type A and Type B virus include two parameters
(alpha and beta). For the risk model input, a set of 1,000 pairs of parameters were developed for both
Type A and Type B viruses from the challenge studies (see Section 5.2.4.1 in Chapter 5). In each iteration
of the Monte Carlo simulation model, a parameter pair (alpha, beta) is selected for Type A and another is
independently selected for Type B.
The uncertainties in the primary morbidity factors are incorporated as uniform distributions that
differ for the two virus types and by certain age groups for each virus. Since these factors are applied to
the number of cases of infection, higher morbidity factor values will result in higher cases of illness.
Economic Analysis for the 5-45 October 2006
Final Ground Water Rule
-------
For Type B viruses there is also uncertainty in the secondary spread factor, which is included as a
symmetrical triangular distribution. As with the primary morbidity factors, higher values for this factor
will lead to increases in the number of illnesses estimated.
Approach to Sensitivity Analysis
A sensitivity analysis was performed to evaluate the relative influence of the selected
uncertain inputs on the range of estimates of baseline cases of illness. Because of the complexity of the
risk assessment simulation model and the total number (2,376) of strata of well types and population age
groups, this sensitivity analysis was limited to one specific age group (20-24 year olds) served by CWSs,
but it does consider all well-types and size categories. A separate model run was performed using 100
uncertainty loop iterations and 250 inner loop iterations. The output captured the number of cases of
illness estimated for each virus for each of the 100 loops, as well as the 100 specific values selected for the
uncertain inputs in each iteration.
For the uncertain Pwelland Psampie occurrence inputs, a metric was used to combine the effect of the
selected values since these inputs are linked and have an inverse relationship as described above. The
metric used was the product of Pwell and the expected value of the Psampie distribution for that iteration,
referred to here as PW*PS. The reason for using this metric was that the product of these values reflects the
combination of the fraction of wells that are virally contaminated and the fraction of time that
contaminated wells have virus present. Higher values of PW*PS would be expected to result in a higher
number of cases of illness.
For the uncertain inputs for the parameters for the infectivity dose-response equations, a metric
was also used to reflect the combined effect of the alpha and beta parameters. For the selected parameters
in a given iteration, the probability of infection from an exact dose of one viral infectious unit was
computed.
For the other uncertain inputs, the actual values selected for the given iteration were used.
The Pearson Correlation Coefficient was computed based on the uncertain input values (or metric
is described above) and the estimated annual cases of illness obtained. Pearson Correlation Coefficients
are always values that fall between -1 and +1. Positive values indicate that increases in the uncertain input
values lead to increases in the estimates of cases; negative values imply that increases in the input values
lead to decreases in the estimates of cases. The larger the absolute value of the coefficient, the more
pronounced is its influence on the results.
The results obtained are summarized below and shown graphically in Exhibits 5.15a and 5.15b for
Type A and Type B viruses, respectively.
Virus Type A:
PW*PS: 0.323
Percent More Vulnerable Wells: 0.114
Infectivity Dose-Response Parameters: 0.531
Primary Morbidity Factor: 0.384
Economic Analysis for the 5-46 October 2006
Final Ground Water Rule
-------
Virus Type B:
PW*PS: 0.054
Percent More Vulnerable Wells: -0.027
Infectivity Dose-Response Parameters: 0.033
Primary Morbidity Factor: 0.114
Secondary Spread Factor: -0.025
These results indicate that for Type A viruses, the uncertainty in the dose-response parameters is
the main factor affecting the number of illnesses, while for Type B viruses it is the Primary Morbidity
Factor. It should be pointed out, however, that because the values of all of these correlation coefficients
are relatively low, none would be considered strong drivers of the outcome individually. Rather, it would
appear that it is the random combinations of multiple uncertain input factors that lead to higher (or lower)
estimates of cases of illness.
Exhibit 5.15a Summary of Pearson Correlation Coefficients for Uncertain Inputs,
Virus Type A
Summary of Pearson Correlation Coefficients for Uncertain Inputs - Virus Type B
-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Pen
Primary Mo
Infectivity Do;
Parar
Secondary S
ent More Vuln
bidity Factor
Pw*Ps
e-Response
iters
•
D
D
Dread Facton
erableWellJ
Economic Analysis for the
Final Ground Water Rule
5-47
October 2006
-------
Exhibit 5.15b Summary of Pearson Correlation Coefficients for Uncertain Inputs,
Virus Type B
Summary of Pearson Correlation Coefficients for Uncertain Inputs - Virus Type A
-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Infectivity Do;
e-Response
Primary Mo bidity Factor
5.2.5.6 Methodology for Estimating Risk Reductions
The methodology for estimating the reduction in risk for the regulatory alternatives builds upon
the approach and assumptions used to establish the baseline risk as described in the preceding section. The
primary difference between the modeling for estimating the baseline risk model and the modeling for
estimating the risk reduction from a given regulatory alternative is that the latter incorporates a change in
the concentration of viral pathogens reaching the finished drinking water of the exposed population.
These changes reflect either a reduction in pathogen concentration between source water and finished
water due to disinfection or the complete elimination of the pathogen in the finished water from other
non-treatment corrective actions addressing the source water contamination. In addition to accounting for
the magnitude of pathogen exposure reduction, an important component of the risk reduction modeling is
to account for the timing of when those reductions occur over a 25 year analysis timeframe following
promulgation of the rule.
As discussed in the description of the baseline risk analysis in Section 5.2.5.1, each well in the
simulation process is designated as either having a virus present at some time or never having a virus
present based on the Pwell probability. Also, for those wells having some viral occurrence, values are
assigned for Psample and for the virus concentration. The risk reduction part of the model uses the exact
same simulated wells as those generated in the baseline risk part of the model.
Economic Analysis for the
Final Ground Water Rule
5-48
October 2006
-------
For the sake of efficiency in implementing the simulation modeling process, those wells
designated as never having a virus present are recognized as having zero risk reduction potential and are
counted as such in the model outputs, but are not run through the detailed steps of the risk reduction
model.
For those wells that do have a virus present, the risk reduction model answers the following three
questions:
1) Is a corrective action performed on this well as a result of the regulatory alternative being
considered?
2) What is the finished water virus concentration following corrective action?
3) In what year following rule implementation is the corrective action performed?
The risk reduction model then processes the reduced virus concentrations through the dose
response functions for infectivity, morbidity and mortality as in the baseline risk assessment.
In the baseline risk analysis, the primary outputs are estimates of annual cases of illness and deaths
due to endemic infection from Type A and Type B viruses, and these are assumed to be the same for each
of the 25 years following rule promulgation. The outputs from the risk reduction model are the same - the
cases of illness and death that remain in each of the 25 years following rule promulgation. These are the
remaining cases each year resulting from the virus concentrations that remain after the corrective actions
are performed. The risk reductions, in terms of cases of illness and deaths avoided, are then obtained by
subtracting the cases remaining after the rule from the baseline cases.
Exhibit 5.16 shows an example of what the model might predict for a particular well with a viral
pathogen present. In this example, the baseline annual cases of illness resulting from the presence of a
pathogen in the undisinfected source water is assumed to be 1,000. The baseline risk shows these 1,000
cases each year for the 25 year period. If 4-log disinfection is implemented as a corrective action after
year 10 as a result of the rule, the remaining number of baseline cases would be 1, beginning in year 11
and continuing through year 25. The cases avoided each year for this well are, then, 0 for the first 10 years
and 999 for years 11 through 25.
Economic Analysis for the 5-49 October 2006
Final Ground Water Rule
-------
Exhibit 5.16 Example of Baseline Cases, Cases Remaining, and Cases Avoided
at the Well Level
Year
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Baseline
Cases
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
Cases
Remaining
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1,000
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Cases
Avoided
0
0
0
0
0
0
0
0
0
0
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
Estimates of cases avoided calculated for all of the individual wells are then aggregated to arrive at
the total national estimates of risk reduction. In addition, some of the assumptions and data used in the
risk reduction model are uncertain and are therefore input as uncertainty distributions. As a result of the
uncertainty reflected in those inputs, together with the uncertainty reflected in other inputs to the baseline
risk model that are also carried into the risk reduction model, the output of the model is a range of values
of cases avoided. The range is used by EPA to determine the expected value and the 90 percent
confidence bounds on that expected value.
The following sections describe in more detail the specific assumptions and inputs-including
considerations of uncertainty-that are used to model risk reduction for the sanitary survey and triggered
monitoring components of the GWR.
Economic Analysis for the
Final Ground Water Rule
5-50
October 2006
-------
Sanitary Surveys
The Sanitary Survey component of the rule applies to all ground water wells, including those wells
that are currently disinfecting and meeting 4-log reductions. As described in the baseline information in
Chapter 4 and in the estimation of the baseline risk, all wells are stratified into 8 major categories that
reflect the 2x2x2 (=8) combinations of the well characteristics of: Disinfecting or Nondisinfecting; More
Vulnerable or Less Vulnerable; and Proper or Improper well construction.
As discussed in Chapter 4 (Section 4.3.3), the fraction of wells considered More Vulnerable varies
from 0% to approximately 7% as a function of systems system size and type. The average (weighted by
number of systems) is that about 2.5% are in the More Vulnerable stratum (the remaining 97.5% are,
therefore, in the Less Vulnerable stratum). To estimate the benefits from correcting significant
deficiencies, each of these strata must be further defined in regards to well construction.
Construction status of a well (i.e., whether a well is properly or improperly constructed) is
estimated based on ASDWA survey data (ASDWA, 1997). Those wells identified as improperly
constructed are likely to be identified by states during a sanitary survey. Different percentages of improper
construction are estimated based on historical total or fecal coliform detections. EPA believes that a
history of detection of these contaminants is indicative of wells that fall into the More Vulnerable
classification within the benefits model. The percentages of properly and improperly constructed wells are
estimated as follows.
Less vulnerable wells - The ASDWA survey of States found that of community GWSs with no TC
or fecal coliform detections, 83.6 percent of the systems had wells that were constructed according to State
standards. Thus, 16.4 percent of systems had wells identified by State officials as not being constructed to
State standards and are considered to be improperly constructed.
More vulnerable wells - The same survey found that of community GWSs with TC detections, but
no fecal coliform detections, 217 of 211 systems had wells constructed according to State standards. A
third group consisted of systems with positive fecal coliform detections, of which 164 of 231 systems had
wells constructed according to State standards. Thus, for systems with TC or fecal coliform detections,
381 of 508 systems (75.0 percent) had wells that were constructed according to State standards. Therefore,
25.0 percent of systems had wells identified by State officials as not being constructed to State standards
and are considered to be improperly constructed.
The combination of the vulnerability and well construction estimates described above can be used
to approximate the percent of all wells having virus present that are expected to be identified and corrected
by sanitary surveys. To arrive at this estimate, the percent of all wells that are improperly constructed is
first calculated as the weighted average for the More and Less Vulnerable strata as:
• More vulnerable (2.5%), Improperly constructed (25.0%):
2.5%* 25.0% = 0.6%
Less vulnerable (97.5%), Improperly constructed (16.4%):
97.5%* 16.4% =16.0%
The total fraction of all wells that are improperly constructed is the sum of these, which is 16.6%.
Economic Analysis for the 5-51 October 2006
Final Ground Water Rule
-------
It is assumed that the sanitary survey provisions of the ground water rule will result in identifying
some (but not all) wells that are improperly constructed. It is assumed that corrective actions will be
performed on those improperly constructed wells. Specifically, EPA has estimated that for those wells that
are improperly constructed and also have a virus present, 50% will be identified and the contamination
eliminated by a corrective action. To recognize the uncertainty in this estimate of the effectiveness of
sanitary surveys, EPA has assumed a uniform distribution of 40% to 60% (mean = 50%). Therefore, as a
central value, it is expected that the sanitary survey and corrective action alternative of the rule will result
in corrective actions being performed and contamination eliminated at approximately half of the 16.6% of
all wells that are improperly constructed, which is 8.3% of all wells. Therefore, it can be estimated that
this rule alternative will result in the reduction of 8.3% of the baseline cases of illnesses and deaths per
year by the 25th year after rule promulgation.
For those wells that have virus present in the source water that are caught by a sanitary survey, the
corrective action is assumed to eliminate the virus completely. Therefore, the finished water concentration
as well as the source water concentration for these wells is set to zero for the risk reduction calculations for
those wells. While the fraction of wells corrected by sanitary survey is the same for both the disinfecting
and the nondisinfecting strata, almost all of the risk reduction (i.e., cases avoided) from sanitary surveys
will be from the nondisinfecting wells. This is because for the disinfecting wells that are currently
achieving 4-log removal, most of the virus present in the source water is already being inactivated so that
the incremental risk reduction from eliminating the source entirely is very small and contributes very little
to the total benefits achieved. Nevertheless, the risk reduction for disinfecting wells that is achieved
through sanitary surveys is calculated in the risk reduction model for completeness sake.
With respect to when this reduction in virus concentration and risk occurs, it is assumed that
identifying the (approximately) 8.3% of wells that are improperly constructed wells and performing a
corrective action on them as a result of sanitary surveys will occur across the entire 25 year analysis
period. It is assumed that no corrective actions will occur in the first three years, and that all of the
corrective actions performed will be evenly distributed across the remaining years. Therefore, in the
benefits model, each well that undergoes a corrective action from a sanitary survey is assigned a year
between year 4 and year 25 with equal probability of it occurring in any one of the years in the time
period.
Triggered Monitoring
The triggered monitoring component of the rule applies only to the nondisinfecting subset of wells
and any wells that are applying disinfectant but not achieving 4-log removal of viruses. For the wells that
are currently achieving 4-log removal, the sanitary survey requirements still apply and the output
developed in that part of the model as described above are retained for those wells
For the triggered monitoring component of the risk reduction analysis, each well goes through a
2-step process. In the first step of the process, estimates are made of the number of TC positives-and
therefore the number of source water indicator samples-that occur during the 22 years between year 4 and
year 25 (it is assumed that no corrective actions will be taken during the first 3 years after rule
implementation). The number of TC positives expected per year for each well of a given type and size are
obtained from the DV data as described in chapter 4 (Section 4.2.7). The total number of TC positives
expected through year 25 is then calculated as the number TC positives per year times 22. If, for example,
a well is in the CWS size 10,000-50,000 category, the DV data indicate that these wells average 2.21 TC
Economic Analysis for the 5-52 October 2006
Final Ground Water Rule
-------
positives per year; over 22 years, then, it is expected that these wells will have 48.6 TC positives and
therefore take up to 49 source water indicator samples between years 4 and 25.
In the second step of the process, a simulation is performed to determine which, if any, of the
indicator samples taken through year 25 is the first positive indicator result. As described in Chapter 4, in
each uncertainty loop of the risk reduction model a set of values for both virus and indicator hit rates is
selected. Included among this set of values is the probability that the first positive of an indicator will
occur on a given assay number (contingent on these assays being performed in wells that are known to
have a virus present at some time). Exhibit 4.27 showed the probability of the first indicator positive
occurring on a given assays for the median and the 5th and 95th percentiles of a sample of 1,000 of these
uncertainty sets of occurrence values. The curve for the median set of values shown in that exhibit
indicates that there is about a 40% probability that the first indicator positive will occur on or before the
49th assay. The data for each uncertainty set provides these cumulative probabilities of observing the first
positive on or before each specific assay number. In the risk reduction model, a random value between 0
and 1 is generated for each well. That value is used as a look-up value to determine what assay number
would produce the first positive.
For example, if the curve shown as the median data set in Exhibit 4.27 were the set of values being
used for a particular uncertainty loop, and the random number between 0 and 1 generated for a well in the
CWS size 10,000-50,000 category were 0.25, the look-up function would indicate that the first indicator
positive would occur on assay number 8. Since these wells are expected to take 48.6 indicator assays over
the 22 year period, the 8th assay would occur in the 6th year (48.6 / 8 ~ 6). Since there are no samples
taken in years 1 through 3, the 6th year of sampling corresponds to year 9 of the 25 year modeling period.
Therefore, this well would be "caught" by triggered monitoring in year 9. This prediction is then
compared to the year in which the well is captured (if it is) by the sanitary survey provisions. The
corrective action is assigned to the rule provision (SS or TM) that occurs the earliest in the 25 year period.
If both occur in the same year, one of the two is selected randomly.
It is important to note that this analysis assumes no correlation between the occurrence of total
coliform in the distribution system and the occurrence of fecal contamination in the well water source. In
fact, the two are positively correlated in systems that do not disinfect. Bacteria in the source water of
systems that do not disinfect can directly cause total coliform positives in samples taken from the
distribution system. Although this relationship is known to exist, EPA has insufficient data to include it in
an occurrence model. As a result, only about 3% of TC positives are estimated to lead to indicator
positives in the triggered monitoring samples. The true rate of indicator positives is expected to be greater,
with more disinfecting systems taking corrective action. Not modeling the relationship between TC
positives and fecal contamination at the source contributes to underestimation of the effectiveness of
triggered monitoring. As a result, both benefits and costs (to a lesser degree) are underestimated.
Based on the number of TC positives expected per well across all well types and sizes, together
with the expected values of indicator positives as a function of assay number across all of the uncertainty
sets available to draw from for the simulation model, it can be estimated that approximately one-third
Economic Analysis for the 5-53 October 2006
Final Ground Water Rule
-------
(33%) of nondisinfecting wells with virus present should be caught and corrected by the triggered
monitoring provision of the rule by the 25th year of the modeling timeframe.7
Because there are some nondisinfecting wells that will be caught in the simulation model by both
sanitary survey and triggered monitoring (where the one occurring earliest is selected for corrective
action), the total wells ultimately caught by both components of this rule alternative will be less than the
sum of the two individual components (i.e., -8.3% for SS and 27.5% for TM). The expected fraction of
nondisinfecting wells that are captured by either SS or TM can be estimated from the sum of these minus
the product (to account for the overlap):
(8.3% + 27.5%) - (8.3% * 27.5%) = 33.5%
Therefore, the SS + TM alternative should, by the 25th year after rule promulgation, result in
corrective actions being performed at approximately 33.5% of all non-disinfecting wells and 8.3% of all
(4-log) disinfecting wells that have virus present in their source water. Because most of the baseline risk
is found in nondisinfecting wells, it is also, therefore, expected that this rule will result in approximately a
33.5% reduction in the baseline cases of illness and death.
5.2.5.7 Results for Risk Reduction for the Final GWR
The final GWR, the risk targeted approach, includes sanitary surveys, triggered monitoring,
corrective action, and compliance monitoring. The estimated reduction in illnesses and deaths are
presented by system size and type in Exhibit 5.17 and in summary in Exhibit 5.18. These values are the
annual average cases avoided across the 25 year analysis timeframe, and they include the initial three years
after implementation when no corrective actions are performed as a result of sanitary surveys or triggered
monitoring.
Exhibit 5.19 provides a summary of the estimated cases avoided per year in the 25th year following
rule implementation. This summary provides an indication of the magnitude of the annual benefits that
can ultimately be achieved by the GWR. The cases of illness avoided shown here, 41,868,
represent an approximately 22.6% reduction of baseline cases of 185,186 shown in Exhibit 5.13. A
detailed breakdown of illnesses and deaths avoided by age group is presented in Appendix B.
7 The effect of taking five repeat samples for any positive indicator sample is not included in this analysis.
Based on repeat sampling, some wells will not be caught as described in this section (i.e., all repeat samples will be
negative) and benefits will not accrue to those wells. The effect of this omission is a slight overestimate of benefits,
however, the number of systems with five negative repeat samples is expected to be small.
Economic Analysis for the 5-54 October 2006
Final Ground Water Rule
-------
Exhibit 5.17 Annual Viral Illnesses and Deaths Avoided for the GWR by System
Size and Type
Type A Viruses
Illnesses
A
Deaths
B
Type B Viruses
Illnesses
C
Deaths
D
Total
Illnesses
E=A+C
Deaths
F=B+D
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-50K
50,001-100K
100.001-1M
>1 Million
All Sizes
338
783
752
1,844
4,619
4,394
8,832
8,336
0
29,900
0.00
0.01
0.00
0.01
0.03
0.03
0.06
0.06
0.00
0.20
26
57
55
132
327
338
582
615
0
2,131
0.01
0.01
0.01
0.03
0.06
0.07
0.12
0.12
0.00
0.42
363
840
807
1,977
4,946
4,732
9,414
8,951
0
32,031
0.0
0.0
0.0
0.0
0.1
0.1
0.2
0.2
0.0
0.6
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-50K
50,001-100K
100.001-1M
>1 Million
All Sizes
134
428
395
489
241
151
52
93
0
1,983
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.01
6
19
22
35
14
8
3
4
0
111
0.00
0.00
0.00
0.01
0.00
0.00
0.00
0.00
0.00
0.02
141
447
416
524
255
159
55
97
0
2,094
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-50K
50,001-100K
100.001-1M
>1 Million
All Sizes
Total
1,806
2,555
1,081
913
469
491
75
170
0
7,560
39,442
0.01
0.02
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.05
0.26
40
62
30
24
12
11
1
3
0
184
2,426
0.01
0.01
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.04
0.48
1,846
2,617
1,111
937
481
501
76
173
0
7,743
41,868
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.7
Note: Detail may not add to totals due to independent rounding. The figures presented in this exhibit represent only the
quantifiable benefits of the GWR. The unquantified benefits are expected to comprise a significant portion of the overall
benefits of the Final Rule and are presented in Section 5.4.
Source: Appendix B
Economic Analysis for the
Final Ground Water Rule
5-55
October 2006
-------
Exhibit 5.18 Summary of Annual Viral Illnesses and Deaths Avoided for the GWR
Virus type
Type A
TypeB
Total
Illnesses per Year
Mean
39,442
2,426
41,868
5th Percentile
10,093
181
10,274
95th Percentile
79,925
8,114
88,039
Deaths per Year
Mean
0.3
0.5
0.7
5th Percentile
0.1
0.0
0.1
95th Percentile
0.5
1.6
2.1
Note: Details may not add to totals due to independent rounding and independent statistical analyses.
Source: Appendix B
Exhibit 5.19 Viral Illnesses and Deaths Avoided in the 25th Year
after Implementation of the GWR
Virus type
Type A
TypeB
Total
Illnesses per Year
Mean
59,126
3,583
62,709
5th Percentile
15,387
259
15,646
95th Percentile
121,578
12,372
133,951
Deaths per Year
Mean
0.4
0.7
1.1
5th Percentile
0.1
0.0
0.1
95th Percentile
0.8
2.4
3.2
Note: Details may not add to totals due to independent rounding and independent statistical analyses.
Source: Appendix B
5.2.5.8 Results for Reduction in Individual Risks for the Final GWR
Exhibits 5.20a and 5.20b present the distributions of individual annual risks of infection for the
population using ground water for Type A and Type B viruses, respectively. These graphs show the risks
separately for individuals consuming water from CWS, NTNCWS and TNCWS systems. Baseline risks
and annual risks following the full implementation (25 years) of the rule alternatives are presented. Note
that curves that are further to the right indicate more risk, those to the left indicate less risk.
In general, the graphs show a relatively small reduction in individual risk from the Sanitary Survey
and Corrective Action alternative, somewhat larger risk reductions for the Risk Targeted and Multi-Barrier
alternatives, and substantially greater reductions for the 4-log across-the-board disinfection alternative.
These changes in individual risk track well with the estimated reductions in annual cases of illness and
deaths for the various rule alternatives presented in previous sections. In addition, the annual individual
infection risks are substantially higher for Type A than Type B viruses, which corresponds to the
differences in the dose response functions for the two types of viruses.
Also note that these graphs imply that the annual individual baseline risks for CWS consumers is
lower than that for the noncommunity water systems, and that individual risk for NTNCWSs is slightly
higher than that for TNCWSs. This can be understood in terms of the differences in current disinfection
and differences in water consumption days and amounts. (Virus occurrence is assumed to be the same in
all system types in the primary analysis.) Although there is more water consumption both in terms of days
of consumption and volume per day per individual in CWSs than in either NTNCWS or TNCWS (which
would tend to push individual risk up), a much larger portion of the CWS wells are currently disinfecting
than the NTNCWS or TNCWS wells, which pushes the overall individual risk distribution for CWS down.
Between NTNCWSs and TNCWSs, both the percent of wells disinfecting and water consumption are
Economic Analysis for the
Final Ground Water Rule
5-56
October 2006
-------
lower in the TNCWSs than in NTNCWSs and this is reflected in the comparing these baseline graphs.
(Note that for the 4-log rule alternative which results in all wells in all system types being disinfected, a
comparison of the risk curves across the three types of systems shows the greatest risk in CWS, slightly
less in NTNCWS, and much less in TNCWS. This is because with this alternative, only differences in
water consumption days and volume per day drive the individual risk levels for consumers at these
different systems types.
Economic Analysis for the 5-57 October 2006
Final Ground Water Rule
-------
Exhibit 5.20a Comparison of Average Annual Individual Infectivity Risk
Distributions for Baseline and Rule Alternatives for Type A Viruses
I
_L JL
///
T/7/
1 OOE-06 1 OOE-05 1 OOE-04 1 OOE-03
Annual Individual Risk of Infection
1 .OOE-08 1 .OOE-07 1 .OOE-06 1 .OOE-05 1 .OOE-04 1 .OOE-03 1 .OOE-02 1 .OOE-01 1 .OOE-KJO
Annual Individual Risk of Infection
7Z
77
77
JOE-08 1.OOE-07 1.OOE-06 1.OOE-05 1.OOE-04 1.OOE-03 1.OOE-02 1.OOE-01 1.00E400
Annual Individual Risk of Infection
Economic Analysis for the
Final Ground Water Rule
5-58
October 2006
-------
Exhibit 5.20b Comparison of Average Annual Individual Infectivity Risk
Distributions for Baseline and Rule Alternatives for Type B Viruses
I///
TT7L
I//
LlL
1 OOE-08 1 OOE-07 1 OOE-06 1 OOE-05 1 OOE-04 1 OOE-03 1 OOE-02 1 OOE-01 1 OOE+00
Annual Individual Risk of Infection
1 .OOE-08 1 .OOE-07 1 .OOE-06 1 .OOE-05 1 .OOE-04 1 .OOE-03 1 .OOE-02 1 .OOE-01 1 .OOE-KJO
Annual Individual Risk of Infection
1 .OOE-08 1 .OOE-07 1 .OOE-06 1 .OOE-05 1 .OOE-04 1 .OOE-03 1 .OOE-02 1 .OOE-01 1 .OOE-KJO
Annual Individual Risk of Infection
Economic Analysis for the
Final Ground Water Rule
5-59
October 2006
-------
5.2.5.9 Potential Increases in Health Risks
It is unlikely that the GWR will result in a significant increase in risk from other contaminants,
although adding disinfection to currently non-disinfecting systems could result in some increased risk.
When disinfection is first introduced into a previously undisinfected system, the disinfectant can react with
pipe scale causing increased risk from some contaminants and water quality problems. Contaminants that
could be released include lead, copper, and arsenic. It could also possibly lead to a temporary
discoloration of the water as the scale is loosened from the pipe. These risks can be addressed by
gradually phasing in disinfection to the system, by targeted flushing of distribution system mains, and by
maintaining a proper corrosion control program.
Using a chemical disinfectant could also result in an increased risk from disinfection byproducts
(DBFs). Risk from DBFs has already been addressed in the Stage 1 Disinfection Byproducts Rule (DBPR)
(USEPA, 1998d) and additional consideration of DBP risk has been addressed in the recently published
final Stage 2 DBPR (USEPA, 2006e). In general, GWSs are less likely to experience high levels of DBFs
than surface water systems, because they have lower levels of naturally occurring organic materials
(generally represented by total organic carbon (TOC)) that contribute to DBP formation. For the most
part, GWSs with high levels of TOC in their ground water source are located in States that already require
GWSs to disinfect, therefore decreasing the chance that significant disinfection byproduct problems would
result from this rule.
Economic Analysis for the 5-60 October 2006
Final Ground Water Rule
-------
5.3 Monetized Benefits from Reduction in Exposure to Waterborne Pathogens
Once the annual illnesses and deaths avoided as a result of the GWR implementation are
estimated using the risk model described in the previous sections, monetary unit values can be applied to
these estimates to establish the monetary benefits attributable to the rule. As discussed earlier, only a
portion of the illnesses and deaths avoided are quantified. Further, the only benefits estimated and
monetized for the main analysis in this EA are acute health effects of viral infections. The following
sections explain how the value of reductions in illnesses and deaths was derived and presents results of
the monetized benefits calculations.
5.3.1 Value of Reduction in Type A and Type B Virus Cases
5.3.1.1 Value of Viral Illnesses Avoided
The goal of this analysis is to provide as complete an accounting as possible of the social welfare
impacts of the regulatory requirements under consideration. Based on the principles of welfare
economics, the preferred approach for valuing reductions in the risk of Type A and Type B virus
morbidity is to rely on estimates of willingness to pay (WTP) for these risk reductions. However, there
are no direct studies of WTP for avoided morbidity from the viral illnesses considered in the benefits
analysis. As a minimum estimate of that value, this analysis estimates the value of averted morbidity
risks based on the (1) avoided medical costs and (2) the value of avoided time losses for the illness they
cause. The rationale for and limitations of this approach are discussed below, and greater detail is
provided in Appendix A.
The calculation of medical costs includes the costs of medical services and medications received
by ill individuals. The assumption behind using these costs as a benefit measure is that a policy that
reduces the incidence of illness will yield benefits at minimum equal to the costs avoided. Cost of illness
(COI) estimates, however, may significantly underestimate individual willingness to pay for a variety of
reasons. In particular, these estimate do not: (1) address the value of avoiding pain and suffering; (2)
include costs that individuals incur to avoid the illness (i.e., defensive or averting expenditures); (3)
reflect aversion to risk (the fear of becoming ill); (4) consider ex ante values (they are based on ex post
costs); and (5) consider whether treatment returns individuals to their original health state (i.e., is
equivalent to avoiding the illness entirely).
A number of researchers have explored the relationship between the COI and individual WTP for
risk reductions. This research suggests that the ratio of these two types of values varies greatly depending
on the nature of the health effect, the characteristics of the individuals studied, and the factors included in
the construction of each estimate. Comparison studies result in WTP to COI ratios ranging from about a
factor of 2 to as much as a factor of 79 (in one case); many of the ratios are between 3 and 6.8 In other
words, the COI estimates were typically one-third to one-sixth of the WTP estimates, but the ratio varied
greatly.
See Appendix B of EPA's Handbook for Non-Cancer Health Effects Valuation (USEPA, 2000c) for a review of these
studies.
Economic Analysis for the October 2006
Final Ground Water Rule 5-61
-------
In some cases, COI studies include indirect, as well as direct costs. These indirect costs usually
include lost earnings due to missed market work time, and may also include costs associated with reduced
productivity while at work and/or lost nonmarket work time (e.g., child care or housekeeping).
Typically, these costs are estimated using the human capital approach, which focuses on the value of
goods and services that are bought and sold in the marketplace and ignores other aspects of time use that
affect individual well-being.
The analysis of Type A and Type B virus-related morbidity uses two measures of the
COI-referred to in this EA as Traditional and Enhanced. Both approaches include direct medical costs
and the value of lost work time, but differ in the assessment of value of lost work time. They both
consider the impact of time losses on foregone market production, which affects the individual worker
(e.g., in terms of lost income) as well as other members of society (who benefit from the availability of
the goods or services produced as well as the taxes paid), and foregone nonmarket (household and
volunteer) production, which affects the individual and other household members and often has impacts
outside the home. The Traditional COI includes nonmarket (unpaid) work time based on replacement
costs and does not include a value for the time lost of children under 16 years of age. The other approach,
the Enhanced COI, values nonmarket work time based on opportunity costs and places a value on the lost
time of children. Both approaches also include values for the nonmarket time lost by friends or family
members caring for those who are sick,9 but the approaches use different values for this lost time.
The Enhanced COI also includes the value of lost leisure time and lost productivity—the reduced
utility (or sense of well-being) associated with decreased enjoyment of time spent in both market and
nonmarket activities. The Enhanced COI is an attempt to more completely measure the loss of welfare
from an illness.
A search of the literature suggests that researchers have not attempted to estimate directly (e.g.,
through surveys) the difference between the value of time in a well state compared to time in an ill state.
This analysis relies instead on wage and compensation data to estimate the opportunity costs of time
usage. This approach recognizes that, because resources are limited, any decision to use resources for one
purpose means that they cannot be used for other purposes. One minimal measure of the value of a
resource, therefore, is the value of its next best use.
The application of the opportunity cost approach to paid work time is relatively clear, since
compensation can be used to estimate these costs. For other (unpaid) time spent in nonmarket work or
leisure activities, wage data are also used based on the assumption that (at the margin) the wage
represents the opportunity cost of engaging in such activities.
More precisely, lost market work is valued at the median gross (pre-tax) wage rate plus benefits,
also referred to as total compensation or employer's costs. This approach is most representative of the
full social impact of lost work because it incorporates both the loss to the individual in terms of lost
income and the loss to society in terms of reduced tax revenue or decreased production of goods and
services. Lost nonmarket work and leisure time is valued at the median net (post-tax) wage rate. This
approach reflects the assumption that, at the margin, an individual will choose to engage in nonmarket
work or leisure activities only if the value of these activities exceeds the wage rate that the individual
9 Paid care is included in the medical cost component of the analysis and hence is not discussed in the
discussion of time losses.
Economic Analysis for the October 2006
Final Ground Water Rule 5-62
-------
would otherwise earn. Sleep time presents special problems in this analysis, both because data on the
effect of Type A and Type B virus—related morbidity on the amount or quality of sleep time is not
available and because current literature on valuing lost time has not settled on an accepted valuation
method. This analysis, thus, conservatively assumes that lost sleep time has zero value.
These values are applied to both complete losses of time (time spent in illness-related activities
rather than normal activities) as well as to partial losses (time spent in normal activities that is less
productive or pleasurable than in the absence of illness). In the latter case, however, the dollar value of
the loss is prorated to reflect the fact that the individual does not completely lose the productivity or
utility associated with the activity. These values are applied to such time losses incurred by the ill
individual.
The use of medical costs and the opportunity cost of time losses to value morbidity related to
Type A and Type B viruses may understate the value of these risk reductions for a variety of reasons.10
As noted earlier, COI estimates generally understate WTP for a variety of reasons, e.g., because they
exclude consideration of the value of avoided pain and suffering or of risk aversion. In addition, the use
of wage and compensation data to value lost time may understate the utility of time spent in its preferred
use. The use of wage rates may understate the total utility associated with an activity even in the case of
paid work, because individuals may derive intrinsic pleasure from the activity above and beyond the
income they receive. For nonmarket work and leisure, the value of the activity to the individual may
exceed the opportunity cost for similar reasons. In addition, nonmarket work and other activities can
provide benefits to other members of society that are not reflected in the individual wage rate. Finally,
this approach does not include the value of lost sleep time.
In addition, relying on wage data for valuing lost time presents difficulties in the case of
individuals for whom these data are not available, such as children, the unemployed, and those out of the
labor market. For the Enhanced COI approach, all lost time of children is valued at the median post-tax
wage rate. No estimates exist, however, for the indirect cost of illness. EPA transferred the method used
for days lost due to illness (equal to the duration of illness) to children. EPA further assumes that a
caretaker stays home with these children, introducing additional lost caretaker days. The rationale for this
approach is discussed in more detail below. It is unclear whether this approach under- or overstates the
value of time losses for the individuals in these other categories, given the available information on these
values. However, the Agency's Children's Health Valuation Handbook states: "To the extent that a
caregiver is more likely to be involved when a child is recuperating, the total value of lost time is likely to
be higher for a child's illness than for an adult's." (USEPA, 2002)
COI Calculations
The primary risk of illness that the GWR addresses is from endemic exposure to Type A and
Type B viruses and the resulting acute cases of illnesses. Many elements of the COI come from a
10 There are a number of other simplifying assumptions inherent in the application of this approach that
may lead it to under- or overstate the value of time losses, related to factors such as the functioning of the labor
market, the treatment of individuals who are not labor force participants, the use of average or median (rather than
marginal) earnings data, and the possibility that substitute activities (e.g., watching TV instead of normal activities)
have some positive value. It is unclear whether, in total, these practical limitations serve to increase or decrease the
bias that results from the sources discussed in this paragraph.
Economic Analysis for the October 2006
Final Ground Water Rule 5-63
-------
literature review of current medical studies; other elements have been assumed. The literature review
yielded information on the duration of illness from Type A and Type B viruses, the type of medical care
sought, if any, and the costs associated with these services. The data from the literature review, as well as
assumptions made in this analysis, are described in Appendix A.
The computation of COI involves two broad categories of costs—direct and indirect medical
costs. All costs are updated to a common year (2003) used as the starting point for projecting benefits
into future time periods. For Type A viruses, each cost component has a separate estimate made based on
age and the health state of the individual (healthy or immunocompromised). For Type B viruses, cost
components have separate estimates based both on age and on the type of care required: no medical care
(93% of cases); outpatient care (6% of cases); or inpatient care (1% of cases). Detail on these breakouts
is also provided in Appendix A. The next two subsections discuss the details of the computations used to
derive direct and indirect medical costs.
Direct Medical Costs
For both the Enhanced COI and Traditional COI, the cost for a case of Type A or Type B viral
illness is derived by summing the costs of outpatient and inpatient care. Outpatient care consists of an
initial physician visit ($114.55) and product of the cost of each follow-up visit ($66.18) and the number
of follow-up visits. Multiplying this sum by the percentage of patients that utilize outpatient services
yields the weighted unit cost of outpatient care. The cost of inpatient care consists of the costs of the
initial doctor visit in the hospital ($152.87), any follow-up visits ($52.25), and the hospital charges
(calculated on a per day basis, with costs ranging from $1,007 per day to $4,870 per day). As with
outpatient costs, multiplying the sum of doctor visits and hospital charges by the percentage of patients
who require inpatient care yields the weighted unit cost of inpatient care. The sum of the weighted unit
costs of outpatient and inpatient care equals the weighted direct costs.
Exhibits 5.2 la - 5.21b and 5.22a - 5.22b show the weighted direct medical costs per case of Type
A viral illness, ranging from an average cost of $0 (for healthy patients, 5 years or older) to an average
cost of $4,486 (for immunocompromised patients younger than 5 years old). The weighted direct medical
costs per case of Type B viral illness (Exhibits 5.21c-5.21e and 5.22c-5.22e) range from an average of $0
(for patients requiring no medical care) to $23,431 (for patients less than 1 year old requiring inpatient
care). Costs for doctor visits, follow-up visits, and hospital stays constitute the direct medical costs.
Costs for initial and follow-up doctor visits are obtained in 2000 dollars and updated to 2003 dollars using
the consumer price index (CPI) for medical care services.
Economic Analysis for the October 2006
Final Ground Water Rule 5-64
-------
Exhibit 5.21 a Estimates for Average Cost and Average Cost per Healthy Patient
of Type A Illness, by Age (Enhanced COI)
Cost Category
Symptom duration--
reported days (range) [A]
Average Cost Per Patient
2003$
<2 years
3
2 to 4 years
3
5 to 15 years
3
>16 years
2.5
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
0.5
(0-9)
$66.18
$148
$115 -$710
14%
$21
$16 - $101
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
14%
$21
$16 - $99
$114.55
N/A
$66.18
$115
0.0%
$0
$114.55
N/A
$66.18
$115
0.0%
$0
Inpatient costs
Duration of hospital stay (days) [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
5
$1,007.19
$5,036
$152.87
4
$52.25
$5,398
1 .4%
$76
$97
$16 - $101
5
$1,007.19
$4,029
$152.87
4
$52.25
$4,391
1 .4%
$61
$82
$16 - $99
N/A
$1,007.19
$4,029
$152.87
N/A
$52.25
$4,338
0.0%
$0
$0
$0- $0
N/A
$1,007.19
$4,029
$152.87
N/A
$52.25
$4,338
0.0%
$0
$0
$0- $0
Indirect Costs
Value of lost patient day [D]
Lost patient days (No Medical
Care/Outpatient) [A]
Lost patient days (Inpatient) [A]
Value of caregiver day [D]
Caregiver days (No Medical
Care/Outpatient) [A]
Caregiver days (Inpatient) [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$199.36
3
5
$227.79
3
5
$59.81
0
$1,293
$1,390
$1,310 - $1,395
$199.36
3
5
$227.79
3
5
$59.81
0
$1 ,293
$1,376
$1,309 - $1,393
$199.36
1.5
N/A
$227.79
1.5
N/A
$59.81
0
$641
$641
$227.79
1 (15.3%), 0(84.7%)
[Note 1]
N/A
$227.79
0
N/A
$68.34
1
$103
$103
Note 1: For the healthy population >16 years, the severity of symptom manifestation is dependent on the rotavirus
strain and can be divided into two groups. The G2 & G9 strains represent 15.3% of illnesses, while all other strains
comprise the remaining 84.7%.
Sources:
[A] Appendix A, Exhibit A. 1
[B] Appendix A, Exhibit A.2
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($227.79) multiplied by percent loss productivity (30 percent, rounded
from Harrington et al. 1991)
Economic Analysis for the
Final Ground Water Rule
October 2006
5-65
-------
Exhibit 5.21 b Estimates for Average Cost and Average Cost per
Immunocompromised Patient of Type A Illness, by Age (Enhanced COI)
Cost Category
Symptom duration-
reported days (range) [A]
Average Cost Per Patient
2003$
<2 years
5
2 to 4 years
3
5 to 15 years
3
>16 years
2.5
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
100%
$148
$115 - $710
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
100%
$148
$115 - $710
$114.55
0
$66.18
$115
100.0%
$115
$114.55
0
$66.18
$115
100.0%
$115
Inpatient costs
Duration of hospital stay (days) [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
5
$1,007.19
$4,029
$152.87
4
$52.25
$4,338
100%
$4,338
$4,486
$4,453 - $5,049
3
$1,007.19
$4,029
$152.87
2
$52.25
$4,338
100%
$4,338
$4,486
$4,453 - $5,049
3
$1,007.19
$4,029
$152.87
2
$52.25
$4,338
100%
$4,338
$4,453
2.5
$1,007.19
$4,029
$152.87
2
$52.25
$4,338
100%
$4,338
$4,453
Indirect Costs
Value of lost patient day [D]
Lost patient days [A]
Value of caregiver day [D]
Caregiver days [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$199.36
5
$227.79
5
$59.81
0
$2,136
$6,622
$6,589 - $7,184
$199.36
3
$227.79
3
$59.81
0
$1,281
$5,767
$5,734 - $6,330
$199.36
3
$227.79
3
$59.81
0
$1,281
$5,734
$227.79
2.5
$227.79
0
$68.34
0
$569
$5,022
Sources:
[A] Appendix A, Exhibit A. 1
[B] Appendix A, Exhibit A.2
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($227.79) multiplied by percent loss productivity (30 percent, rounded
from Harrington et al. 1991)
Economic Analysis for the
Final Ground Water Rule
October 2006
5-66
-------
Exhibit 5.21 c: Estimates for Average Cost and Average Cost per Case of Type B
Illness Requiring No Medical Care, by Age (Enhanced COI)
Cost Category
symptom duration-
reported days (range) [A]
Average Cost per Patient
2003$
<1 Year
3(1-6)
1 to 4 years
3(1-6)
5 to 15 years
3(1-6)
>16 years
3(1-6)
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
NA
NA
NA
NA
0.0%
$0.00
NA
NA
NA
NA
0.0%
$0.00
NA
NA
NA
NA
0.0%
$0.00
NA
NA
NA
NA
0.0%
$0.00
Inpatient costs
Duration of hospital stay [A]
Hospital cost per day
Total hosptial cost
Initial physician visit
Average # of follow-up visits [A]
Cost per follow-up
Total Unit Cost
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
Indirect Costs
Value of lost patient day [D]
Lost patient days [A]
Value of caregiver day [D]
Caregiver days [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs1 [C]
Total Cost of Illness [C]
$199.36
3(1-6)
$227.79
3(1-6)
$59.81
0
$1,281
$427 - $2,563
$1,281
$427 - $2,563
$199.36
3(1-6)
$227.79
3(1-6)
$59.81
0
$1,281
$427 - $2,563
$1,281
$427 - $2,563
$199.36
1.5
$227.79
1.5
$59.81
0
$641
$641
$227.79
1.15
$227.79
0
$68.34
1.09
$336
$336
1 Total Indirect Costs is shown as a range for those age categories for which the duration of symptoms, and hence
lost patient days, was a range determined from best available data.
Sources:
[A] Appendix A, Exhibit A.3
[B] Appendix A, Exhibit A.4
[C] Calculations based on data presented in the table
[D] Appendix A, Exhibit A.7.E] Weighted average per day value of time ($227.79) multiplied by percent loss
productivity (30 percent, rounded from Harrington et al. 1991)NA: Not applicable to mild cases.
Economic Analysis for the
Final Ground Water Rule
October 2006
5-67
-------
Exhibit 5.21 d Estimates for Average Cost and Average Cost per Case of Type B
Illness Requiring Outpatient Care, by Age (Enhanced COI)
Cost Category
symptom duration-
reported days (range) [A]
Average Cost per Patient
2003$
<1 Year
5(2-12)
1 to 4 years
5(2-10)
5 to 15 years
5 (2-7)
>16 years
5 (2-9)
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100.0%
$181
$114.55
1
$66.18
$181
100.0%
$181
Inpatient costs
Duration of hospital stay [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
Indirect Costs
Value of lost patient day D
Lost patient days A
Value of caregiver day [D]
Caregiverdays [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$199.36
5(2-12)
$227.79
5(2-12)
$59.81
0
$2,136
$ 854 - $ 5,126
$2,316
$ 1,035 - $ 5,307
$199.36
5(2-10)
$227.79
5(2-10)
$59.81
0
$2,136
$ 854 - $ 4,272
$2,316
$1,035 - $ 4,452
$199.36
5 (2-7)
$227.79
5 (2-7)
$59.81
0
$2,136
$ 854 - $ 2,990
$2,316
$1,035 - $ 3,171
$227.79
5 (2-9)
$227.79
0
$68.34
0
$1,139
$456 - $2,050
$1 ,320
$ 636 - $ 2,231
Sources:
[A] Appendix A, Exhibit A.3
[B] Appendix A, Exhibit A.4
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($227.79) multiplied by percent loss productivity (30 percent, rounded
from Harrington et al. 1991)
* No data available
Economic Analysis for the
Final Ground Water Rule
October 2006
5-68
-------
Exhibit 5.21 e Estimates for Average Cost and Average Cost per Case of Type B
Illness Requiring Inpatient Care, by Age (Enhanced COI)
Cost Category
symptom duration--
reported days (range) [A]
Average Cost per Patient
2003$
<1 Year
7(2-14)
1 to 4 years
7(2-14)
5 to 15 years
7(2-14)
>1 6 years
7(2-14)
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100.0%
$181
$114.55
1
$66.18
$181
1 00.0%
$181
Inpatient costs
Duration of hospital stay--
reported days [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
4.7
$4,869.90
$22,889
$152.87
4
$52.25
$23,250
100%
$23,250
$23,431
4.7
$1,839.15
$8,644
$152.87
4
$52.25
$9,006
100%
$9,006
$9,187
2.5
$1,839.15
$4,598
$152.87
2
$52.25
$4,855
1 00%
$4,855
$5,036
2.5
$1,839.15
$4,598
$152.87
2
$52.25
$4,855
1 00%
$4,855
$5,036
Indirect Costs
Value of lost patient day [D]
Lost patient days [A]
Value of caregiver day [D]
Caregiver days [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$199.36
7(2-14)
$227.79
7(2-14)
$59.81
0
$2,990
$ 854 - $ 5,980
$26,421
$ 24,285 - $ 29,411
$199.36
7(2-14)
$227.79
7(2-14)
$59.81
0
$2,990
$ 854 - $ 5,980
$12,177
$ 10,041 - $ 15,167
$199.36
7(2-14)
$227.79
7(2-14)
$59.81
0
$2,990
$ 854 - $ 5,980
$8,026
$5,890 - $ 11,016
$227.79
7(2-14)
$227.79
0
$68.34
0
$1,595
$456 - $3,189
$6,631
$5,492 - $ 8,225
Sources:
[A] Appendix A, Exhibit A.3
[B] Appendix A, Exhibit A.4
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($227.79) multiplied by percent loss productivity (30 percent, rounded
from Harrington et al. 1991)
* No data available
Economic Analysis for the
Final Ground Water Rule
October 2006
5-69
-------
Exhibit 5.22a Estimates for Average Cost and Average Cost per Healthy Patient
of Type A Illness, by Age (Traditional COI)
Cost Category
Symptom duration-
reported days (range) [A]
Average Cost Per Patient
2003$
<2 years
3
2 to 4 years
3
5 to 15 years
3
>16 years
2.5
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
14%
$21
$16 - $101
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
14%
$21
$16 - $99
$114.55
N/A
$66.18
$115
0.0%
$0
$114.55
N/A
$66.18
$115
0.0%
$0
Inpatient costs
Duration of hospital stay (days) [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
5
$1,007.19
$5,036
$152.87
4
$52.25
$5,398
1 .4%
$76
$97
$16 - $101
5
$1,007.19
$4,029
$152.87
4
$52.25
$4,391
1 .4%
$61
$82
$16 - $99
N/A
$1,007.19
$4,029
$152.87
N/A
$52.25
$4,338
0.0%
$0
$0
$0- $0
N/A
$1,007.19
$4,029
$152.87
N/A
$52.25
$4,338
0.0%
$0
$0
$0- $0
Indirect Costs
Value of lost patient day [D]
Lost patient days (No Medical
Care/Outpatient) [A]
Lost patient days (Inpatient) [A]
Value of caregiver day [D]
Caregiver days (No Medical
Care/Outpatient) [A]
Caregiver days (Inpatient) [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$0.00
3
5
$85.12
3
5
$0.00
0
$258
$354
$274 - $359
$0.00
3
5
$85.12
3
5
$0.00
0
$258
$340
$274 - $357
$0.00
1.5
N/A
$85.12
1.5
N/A
$0.00
0
$128
$128
$85.12
1 (15.3%), 0 (84.7%)
[Note 1]
N/A
$85.12
0
N/A
$25.54
1
$39
$39
Note 1: For the healthy population >16 years, the severity of symptom manifestation is dependent on the rotavirus
strain and can be divided into two groups. The G2 & G9 strains represent 15.3% of illnesses, while all other strains
comprise the remaining 84.7%.
Sources:
[A] Appendix A, Exhibit A. 1.
[B] Appendix A, Exhibit A.2.
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($85.12) multiplied by percent loss productivity (30 percent, rounded from
Harrington et al. 1991)
Economic Analysis for the
Final Ground Water Rule
October 2006
5-70
-------
Exhibit 5.22b Estimates for Average Cost and Average Cost per
Immunocompromised Patient of Type A Illness, by Age (Traditional COI)
Cost Category
Symptom duration-
reported days (range) [A]
Average Cost Per Patient
2003$
<2 years
5
2 to 4 years
3
5 to 15 years
3
>16 years
2.5
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
100%
$148
$115 - $710
$114.55
0.5
(0-9)
$66.18
$148
$115 - $710
100%
$148
$115 - $710
$114.55
0
$66.18
$115
100.0%
$115
$114.55
0
$66.18
$115
100.0%
$115
Inpatient costs
Duration of hospital stay (days) [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
5
$1,007.19
$4,029
$152.87
4
$52.25
$4,338
100%
$4,338
$4,486
$4,453 - $5,049
3
$1,007.19
$4,029
$152.87
2
$52.25
$4,338
100%
$4,338
$4,486
$4,453 - $5,049
3
$1,007.19
$4,029
$152.87
2
$52.25
$4,338
100%
$4,338
$4,453
2.5
$1,007.19
$4,029
$152.87
2
$52.25
$4,338
100%
$4,338
$4,453
Indirect Costs
Value of lost patient day D
Lost patient days A
Value of caregiver day [D]
Caregiver days [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$0.00
5
$85.12
5
$0.00
0
$426
$4,912
$4,879 - $5,474
$0.00
3
$85.12
3
$0.00
0
$255
$4,741
$4,708 - $5,304
$0.00
3
$85.12
3
$0.00
0
$255
$4,708
$85.12
2.5
$85.12
0
$25.54
0
$213
$4,666
Sources:
[A] Appendix A, Exhibit A. 1.
[B] Appendix A, Exhibit A.2.
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($85.12) multiplied by percent loss productivity (30 percent, rounded from
Harrington et al. 1991)
Economic Analysis for the
Final Ground Water Rule
October 2006
5-71
-------
Exhibit 5.22c Estimates for Average Cost and Average Cost per Case of Type B
Illness Requiring No Medical Care, by Age (Traditional COI)
Cost Category
Symptom duration-
reported days (range) [A]
Average Cost per Patient
2003$
<1 Year
3(1-6)
1 to 4 years
3(1-6)
5 to 15 years
3(1-6)
>16 years
3(1-6)
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
NA
NA
NA
NA
0.0%
$0.00
NA
NA
NA
NA
0.0%
$0.00
NA
NA
NA
NA
0.0%
$0.00
NA
NA
NA
NA
0.0%
$0.00
Inpatient costs
Duration of hospital stay [A]
Hospital cost per day
Total hosptial cost
Initial physician visit
Average # of follow-up visits [A]
Cost per follow-up
Total Unit Cost
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$0
Indirect Costs
Value of lost patient day [D]
Lost patient days [A]
Value of caregiver day [D]
Caregiver days [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$0.00
3(1-6)
$85.12
3(1-6)
$0.00
0
$255
$85 - $511
$255
$85 - $511
$0.00
3(1-6)
$85.12
3(1-6)
$0.00
0
$255
$85 - $511
$255
$85 - $511
$0.00
1.5
$85.12
1.5
$0.00
0
$128
$128
$85.12
1.15
$85.12
0
$25.54
1.09
$126
$126
Sources:
[A] Appendix A, Exhibit A.3.
[B] Appendix A, Exhibit A.4.
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($85.12) multiplied by percent loss productivity (30 percent, rounded from
Harrington et al. 1991)
NA: Not applicable to mild cases.
Economic Analysis for the
Final Ground Water Rule
October 2006
5-72
-------
Exhibit 5.22d Estimates for Average Cost and Average Cost per Case of Type B
Illness Requiring Outpatient Care, by Age (Traditional COI)
Cost Category
Symptom duration-
reported days (range) [A]
Average Cost per Patient
2003$
<1 Year
5(2-12)
1 to 4 years
5(2-10)
5 to 15 years
5 (2-7)
>16 years
5 (2-9)
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100.0%
$181
$114.55
1
$66.18
$181
100.0%
$181
Inpatient costs
Duration of hospital stay [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
NA
NA
NA
NA
NA
NA
NA
0%
$0.00
$181
Indirect Costs
Value of lost patient day [D]
Lost patient days [A]
Value of caregiver day [D]
Caregiverdays [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$0.00
5(2-12)
$85.12
5(2-12)
$0.00
0
$426
$170 - $1,021
$606
$351 - $1,202
$0.00
5(2-10)
$85.12
5(2-10)
$0.00
0
$426
$170 - $851
$606
$351 - $1,032
$0.00
5 (2-7)
$85.12
5 (2-7)
$0.00
0
$426
$170 - $596
$606
$351 - $777
$85.12
5 (2-9)
$85.12
0
$25.54
0
$426
$170 - $766
$606
$351 - $947
Sources:
[A] Appendix A, Exhibit A.3.
[B] Appendix A, Exhibit A.4.
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($85.12) multiplied by percent loss productivity (30 percent, rounded from
Harrington et al. 1991)
* No data available.
Economic Analysis for the
Final Ground Water Rule
October 2006
5-73
-------
Exhibit 5.22e Estimates for Average Cost and Average Cost per Case of Type B
Illness Requiring Inpatient Care, by Age (Traditional COI)
Cost Category
Symptom duration-
reported days (range) [A]
Average Cost per Patient
2003$
<1 Year
7(2-14)
1 to 4 years
7(2-14)
5 to 1 5 years
7(2-14)
>16 years
7(2-14)
Direct Costs
Outpatient costs
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Outpatient [A]
Weighted Unit Cost [C]
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100%
$181
$114.55
1
$66.18
$181
100.0%
$181
$114.55
1
$66.18
$181
100.0%
$181
Inpatient costs
Duration of hospital stay--
reported days [A]
Hospital cost per day [B]
Total hosptial cost [C]
Initial physician visit [B]
Average # of follow-up visits [A]
Cost per follow-up [B]
Total Unit Cost [C]
Percent Inpatient [A]
Weighted Unit Cost [C]
Total Weighted Direct Costs [C]
4.7
$4,869.90
$22,889
$152.87
4
$52.25
$23,250
100%
$23,250
$23,431
4.7
$1,839.15
$8,644
$152.87
4
$52.25
$9,006
100%
$9,006
$9,187
2.5
$1,839.15
$4,598
$152.87
2
$52.25
$4,855
100%
$4,855
$5,036
2.5
$1,839.15
$4,598
$152.87
2
$52.25
$4,855
100%
$4,855
$5,036
Indirect Costs
Value of lost patient day [D]
Lost patient days [A]
Value of caregiver day [D]
Caregiverdays [A]
Value of lost productivity day [E]
Lost productivity days [A]
Total Indirect Costs [C]
Total Cost of Illness [C]
$0.00
7(2-14)
$85.12
7(2-14)
$0.00
0
$596
$170 - $1,192
$24,027
$23,601 - $24,623
$0.00
7(2-14)
$85.12
7(2-14)
$0.00
0
$596
$170 - $1,192
$9,782
$9,357 - $10,378
$0.00
7(2-14)
$85.12
7(2-14)
$0.00
0
$596
$170 - $1,192
$5,632
$5,206 - $6,228
$85.12
7(2-14)
$85.12
0
$25.54
0
$596
$170 - $1,192
$5,632
$5,206 - $6,228
Sources:
[A] Appendix A, Exhibit A.3.
[B] Appendix A, Exhibit A.4.
[C] Calculations based on data presented in the table.
[D] Appendix A, Exhibit A.7
[E] Weighted average per day value of time ($85.12) multiplied by percent loss productivity (30 percent, rounded from
Harrington et al. 1991)
* No data available.
Indirect Medical Costs
For the Enhanced COI, the total indirect cost associated with a case of Type A viral illness ranges
from an average of $103 (for healthy patients 16 years old and older) to $2,136 (for patients under 2 years
of age). Indirect costs associated with cases of Type B viral illness range from $336 (for patients 16 years
Economic Analysis for the
Final Ground Water Rule
October 2006
5-74
-------
old and older requiring no medical care) to $2,990 (for patients under 16 years of age requiring inpatient
care). Total indirect cost is the sum of the value of patient days lost, the value of productivity lost, and
the value of care giver days lost. Exhibit 5.21 also includes this information. The value of a lost day is
set at $227.79 for adults 16 years of age and older, and $199.36 for children under 16 years of age. As
described in Appendix A, the figure for adults is calculated by summing the product of hours of market
work per day and the median gross (pre-tax) wage and benefits ($20.82), the product of hours of
nonmarket work per day and the median post-tax wage ($12.46), and the product of hours of leisure time
per day and the median post-tax wage ($12.46), and the product of hours of sleep per day and a rate of
zero. The figure for children uses a median post-tax wage ($12.46) for all lost time (16 waking hours).
For the Traditional COI, the total indirect medical cost associated with a case of Type A viral
illness ranges from an average of $39 (for healthy patients 16 years old and older) to $426 (for
immunocompromised patients 2 years of age and younger). Indirect costs associated with cases of Type
B viral illness range from $126 (for patients 16 years old and older requiring no medical care) to $596
(for patients requiring inpatient care). Total indirect cost is the sum of the value of patient days lost, the
value of productivity lost, and the value of care giver days lost. Exhibit 5.22 also includes this
information. The value of a lost day for adults and caregivers is set at $85.12. As shown in Appendix A,
this figure is calculated by summing the product of hours of market work per day and the median gross
(pre-tax) wage and benefits ($20.82), and the product of hours of nonmarket work per day and the median
post-tax wage ($6.23).
While the methodology used to derive the indirect cost of illness for adults is straightforward, the
valuation of children's time presents unique problems. The best approach when valuing children's health
effects is the use of child-specific valuations of these effects. For direct costs, EPA has used such
valuations. Indirect costs, however, prove more challenging. As noted in the Children's Health
Valuation Handbook (USEPA, 2002), "[children's] time lost to sickness also has value, although no direct
measure exists for this loss." In this instance, the Handbook states that, "as a second-best option,
...transfer benefit values estimated for adults to children." The Enhanced COI uses this guideline, in
conjuncture with Executive Order 13045 ("Protection of Children from Environmental Health Risks and
Safety Risks"), and assumes a day lost due to illness for the duration of illness for patients younger than
16 years to be valued at $199.36 (based on the median post-tax wage). In contrast, the Traditional COI
assigns no value for a lost day for children under 16 years of age. Both the Traditional and Enhanced
COI approaches assume that a caretaker stays home with these children, introducing additional lost
caretaker days for each lost patient day. According to the Handbook, "the productivity loss of both
affected individuals should be included in the valuation estimate of a child's illness." The number of
days lost entirely to illness, either by the patient or care giver, is multiplied by $227.79 (for the Enhanced
COI) or $85.12 (for the Traditional COI), the average value of a lost day.
Often patients return to work or school while still experiencing symptoms that affect their
productivity. Because patients are only fractionally as productive at work as well people, the loss
associated with the less productive days (lost productivity) is a portion of the value of a full lost day,
specifically 30 percent11 (rounded from Harrington, 1991). Since there is only a fractional productivity
loss, the days with lessened productivity are multiplied by $68.34 (30 percent of $227.79 for Enhanced
COI) or $25.54 (30 percent of $85.12 for Traditional COI). No productivity losses are assigned to
children under 16 years of age under either the Traditional or Enhanced COI approaches.
See Appendix A.
Economic Analysis for the October 2006
Final Ground Water Rule 5-75
-------
5.3.1.2 Value of Mortality Avoided
Benefits of the GWR also derive from avoiding fatalities due to Type A and Type B virus
infections. The Value of a Statistical Life (VSL) is used to measure the value of these benefits. The VSL
represents an estimate of the monetary value of reducing risks of premature death. The VSL, therefore, is
not an estimate of the value of saving a particular individual's life. The value of a "statistical" life
represents the sum of the values placed on small individual risk reductions across an exposed population.
For example, if a regulation were to reduce the risk of premature death from Type B viral infection by
1/1,000,000 for one million exposed individuals, the regulation would "save" one statistical life
(1,000,000 X 1/1,000,000). If each of the 1,000,000 people were willing to pay $5 to achieve the risk
reduction anticipated from the regulation, the VSL would be $5 million ($5 X 1,000,000).
An EPA study characterized the range of possible VSL values as a Weibull distribution with a
mean of $4.8 million (1990 price level) based on 26 individual study estimates (USEPA 1997b). This
represents the value recommended for use in benefits analyses in EPA's Guidelines for Preparing
Economic Analyses (USEPA 2000e) and endorsed by the Science Advisory Board (SAB) Arsenic review
panel (USEPA 200 Ib). For purposes of the GWR benefits analysis, the VSL Weibull distribution (with
parameters of location = 0, scale = 5.32, shape = 1.51) was incorporated into the benefits model Monte-
Carlo simulation. This enables quantification of the uncertainty surrounding benefits estimates derived
from the VSL. The mean VSL, after all adjustments were made, was $7.4 million in year 2003 dollars
(using a CPI adjustment factor). These adjustments are explained further in the following sections, and
the mean VSL by year of the analysis can be found in Appendix B (Exhibit B.6) of this EA.
5.3.1.3 Measuring Benefits Over the GWR Implementation Schedule
In order to extract benefits data from the model and present these benefits in comparable terms to
a similarly calculated stream of costs, it is necessary to calculated the present value of all benefits over the
lifetime of the implementation schedule. GWR implementation occurs over several years as States and
PWSs learn the requirements, inform their staffs, perform sanitary surveys, and implement source water
and compliance monitoring. A 25-year horizon was chosen for this analysis, because most treatment
technologies evaluated in this EA are estimated to have 20-year life-cycle. In addition, systems have
several years to begin treatment associated with the GWR. Calculating a shorter time frame would include
less of the complete value associated with the cost of technologies. A complete schedule of when costs
and benefits are estimated to be incurred is presented in Appendix B.
5.3.1.4 Adjustments for Income Elasticity
Although the price level (year 2003) is held constant across all benefits projections, real income
increases overtime, and therefore benefit values in future years are adjusted to reflect income elasticity or
income growth, depending on the benefits category being assessed. Benefits based on potentially fatal
health effects are adjusted for income elasticity and income growth. Benefits based on the value of lost
time are adjusted for income growth, but not income elasticity. This section describes how these
adjustments are carried out. Benefits derived from medical costs, the third broad category of benefits, are
adjusted for neither income elasticity nor income growth.
Economic Analysis for the October 2006
Final Ground Water Rule 5-76
-------
In the case of avoided-death benefits, income elasticity adjustments are applied to values in future
years. In general, income elasticity represents changes in valuation in relation to changes in real income.
For example, if, for every 1 percent increase in real income, a particular consumer's willingness to pay for
a particular item increases by 1 percent, this would be represented by an income elasticity of one. For
most items, income elasticity values are actually less than one, reflecting slower growth in willingness to
pay than in income.
In order to apply the income elasticity values in the benefits model, they must be combined with
projections of real income growth over the time frame for analysis. To accomplish this, population and
real gross domestic product (GDP) projections are combined to calculated per capita real GDP values12
(see Appendix B, Exhibit B.5-income elasticity calculations). Percent changes in these values overtime
can then be combined with income elasticity figures to derive a single adjustment factor.13 Given any two
points in time, this factor is calculated as follows:
Income elasticity adjustment factor = (E Ij - E I2 -12 - Ii) / (EI2 -E Ij -12 -10 where:
E = income elasticity
I] = real income (per capita GDP) in the base year
I2 = real income (per capita GDP) in the year of analysis
When applying this formula, income elasticity adjustment factors are calculated from the same
base year as the values subject to adjustment. In this case, income elasticity factors for fatal cases of viral
illness are calculated from a 1990 base year (Ij = 1990 in the above formula) because that is the base year
used in the study from which VSL estimates are derived.14
Kleckner and Neuman (2000) identified published studies from which elasticity values could be
derived for potentially fatal health effects. They suggest a triangular distribution with a mode of 0.40,
and endpoints at 0.08 and 1.00. In the Monte-Carlo simulation that assigns dollar values to benefits,
income elasticity values (E in the above equation) are drawn from this probability distribution. Based on
this formula and inputs, income elasticity factors are computed and applied to avoided-death benefits in
future years. At the average income elasticity value (0.49), the income elasticity factors applied range
from 1.213 (2008) to 1.445 (2029).
12 Ideally, income elasticity and income growth measurements would be calculated using real per capita
personal income growth. Real per capita GDP, however, is used as a proxy for real per capita personal income
growth owing to lack of appropriate data projections for real personal income growth. Historical data suggests that
GDP and personal income grow at similar rates (i.e., Table B-31 of the 2002 Economic Report of the President
shows that both real per capita GDP and disposable personal income grew at an average annual rate of 2.3 percent
between 1959 and 2000).
13 See Appendix A of Kleckner and Neuman (2000) for additional information on the derivation and
application of income elasticity adjustments.
14 The distribution of VSL values used in this EA is derived based on a meta-analysis of 26 different VSL
studies, all representing different year price levels. These price levels were updated to a common 1990 price level as
part of the analysis in "The Benefits and Costs of the Clean Air Act, 1970-1990" (USEPA, 1997c), from which the
distribution used in this EA is taken.
Economic Analysis for the October 2006
Final Ground Water Rule 5-77
-------
The second type of adjustment for income growth is applied to the portion of future benefits
derived from the value of lost time. The methodology here is more straightforward. The same per capita
GDP values referenced above (and presented in Appendix B) are employed to compute simple ratios
(income growth factors) between the future year and the year 2003 (the baseline year for COI
calculations). Lost time benefits values are then multiplied by these factors, which range from 1.15 in
2008 to 1.64 in 2029.
5.3.1.5 Present Value of Future Benefits
To allow comparison of future streams of costs and benefits, it is common practice to adjust both
streams to a present value (PV) using a social discount rate. This process takes into account the time
preference that society places on expenditures and benefits and allows comparison of cost and benefit
streams that vary over a given time period.15 A present value for any future period can be calculated
using the following equation:
PV = Vt/(l+R)t
where: t = the number of years from the reference period (year 0 of the benefits stream)
R = social discount rate
Vt = the benefits occurring t years from the reference period
There is much discussion among economists of the proper social discount rate to use for policy
analysis. For this EA, therefore, PV calculations are made using two social discount rates thought to best
represent current policy evaluation methodologies, 3 and 7 percent. Historically, the use of 3 percent is
based on rates of return on relatively risk-free investments, as described in the Ex anteGuidelinesfor
Preparing Economic Analyses (USEPA, 2000e). The rate of 7 percent is a recommendation of the Office
of Management and Budget (OMB) as an estimate of "before-tax rate of return to incremental private
investment" (USEPA, 1996c). To allow evaluation on an annual basis, the total PV of benefits are
annualized using the same social discount rates.
5.3.2 Summary of Quantified Benefits of GWR
The risk assessment methodology described in this chapter estimates quantified benefits of
reducing endemic acute infections caused by a portion of Type A and Type B viruses. Exhibits 5.23a-b
provide a summary of the cumulative monetary benefits estimated for the GWR for all system sizes and
categories (CWSs, NTNCWSs, and TNCWSs). Nonqualified benefits are discussed in Section 5.4.
The costs for rule alternatives are presented in Chapter 6, and cost/benefit comparisons are evaluated in
Chapter 8.
See EPA's Guidelines for Preparing Economic Analyses (USEPA, 2000e) for a full discussion of the use of social
discount rates in the evaluation of policy decisions.
Economic Analysis for the October 2006
Final Ground Water Rule 5-78
-------
Exhibit 5.23a Annualized Quantified Benefits of Illnesses and Deaths Avoided,
Final Rule, Enhanced COI, All Systems by System Size and Type
($Millions, 2003)
System Size
(Population Served)
3% Discount Rate
Mean
Value
90 Percent
Confidence Bound
Lower
(5th %ile)
Upper
(95th %ile)
7% Discount Rate
Mean
Value
90 Percent
Confidence Bound
Lower
(5th % Me)
Upper
(95th %ile)
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
1 00,001 -1M
>1 Million
All Sizes
$ 0.2
$ 0.4
$ 0.4
$ 1.0
$ 2.5
$ 2.4
$ 4.7
$ 4.3
$ 0.0
$ 16.0
$ 0.1
$ 0.2
$ 0.1
$ 0.3
$ 0.9
$ 0.8
$ 1.5
$ 1.5
$ 0.0
$ 5.4
$ 0.5
$ 1.0
$ 1.0
$ 2.4
$ 5.7
$ 5.3
$ 11.0
$ 10.1
$ 0.0
$ 37.0
$ 0.2
$ 0.4
$ 0.3
$ 0.8
$ 2.1
$ 2.1
$ 4.1
$ 3.8
$ 0.0
$ 13.7
$ 0.0
$ 0.1
$ 0.1
$ 0.3
$ 0.7
$ 0.7
$ 1.3
$ 1.3
$ 0.0
$ 4.6
$ 0.3
$ 0.8
$ 0.8
$ 2.0
$ 4.8
$ 4.6
$ 9.6
$ 8.8
$ 0.0
$ 31.6
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
1 00,001 -1M
>1 Million
All Sizes
$ 0.1
$ 0.2
$ 0.2
$ 0.2
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$
$ 0.9
$ 0.0
$ 0.1
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.3
$ 0.1
$ 0.4
$ 0.5
$ 0.6
$ 0.3
$ 0.2
$ 0.1
$ 0.1
$
$ 2.2
$ 0.1
$ 0.2
$ 0.2
$ 0.2
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$
$ 0.8
$ 0.0
$ 0.1
$ 0.0
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.2
$ 0.1
$ 0.4
$ 0.4
$ 0.5
$ 0.2
$ 0.1
$ 0.0
$ 0.1
$
$ 1.8
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
1 00,001 -1M
>1 Million
All Sizes
TOTAL
$ 0.6
$ 0.9
$ 0.4
$ 0.3
$ 0.2
$ 0.2
$ 0.0
$ 0.1
$
$ 2.7
$ 19.7
$ 0.2
$ 0.3
$ 0.1
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$
$ 0.8
$ 6.5
$ 1.4
$ 2.2
$ 1.0
$ 0.8
$ 0.4
$ 0.4
$ 0.1
$ 0.1
$
$ 6.2
$ 45.4
$ 0.5
$ 0.8
$ 0.3
$ 0.3
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$
$ 2.3
$ 16.8
$ 0.1
$ 0.2
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.7
$ 5.5
$ 1.1
$ 1.8
$ 0.8
$ 0.6
$ 0.3
$ 0.3
$ 0.0
$ 0.1
$
$ 5.1
$ 38.6
Detail may not add to totals due to independent rounding. The Traditional COI only includes valuation for
medical costs and lost work time (including some portion of unpaid household production). The Enhanced COI
also factors in valuations for lost personal time (non-work time) such as childcare and homemaking (to the extent
not covered by the Traditional COI), time with family, and recreation, and lost productivity at work on days when
workers are ill but go to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified
benefits are expected to comprise a significant portion of the overall benefits of the Rule and are presented in
Section 5.4.
Source: Appendix C
Economic Analysis for the
Final Ground Water Rule
October 2006
5-79
-------
Exhibit 5.23b Annualized Quantified Benefits of Illnesses and Deaths Avoided,
Final Rule, Traditional COI, All Systems, by System Size and Type
($Millions, 2003)
System Size
(Population Served)
3% Discount Rate
Mean
Value
90 Percent
Confidence Bound
Lower
(5th % Me)
Upper
(95th %ile)
7% Discount Rate
Mean
Value
90 Percent
Confidence Bound
Lower
(5th % Me)
Upper
(95th %ile)
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
10,001-SOK
50.001-100K
1 00,001 -1M
>1 Million
All Sizes
$ 0.1
$ 0.2
$ 0.2
$ 0.5
$ 1.3
$ 1.2
$ 2.4
$ 2.2
$ 0.0
$ 8.2
$ 0.0
$ 0.1
$ 0.0
$ 0.1
$ 0.3
$ 0.3
$ 0.5
$ 0.5
$ 0.0
$ 1.9
$ 0.3
$ 0.6
$ 0.6
$ 1.4
$ 3.4
$ 3.2
$ 6.7
$ 6.0
$ 0.0
$ 22.3
$ 0.1
$ 0.2
$ 0.2
$ 0.4
$ 1.1
$ 1.1
$ 2.1
$ 1.9
$ 0.0
$ 7.1
$ 0.0
$ 0.0
$ 0.0
$ 0.1
$ 0.3
$ 0.3
$ 0.4
$ 0.5
$ 0.0
$ 1.6
$ 0.2
$ 0.5
$ 0.5
$ 1.2
$ 2.8
$ 2.8
$ 5.9
$ 5.3
$ 0.0
$ 19.1
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
10,001-SOK
50.001-100K
1 00,001 -1M
>1 Million
All Sizes
$ 0.0
$ 0.1
$ 0.1
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$
$ 0.5
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.1
$ 0.1
$ 0.3
$ 0.3
$ 0.3
$ 0.1
$ 0.1
$ 0.0
$ 0.1
$
$ 1.3
$ 0.0
$ 0.1
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.4
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.1
$ 0.1
$ 0.2
$ 0.2
$ 0.3
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$
$ 1.0
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
10,001-SOK
50.001-100K
1 00,001 -1M
>1 Million
All Sizes
TOTAL
$ 0.3
$ 0.4
$ 0.2
$ 0.2
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$
$ 1.3
$ 10.0
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.3
$ 2.2
$ 0.7
$ 1.2
$ 0.6
$ 0.4
$ 0.2
$ 0.2
$ 0.0
$ 0.1
$
$ 3.4
$ 27.0
$ 0.3
$ 0.4
$ 0.2
$ 0.1
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$
$ 1.1
$ 8.6
$ 0.1
$ 0.1
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$ 0.0
$
$ 0.2
$ 1.9
$ 0.6
$ 1.0
$ 0.5
$ 0.4
$ 0.2
$ 0.2
$ 0.0
$ 0.1
$
$ 2.8
$ 22.9
Detail may not add to totals due to independent rounding. The Traditional COI only includes valuation for
medical costs and lost work time (including some portion of unpaid household production). The Enhanced COI
also factors in valuations for lost personal time (non-work time) such as childcare and homemaking (to the extent
not covered by the Traditional COI), time with family, and recreation, and lost productivity at work on days when
workers are ill but go to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified
benefits are expected to comprise a significant portion of the overall benefits of the Rule and are presented in
Section 5.4.
Source: Appendix C
Economic Analysis for the
Final Ground Water Rule
October 2006
5-80
-------
5.3.3 Quantified Benefits to Sensitive Subpopulations
Exhibits 5.23a and 5.23b presented the annualized benefits of the GWR across the general
population. However, a large portion of these benefits will be realized by sensitive subpopulations. As
described previously in section 5.2.2.2, EPA is considering three sensitive populations for the purpose of
the GWR: the immunocompromised, children (< 5 years old), and the elderly (> 65 years old). Exhibit
5.24 presents the benefits of the GWR to sensitive populations.
Exhibit 5.24 Annual Illnesses and Deaths Avoided at Full Implementation,
and Quantified Benefits of the GWR in Sensitive Populations
Population
US Census1
Population
Potentially
Affected
(served by
GW systems)
No. of
Illnesses
Avoided
No. of
Deaths
Avoided
Annual Benefits
Using 3 %
Discount Rate
and Enhanced
COI2
($Millions)
Immunocompromised (0.3%)
All ages
844,266| 342,772| 126| 0.002| $0.62
Nonimmunocompromised Sensitive (99.7%)
Elderly ( >65 yrs)
Children ( <5 yrs)
Children ( <5 yrs) Type A:
<2yrs
2-4 yrs
Children ( <5 yrs) Type B:
<1 month
1 month - <1 yr
1 -4 yrs
Total Nonimmunocompromised Sensitive
Total
34,791,627
19,079,280
7,587,829
11,491,436
315,541
3,470,952
15,292,787
53,870,907
54,715,173
14,125,401
7,746,187
3,080,659
4,665,523
128,110
1,409,207
6,208,872
37,363,958
37,706,730
5,559
2,780
1,684
904
3
37
151
8,339
8,465
0.10
0.06
0.01
0.01
0.00
0.01
0.03
0.15
0.15
$1.55
$4.99
$2.81
$1.51
$0.01
$0.13
$0.52
$7.17
$7.79
Notes: Detail may not sum due to independent statistical analyses and rounding. The figures presented in this
exhibit represent only the quantifiable benefits of the GWR. The nonquantified benefits are expected to comprise a
significant portion of the overall benefits of the Rule and are presented in Section 5.4. The Immunocompromised
population includes bone marrow transplant recipients, AIDS patients, and organ transplant patients.
1 The U.S. Census data is modified to show the number of people that are estimated to be immunocompromised (0.3
% of the population) and not immunocompromised (99.7% of the population). Therefore, the U.S. Census population
estimates shown above for the Elderly and Children categories are 99.7% of the estimates shown in Exhibit 5.4.
2The Enhanced COI factors in valuations for lost personal time (non-work time) such as childcare and homemaking
(to the extent not covered by the Traditional COI), time with family, and recreation, and lost productivity at work on
days when workers are ill but go to work anyway. (The Traditional COI only includes valuation for medical costs and
lost work time (including some portion of unpaid household production).
Source: U.S. Census Population from 2000 Census data (U.S. Bureau of the Census, 2000); Population Potentially
Affected derived from SDWIS (USEPA 2003a); Number of Illnesses Avoided, Deaths Avoided, and Annual Benefits
from GWR Model Output.
Economic Analysis for the
Final Ground Water Rule
October 2006
5-81
-------
5.4 Nonquantified Benefits of GWR Provisions
Due to the limited availability of data, EPA was only able to quantify some of the benefits it
believes are associated with the GWR. The EA for the GWR monetizes benefits associated with Type A
viruses represented by rotavirus data and Type B represented by enterovirus or echovirus data. As
discussed in Section 5.1., the EA quantifies only the endemic, acute illnesses due to rotavirus and
enterovirus. Other benefits, not quantified, are discussed in this section. As discussed in Section 5.2 and
5.3, even in the case where benefits are quantified, data limitations remain and certain factors used in the
analysis have significant associated uncertainties.
The nonquantified health benefits are: 1) decreased incidence of gastroenteritis caused by other
Type A viruses such as norovirus, astrovirus and adenovirus, 2) decreased incidence of other acute
disease endpoints (e.g., hepatitis, conjunctivitis), 3) decreased incidence of chronic illness sequelae
associated with Type B virus (e.g., diabetes, dilated cardiomyopathy, hypertension and reduced kidney
function), 4) decreased incidence of illness and death caused by bacteria, 5) decreased incidence of
waterborne disease outbreaks and epidemic illness, and 6) decreased illness through minimizing
treatment failures or fewer episodes with inadequate treatment.
The nonquantified non-health benefits are: 1) improved perception of ground water quality and
perception about reduced risk associated with PWS wells, 2) reduced use of bottle water and point-of-use
devices, 3) reduced time spent on averting behavior such as obtaining alternative water supplies, and 3)
avoided costs associated with outbreak response.
EPA believes that, collectively, these benefits, both health and non-health, are likely to
substantially exceed those which EPA was able to quantify, and are the primary basis for supporting the
preferred regulatory alternative. Each of these major nonquantified benefits is discussed further below.
5.4.1 Decreased Incidence of Illness Caused by Other Type A Viruses
5.4.1.1 Norovirus
Noroviruses have been suggested as "the single most significant cause of intestinal infectious
disease in the developed world." (Carter, 2005) However, no cell culture method exists to recover
noroviruses in stool or environmental samples, therefore published human infectivity data for norovirus
were not available to use in the EA. This detection method deficiency and resulting lack of published
infectivity data prevented quantification of monetized benefits for norovirus occurrence reduction in the
GWR EA. In this section, the norovirus disease burden is qualitatively discussed together with
information that suggests additional benefit from full implementation of the GWR.
Economic Analysis for the October 2006
Final Ground Water Rule 5-82
-------
Norovirus Disease in the United States
In recent years, numerous common source outbreaks16 have been attributed to norovirus
contamination. CDC retrospectively evaluated 4,050 U.S. common source outbreaks that occurred
during 1998—2000 and determined that at a minimum, 28% could be attributable to norovirus (Turcios, et
al, 2006). In another study of outbreaks occurring during the period from 2000 to 2004, fecal samples
from 226 outbreaks (12 waterborne) of acute gastroenteritis suspected of calicivirus causality were
analyzed by CDC for norovirus and other caliciviruses (Blanton et al, 2006). Caliciviruses (primarily
norovirus but also sapovirus) were detected in 184 (81%) of those outbreaks. These data suggest that
norovirus represents a large component of the total gastroenteritis epidemic disease burden in the United
States. This perspective is further supported by CDC who estimated total U.S. annual illnesses from
Norwalk-like (now called norovirus) viruses of 23,000,000 cases (Mead et al 1999).
The identification of the etiologic agent of a waterborne, or any common source or person-to-
person outbreak depends on a number of factors: the timely recognition of the outbreak and timely stool
sample collection, laboratory capacity and capability, active requests for non-routine tests, and many
other factors (Lee et al., 2002). When routine analysis of stool is requested, the sample is typically tested
only for some enteric bacteria and some protozoan parasites (typically Giardid); testing for viruses and
especially norovirus, is not common. Viral agents are less easily identified than most enteric bacteria and
norovirus is among the most difficult viral agents to identify (Blanton et al., 2006). It is only within the
last ten years that molecular methods have become sufficiently advanced so norovirus strains can be
specifically identified in stool samples. Unlike bacterial agents, no routine, commercially available test
for identifying norovirus exists and therefore ill persons who submit a stool specimen are not routinely
tested for norovirus (Blanton et al., 2006).
Norovirus has a low infectious dose, prolonged asymptomatic shedding, substantial strain
diversity and considerable environmental stability (CDC, 2001a). As a result of a low infectious dose, a
small dose readily allows infection in exposed individuals (CDC, 200la). Prolonged shedding allows
opportunity for individuals to come in contact with norovirus. Norovirus does not provide lasting
immunity upon infection partly because there are many different serotypes (Carter, 2005). The
substantial strain diversity may prevent individuals from acquiring a general immunity that prevents
illness from any norovirus. Environmental stability allows norovirus to survive in ground water and
elsewhere in the environment until acquired by another host (Schwab and Bae, 2005). Thus, norovirus-
contaminated ground water or a norovirus-contaminated surface will continue to maintain infectious
norovirus. Individuals can acquire primary norovirus infection by drinking that ground water or can
acquire secondary infectious norovirus by contact with a contaminated surface or persons.
Although a small percentage (about 20%) of individuals in the United States are likely genetically
immune to norovirus infection (Lindesmith et al., 2003), the remaining population is subject to repeated
episodes of infection and illness. Norovirus is shed in appreciable numbers, at concentrations similar to
enteroviruses (Carter, 2005). Because norovirus is highly infectious (Moe et al., 2001), individuals may
easily acquire infection. Once infected, norovirus is easily transmitted to others. Individuals can easily
16Common source outbreaks arise primarily from food or water (ice is considered to be water by EPA but is
commonly treated as food in outbreak compilations). Propagation by secondary (person-to-person) transmission
may also occur but is not the immediate cause of the outbreak. Outbreaks that are not common source arise and
propagate only by person-to-person transmission.
Economic Analysis for the October 2006
Final Ground Water Rule 5-83
-------
spread infection and illness to family members and others (both children and adults) outside the
household by casual contact with asymptomatic carriers who shed for long periods.
Using molecular epidemiology tools, diverse cases may be tracked backwards in time and
exposure history to identify an index case or a common source exposure. One norovirus strain, the
Farmington Hills strain, is identified as a cause of multiple large outbreaks on cruise ships (Widdowson et
al., 2004). The norovirus attack rates (30% of passengers became ill) resulting during cruise ships
outbreaks suggests a defining character, it's capability to spread rapidly and efficiently via secondary
transmission from a primary infected individual. This norovirus characteristic has significance when
estimating the number of secondary cases which is illustrated further in Appendix J.
Norovirus Ground Water Outbreaks
Norovirus is recognized as the confirmed cause for some ground water outbreaks and is suspected
in many others. In general, the etiological agent of many common source gastroenteritis outbreaks has
not been identified. For example, of greater than 2,500 outbreaks reported to CDC from 1993 to 1997,
68% were of "unknown etiology" (Widdowson et al, 2005). Similarly, most common source outbreaks
resulting from ground water fecal contamination are of unknown etiology. Of 342 ground water
outbreaks in the United States between 1971 and 1994, 212 were of "undetermined" etiology (Craun and
Calderon, 1997).
Two recent ground water outbreaks in Wyoming were both recognized as norovirus outbreaks. In
South Pass, Wyoming, an epidemiological investigation linked illness to the well water and ice
(Parshionkar et al, 2003). A total of 84 illnesses were identified. The well supplied a TNC PWS system
producing water from a sensitive (fractured bedrock) aquifer. A chlorination device malfunctioned due to
poor maintenance. In Big Horn, Wyoming, 35 illnesses were identified due to fecal contamination of an
unchlorinated PWS well in a sensitive (fractured bedrock) aquifer (Anderson et al, 2003). In 1989,
norovirus was identified in an outbreak with 110 illnesses in Sedona, AZ (Lawson et al, 1991). Most
recently, norovirus illness and campylobacterois were the most common illnesses associated with fecal
contamination of PWS wells on South Bass Island, OH in 2004 (Ohio EPA, 2005). It is estimated that
about 1450 individuals became ill from consuming well water tapping a sensitive aquifer with widespread
fecal contamination (many PWS as well as some private wells were fecally contaminated). Other
norovirus ground water outbreaks are identified from Yukon, Canada (Beller et al., 1997), Braun Station,
TX (D'Antonio et al., 1985), Pierce County, WA (Taylor et al, 1981), Onondaga, NY (Chatterjeee et al.
2004), Monroe County, PA (Wilson et al., 1982), Henderson County, IL (Parsonnet et al., 1989),
southeastern PA (Cannon et al, 1991), South Dakota (CDC, 1988) and elsewhere. Among outbreaks that
have been specifically attributed to norovirus exposure in ground water, most have occurred in sensitive
aquifers (see the GWR Occurrence and Monitoring Background Document for a discussion of outbreaks
in sensitive aquifers).
Norovirus Disease Burden and Severity
The health effects of norovirus illness include acute onset of nausea, vomiting, abdominal
cramps, and diarrhea. Vomiting is relatively more prevalent among children. Many adults experience
vomiting as well as diarrhea. Constitutional symptoms (e.g., headache, fever, chills, and myalgia) are
frequently reported. Although rare, severe dehydration caused by norovirus gastroenteritis can be fatal,
with this outcome occurring among susceptible persons (e.g., older persons with debilitating health
Economic Analysis for the October 2006
Final Ground Water Rule 5-84
-------
conditions). No long-term sequelae of norovirus infection have been reported (CDC, 200la). Duration of
illness is typically 12-60 hours.
Norovirus is suggested in waterborne disease when vomiting is a symptom because norovirus is
fairly unique among the common gastrointestinal agents in that vomiting often accompanies diarrhea. In
a prospective study of Dutch patients with norovirus in a community cohort (not an outbreak), 74% of 99
patients of all ages experienced vomiting. Vomiting enhances the ability of norovirus to spread
efficiently via fomites, suspended particles and aerosols.
The avoided illnesses from Type A viruses determined quantitatively in this EA are based only on
rotavirus infection and illness. There are significant differences between rotaviral and noroviral disease.
The total ground water-borne disease burden would change significantly if norovirus were included in the
quantitative analysis. EPA believes that, if norovirus was included in the quantified benefits, there would
be significantly greater total benefits for two reasons. First, norovirus, unlike rotavirus, is a disease of
older teenagers and adults (Carter, 2005) as well as of children. In a study of 1,484 norovirus patients,
the mean age was 43 years (Fankhauser et al. 2002). Similarly, in a study of 1,010 norovirus cases, the
median age was 47 years (Blanton et al, 2006). This distinction is important; the monetized rotavirus
disease burden in this EA provides only a small benefit for adult rotaviral disease because most adults are
immune to rotavirus. Thus, noroviral disease in adults, which is quite prevalent, would be a significant
additional avoided illness and monetized benefit if norovirus were included in the quantified analysis in
this EA.
Second, unlike rotavirus, adults with norovirus experience vomiting and some experience only
vomiting (CDC, 2001a). Most individuals experienced vomiting in a norovirus outbreak. In an analysis
of symptoms, Widdowson et al (2005) found that more than half the ill persons experience vomiting in
almost all norovirus outbreaks (i.e., in 86 percent of 136 outbreaks, greater than 50 percent of people
experienced vomiting as a clinical symptom).
In monetizing the benefits associated with rotavirus or norovirus, the clinical symptoms are
important. Norovirus is unique among the waterborne etiologic agents in that large proportions of ill
individuals experience vomiting. Because adults with rotavirus experience gastroenteritis only, EPA
assumes that each adult ill with rotavirus has one day of lost productivity and a subset (16%) have one
lost patient day. In contrast, for adults with norovirus illness that manifests as vomiting (with nausea and
gastroenteritis), EPA believes that norovirus likely produces greater lost leisure time and more lost
productivity days, and therefore lost patient days, as compared with rotavirus because vomiting is so
debilitating. Thus, a case of noroviral disease, when monetized, would likely produce greater benefit than
rotaviral disease in adults.
In an analysis of care, Widdowson et al. (2005) compiled data on 3,370 persons affected in 112
norovirus outbreaks in the United States. Of these, 329 (10%) sought care from a physician and 33 (1%)
were hospitalized. Because these data are from outbreak investigations, the outbreak cases may be biased
toward the more severe end of the disease spectrum. It has been hypothesized that outbreaks are more
likely recognized when the disease is more severe. Conversely, endemic cases or unrecognized outbreak
cases may be milder disease. Mead et al. (1999) estimated norovirus hospitalization rates at 0.2% rather
than the 1% value determined from outbreaks.
This EA quantifies the benefits for Type A viruses using data only from rotavirus. Unlike
rotavirus, the norovirus human challenge study data to determine norovirus infectivity are not yet
Economic Analysis for the October 2006
Final Ground Water Rule 5-85
-------
published (Moe, 2001). For this reason alone, rotavirus data are used to quantitatively determine the
benefits for the Type A viruses. As discussed above, EPA believes that the more severe clinical
symptoms associated with norovirus, as compared with rotavirus, suggests that the norovirus disease
burden associated with PWS wells is important, especially for adults.
EPA believes that the disease burden for norovirus and rotavirus is similar, although children are
more likely to experience vomiting from norovirus. However, vomiting in children does not result in
additional benefits when the illness is monetized. In contrast, the person-days ill metric does not account
for the clinical symptoms and disease severity in adults who experience those person-days of illness and
this severity provides additional benefits when monetized. EPA believes that consideration of the more
severe clinical symptoms associated with norovirus illness among adults implies that adults will have
more lost-patient days, as well as more lost productivity days.
Norovirus Cost of Illness
EPA believes that the clinical severity of norovirus illness in adults is significant. While EPA has
data on the proportion of cases with vomiting and on the duration of illness, EPA has no data on duration
of vomiting. Consideration of vomiting for norovirus illness suggests that the lost productivity for
norovirus illness in adults could be much greater than is assumed for rotavirus in the quantified benefits
because vomiting is typically sufficiently debilitating (and infectious to others) that most people cannot
(or should not) work. The EA currently assumes a weighted mean value of 0.16 lost patient days and one
day of lost productivity for adults when ill from rotavirus. EPA believes that this value is low compared
with the likely norovirus illness lost productivity and lost patient days.
5.4.1.2 Other Type A viruses
Hepatitis A Virus
Viral pathogens, other than the enteroviruses, rotavirus and the caliciviruses are also transmitted
by the fecal-oral pathway and thus are potential etiologic agents for ground waterborne disease. Hepatitis
A (HAV) virus is the only waterborne virus that is reportable to CDC. About 28,000 HAV cases are
reported to CDC each year, although that number is expected to decline with time because children are
currently vaccinated for HAV in high risk states and newer recommendations are for increased
vaccination coverage. More generally, however, because HAV is more severe as an adult disease, an
aging U.S. population may have greater disease burden. Mead et al (1999) estimate about 83,000 HAV
cases each year, with a hospitalization rate of 13% and a mortality rate of 0.3%. Like all fecal/oral
pathogens, HAV is acquired through a variety of pathways. Hepatitis has a long incubation period and
the virus remains viable in the environment and especially ground water for months. Thus, the infection
source is often obscure. Ground water HAV outbreaks have been identified (Georgetown, TX (Hejkal et
al, 1982), Racine, MO (Missouri Department of Health, 1992); Lancaster, PA (Bowen and McCarthy,
1983); Quebec, Canada (De Serres et al., 1999); all in sensitive aquifers). HAV is not favored in typical
cell lines used in cell culture and so identification is difficult in environmental samples. As a result of the
relatively large hepatitis A disease burden in locales where hygiene is poor, hepatitis A virus is believed
to be highly infectious, perhaps as infectious as norovirus or rotavirus. HAV tends to present more
serious symptoms than other viruses discussed thus far. Unlike norovirus or rotavirus, HAV infection
confers lifetime immunity. No human dose response study data are available at differing doses. Because
Economic Analysis for the October 2006
Final Ground Water Rule 5-86
-------
of this and the fact that HAV cannot be routinely identified in well water, the benefits of avoiding HAV
disease through full implementation of the GWR cannot be quantified.
Hepatitis E virus
Hepatitis E (HEV) virus is another fecal oral virus that is potential transmitted via ground water.
Using serology and case histories of individual patients, HEV is established as endemic within the United
States (Tsang et al, 2000). However, the data suggest that only one or a few percent of the population has
been infected. Unlike HAV, no ground water outbreaks of HEV have occurred in the United Sates,
although they have occurred elsewhere (China and Somalia). HEV is not culturable and no data on
environmental occurrence in ground water are available. The disease is severe (up to 20% mortality
among pregnant women in developing countries) and thus no human dose response data are available.
Because infectious HEV cannot be identified in well water and no human dose response data are
available, the benefits of avoiding HEV disease through full implementation of the GWR cannot be
quantified.
Adenovirus
The adenovirus are a large group of viruses that produce diverse symptoms. Two adenovirus
serotypes adenovirus 40 and 41 produce primarily enteric symptoms, but several other adenoviruses are
also capable of producing such symptoms. Some cause conjunctivitis. Most significantly, adenoviruses
caused a fatal outcome in otherwise health young males in military settings. (CDC, 200Ib) All
adenoviruses, no matter the infection site and characteristic illness, are shed copiously through the gut and
are thus fecal\oral viruses (Carter, 2005). Adenoviruses are not efficiently recovered using commonly
available cell lines and methods and no human dose response data are available. No waterborne disease
outbreaks have been reported for adenovirus in the United States, but adenovrius was among several
pathogens identified in wells sampled during the South Bass Island, OH outbreak (CDC, 2005). The
GWR does not quantify the benefits from avoided adenovirus illness.
Astrovirus
Astrovirus, like rotavirus, is commonly acquired in child care settings, causes mild disease in
children and most children are exposed at an early age. However, like rotavirus, a small percentage of
that large population suffer more significant health effects and may require in-patient care. Like
rotavirus, the disease burden in older children and adult populations is underestimated because the disease
is mild (Carter, 2005). Astroviruses are shed in stool at large numbers (similar to enteroviruses Carter,
2005) and, in France, prospective epidemiology studies have implicated untreated ground water as a route
of infection (Gofti-Laroche, et al, 2003). Astroviruses are not favored for recovery in environmental
samples using common cell lines and no human dose-response data exist. No waterborne disease
outbreaks have been identified for Astrovirus in the United States. The benefits from avoiding astrovirus
illness via the GWR can only be qualitatively be described.
Reovirus
Reovirus is recovered in environmental samples using the BGM cell line and is commonly found
co-occurring with the enteroviruses in PWS wells. Carducci et al (2002) found that, in some cases,
enterovirus detection was limited because reovirus reproduction was so highly favored. Reovirus is more
closely related to rotavirus and thus has some similar characteristics. Unlike rotavirus, reovirus rarely
Economic Analysis for the October 2006
Final Ground Water Rule 5-87
-------
causes disease but it is now recognized as a human pathogen in children (Tyler et al, 2004). Although
reovirus is probably not a significant component of the total disease burden, it is important because it
likely reduces the enterovirus recovery efficiency17, which is typically not greater than 50 percent under
optimal conditions. No waterborne disease outbreaks have been identified for Reovirus in the United
States.
5.4.2 Decreased Incidence of Other Illness Caused by Type B Viruses
Other Type B viral diseases, such as diabetes or cardiomyopathy, may lead to chronic disease.
Because the causal relationship is not well established and the number of cases associated with drinking
water is unknown, the Agency was not able to quantify benefits from the GWR on reducing chronic
diseases. While this EA does not quantify in dollar terms the benefit of avoiding chronic illness, this
section discusses the potential benefits qualitatively and illustrates the significance of these secondary
benefits.
The total number of people with two types of chronic illnesses, diabetes and heart disease in the
United States is substantial. Between 1990-92, there was an annual average of seven million people with
diabetes (all kinds) and four million people with chronic heart disease (including myocarditis and
cardiomyopathy) (Collins, 1997). Additionally, 3.5 percent of heart disease deaths in 1993 were due to
cardiomyopathy (NHLBI, 1996). The potential benefits of avoiding some of these health effects cannot
be overlooked, and may be significant.
An extensive literature review proved that costs of a single case of diabetes or heart disease are
significant. Cost estimates for a case of diabetes and a case of chronic myocarditis (using the cost per
case of an "average case of heart disease" as a proxy for chronic myocarditis) are presented below to
demonstrate the magnitude of potential benefits per avoided case of chronic illness. Potential
implications of diabetes, myocarditis, and cardiomyopathy through consumption of ground water are
briefly discussed below.
Diabetes
There is considerable information that Type 1 diabetes may be associated with enterovirus
infection, including infection with coxsackievirus and echoviruses (Maria et al. 2005; Vreugdenhil et al,
2000). Epidemiological studies have shown a strong correlation between diabetes and enterovirus
infection. Individual case studies have reported diabetes development after enterovirus infection
(Roivainen, 1998). A mechanism for producing disease suggests that entervirus infection triggers
autoimmunity response. While these data are suggestive, they are not definitive. Recently, Maria et al
(2005) reported simultaneous (on the same day) diabetes onset in mother and son coincident with
enteroviral infection. With these data, it is now clear that at least some Type 1 diabetes cases result from
enterovirus infection. The GWR EA considered the severe acute illnesses resulting from enterovirus
infection in quantitatively determining the benefits (illnesses avoided), but only considers qualitatively
17The presence of reovirus in environmental samples complicates the recovery, or detection and
quantification, of enteroviruses. Additionally, reovirus rarely results in illness. Therefore, reovirus occurrence can
cause underestimates of enteroviral occurrence, and can elevate viral concentrations without adding to the disease
burden associated with a contaminated water source. Hence, it can result in an underestimate of the quantified
benefits that accrue from the GWR.
Economic Analysis for the October 2006
Final Ground Water Rule 5-88
-------
the diabetes cases avoided as a result of corrective actions performed due to enterovirus contamination of
PWS wells.
The most comprehensive work regarding the economic burden of diabetes (both Type I and Type
II) in the United States was conducted for the American Diabetes Association. In their report "Economic
Consequences of Diabetes Mellitus in the United States in 1997," Fox et al. (1998) presented the direct
medical and indirect costs attributable to diabetes, as well as a total and per capita estimate of
expenditures of people with and without diabetes. Improving on their estimates and methodology from
their 1992 effort (Fox et al., 1993), this national prevalence-based COI study also compares the health
care expenditures of diabetics in 1997 to nondiabetics.
The authors created a holistic estimate of the health care expenditures attributable to diabetes in
1997 by including: 1) medical expenditures attributable to diabetes (i.e., the cost due to the excess
prevalence of diabetes related chronic complications and general medical conditions in people with
diabetes), and 2) total medical expenditures incurred among people with diabetes (i.e., the cost for all
services for people with diabetes). Annual per capita expenditure estimates were also calculated and
defined as the sum of the expenditures for diabetics in 1997, divided by the 1997 diabetic population.
The estimates do not, however, include pain and suffering nor do they include lost productive and leisure
time.
The per capita annual medical expenditures for people with diabetes was $13,092 for people with
diabetes versus $3,470 among people without diabetes. Therefore, the annual cost of diabetic care is
$9,622 per person.18 The annual net productivity loss for each person due to diabetes totaled $1,650 for
18-64 year olds and $528 for those 65 and older.19 These are sums of costs attributable to diabetes from
productivity loss from work, from restricted-activity and from bed-disability.
According to the 1980-1987 Hospital Cost and Utilization Project (HCUP), a national sample of
more than 500 hospitals that represent an unweighted 20 percent sample of discharges, the mean age of
diagnosis for a case of diabetes mellitus within the study was 53 (Elixhauser et al., 1993). Assuming that
a patient incurs treatment for diabetes each year throughout the duration of his expected life from age 53
(29.6 years)20, the present value estimate of the direct medical costs and indirect costs of illnesses would
be $227,032 using a 3 percent discount rate and $143,733 using a 7 percent discount rate. This figure
could be even greater if the cost of premature death or pain and suffering were incorporated. While this is
a simple approximation of the magnitude of a COI value for this illness, it captures the lifetime costs of
diabetes in those who survive the first year through their life expectancy period from the age of diagnosis.
1 8
Costs updated from January 1995 dollars to 2003 (annual) dollars using the CPI-U for "medical care
services" (= 278.8 - 219.8 = 1.3).
19 Costs up
177.1-159.1 = 1.1)
20 Life Tabl
1995." (NCHS, 1998)
19 Costs updated from January 1997 dollars to 2003 (annual) dollars using the CPI-U for "all items" (=
20 Life Tables. Table 6-3. "Expectation of Life at Single Years of Age, by Race and Sex: United States,
Economic Analysis for the October 2006
Final Ground Water Rule 5-89
-------
Myocarditis and Dilated Cardiomyopathy
Viral infection of the heart is relatively common and usually of little consequence (Kearney et al.,
2001). However, virus infection can lead to substantial cardiac damage and severe acute heart failure and
can also evolve into chronic heart failure (Kearney et al, 2001). Viral infection is the most common cause
of myocarditis. Viruses for which ground water is one possible route of exposure such as Coxsackie A
and B virus, echovirus and adenovirus can cause myocarditis (Kim et al, 2001; Magnani and Dec, 2006;
Huhn et al. 2005) with Coxsackie B virus the most often (of all viruses) associated with myocarditis
(Kearney et al. 2001). In addition to myocarditis, epidemiological studies from Finland have documented
an association between enterovirus infection and heart attacks (myocardial infarction) in men with no
prior evidence of heart disease (Reunanen et al, 2002).
Mean age of patients with active myocarditis is 42 years (Kearney et al., 2001). Sixty percent of
myocarditis patients had antecedent symptoms indicative of recent infection (Kearney et al. 2001).
Myocarditis accounted for 22% of sudden unexpected death under age 30 and 11% of those between 30
and 40. Mortality is 20% at one year and 56% at four years (unless transplant occurs) (Kearney et al,
2001) largely due to chronic heart failure (dilated cardiomyopathy).
Myocarditis is included in the quantified benefits as a severe illness that can be caused by
echovirus and other enterovirus infection. However, myocarditis can lead to chronic heart disease
(dilated cardiomyopathy) which, like all chronic disease sequelae, are not quantified in the EA benefits.
Myocarditis is a common cause of dilated cardiomyopathy, which in developed countries is the
underlying etiology in about 45% of patients undergoing heart transplants (Kearney et al. 2001). The EA
underestimates the benefits associated with preventing acute viral heart infections by coxsackievirus (and
adenovirus) because it focuses primarily on echovirus disease endpoints. Also and perhaps most
importantly, the EA quantifies only severe acute cases due myocarditis and neglects the chronic disease
sequelae associated with all enteroviral heart infections.
The annual direct COI associated with an "average case of heart disease" is estimated to be
$5,12921. This estimate is derived from data originally computed by of the National Center for Health
Statistics (NCHS) for "heart disease" (which includes International Class of Diseases, 9th Revision
(ICD-9) codes 391-398, 402, 404, 410-416, 420-429)(Hodgson, 1984; Hodgson, 1998). Since no
specific cost data were available for chronic myocarditis, or cardiomyopathy (ICD-9 code 425), annual
21 Cost updated from January 1995 dollars to 2003 (annual) dollars using the CPI-U for "medical care
services" (= 278.8 - 219.8 = 1.3).
Economic Analysis for the October 2006
Final Ground Water Rule 5-90
-------
per capita costs for an "average case of heart disease" were computed using data on "heart disease" in
conjunction with prevalence numbers from the 1995 National Health Interview Survey.22'23
Indirect costs of "other heart disease" were estimated by Cropper and Krupnick (1990) who used
information from the 1978 Social Security Survey of Disabled and Work to model the effects of disease
on labor participation and earnings. Cropper and Krupnick found that the annual indirect cost ranged
from $3,447 to $7,074 depending on the age of the individual and the age of illness onset.24 Again, it is
important to note that these costs do not include pain and suffering.
According to the 1980-1987 HCUP study of 500 hospitals, the mean age of diagnosis for
cardiomyopathy was 60 (Elixhauser et al., 1993). Using this diagnostic category as a proxy for chronic
myocarditis,25 the lifetime COI could be substantial. For example, the present value of both direct and
indirect costs for a patient with the condition would be $61,117 given an average life expectancy of 21.1
years (7 percent discount rate). This figure could be even greater if the costs of lost earnings and of
premature death were incorporated.
Flaccid Paralysis
Flaccid paralysis is a rare but severe consequence of enterovirus infection. Flaccid paralysis
occurs most commonly upon poliovirus infection (about 1 in 200 infections) but vaccination has
eliminated almost all poliovirus cases. However there is no vaccine to prevent flaccid paralysis from
echovirus, coxsackievirus or enterovirus 70 and 71 (Grimwood et al, 2003; Rotbart, 1995) infection.
Flaccid paralysis is more likely with echovirus than coxsackievirus but rare. For example, no flaccid
paralysis cases occurred as the result of an echovirus 18 outbreak in 2001 in 29 cases of viral meningitis
from (surface) drinking water at an Alaska camp (McLaughlin et al, 2004). In an echovirus 33 outbreak
in New Zealand in 2000, one healthy three-year old (of 75 infected persons) suffered flaccid paralysis
(Grimwood et al, 2003); two infants died (Huang, et al. 2003). McMinn et al. 2001 report two cases of
Guillain-Barre paralysis among 14 children neurologically ill from enterovirus 71 during a 1990 outbreak
in Western Australia. For the period 1970-1979, CDC reports 58 paralysis cases due to enteroviruses
with most cases due to echovirus (Moore, 1982). The EA does not explicitly include benefits associated
with preventing flaccid paralysis because the disease is so infrequent.
22 Chronic illness prevalence rates (cases per 1,000 individuals) for "heart disease" were multiplied by the
total U.S. population to obtain the total number of heart disease cases in 1995. The "average cost of heart disease"
per person in 1995 was subsequently calculated by dividing the total cost of heart disease in 1995 by the total
number of heart disease cases in 1995. Prevalence figures were from Current Estimates of the National Health
Interview Survey, 1995 (Benson and Marano, 1998), and the total U.S. population was obtained from the Census
Bureau.
23 Without more detailed information, this simplified method assumes that the cost of any heart disease,
whether ischemic or other, would be the same within this major disease group. This is a major limitation of these
estimates, as hospital costs for coronary heart disease may not be the same for hypertensive disease, for example.
24 Costs updated from January 1977 dollars to 2003 (annual) dollars using the CPI-U for "all items: (=
177.1-58.5 = 3.0).
25 Costs updated from January 1977 dollars to 2003 (annual) dollars using the CPI-U for "all items: (=
177.1-58.5 = 3.0).
Economic Analysis for the October 2006
Final Ground Water Rule 5-91
-------
5.4.3 Decreased Incidence of Bacterial Illness and Death
Although the EA does not quantify the benefits of bacterial illnesses that may be avoided by the
GWR, EPA believes that these benefits could be substantial. EPA believes that if a fecal indicator occurs
in a PWS well it demonstrates that a pathway exists for pathogens to travel from a fecal contaminant
source to the well. Furthermore, if such a pathway exists, EPA believes it is only a matter of time before
pathogens (viral or bacterial) excreted by infected animals or individuals reach the well. Such exposure
might then cause illness among populations using undisinfected water from the well. There is a larger
range for bacterial sources of contamination than sources of viral contamination because bacterial
pathogens that also cause illness in humans can originate from animal reservoirs (infected livestock)
whereas most viruses that cause illness in humans do not. Identifying and eliminating or treating such
sources of contamination under the GWR will result in reduced exposure to bacterial pathogens and
illness avoided. Bacterial contamination can occur in wells in both sensitive and non-sensitive aquifers.
However, because the bacteria are larger than the viruses, they are less mobile in non-sensitive aquifers
and are more likely to occur in wells in sensitive aquifers and in shallow wells in non-sensitive aquifers.
The nonquantified benefits of bacterial illness avoided could be much larger than that predicted for
viruses because the severity of some of the symptoms associated with bacterial illness are worse than
some of the symptoms of viral illness. The following is a discussion of different bacterial pathogens
found in ground water and their associated clinical symptoms if ingestion results in illness.
5.4.3.1 Bacterial Pathogen Occurrence
The bacterial pathogens can be divided into two groups; the frank pathogens such as E. coll
O157:H7 and the opportunistic pathogens, such as Pseudomonas aeruginosa. The opportunistic bacterial
pathogens found in drinking water have been well summarized by Rusin et al (1997) and will be
discussed later when exposure scenarios other than contaminated source water are considered. This
discussion will focus on the frank bacterial pathogens that are found in fecally contaminated ground
water. Unlike the viral pathogens, the frank bacterial pathogens can be excreted by animals (typically
birds and mammals) but some, such as Helicobacter Pylori andLegionellapneumophila are naturally
present in ground water.
E. coli
The group of bacteria known as E. coli contain both pathogenic and non-pathogenic isolates. The
most dangerous E. coli bacteria contain the gene for producing Shiga toxins. E. coli 0157:H7 is the most
widespread shiga-toxin producing E. coli but at least 81 serotypes have been identified (Prager et al,
2005). Release of toxins in the body can result in kidney failure, shock and death in otherwise healthy
individuals, especially small children. Typically, kidney failure occurs in 2-7% of illnesses. Death or
end-stage renal disease occurs in about 12% of patients four years after diarrhea-associated kidney failure
(Garg et al, 2003). Twenty five percent of kidney failure survivors demonstrate long-term renal sequelae
(Garg et al, 2003). For patients with moderate and severe gastroenteritis caused by E. coli, long-term
study shows that they have an increased risk of hypertension and reduced kidney function (Garg et al.
2005). CDC estimates that drinking water is responsible for 15% of the 73,000 illnesses each year from
E. coli O157:H7 in the United States (Rangel, 2005).
Ground water outbreaks due to E. coli O157:H7 are prominent because of the fatal outcomes
associated with those outbreaks. In Walkerton, Ontario, 6 individuals died and 27 developed kidney
Economic Analysis for the October 2006
Final Ground Water Rule 5-92
-------
failure from ground water contaminated with E. coll O157:H7 (and Campylobacter jejuni) (Health
Canada, 2000). The aquifer was sensitive (karst) and fecal contamination often occurred. The outbreak
coincided with a reduction in chlorination treatment coincident with a large fecal contamination event. In
Washington County, NY, two individuals (including an otherwise healthy two year old child) died from
E. coll O157:H7 contamination of a county fair water supply system. The aquifer was thin and shallow
but not sensitive. (The fair ground's water supply system was not recognized as a PWS system because it
was used for less than 60 days each year.) Four individuals died in Cabool, MO due to E. coll O157:H7
(Swerdlow et al, 1992). This aquifer was sensitive (karst) and multiple illnesses (above normal levels)
occurred before a water main break so the outbreak was likely due to source water contamination.
Another E. coll O157:H7 outbreak occurred in ground water in Minnesota but no kidney failures resulted.
Although manure is often considered to be the source of shiga-toxin producing E. coll, they have
also been isolated from municipal sewage (Holler et al, 1999). E. coll O157:H7 was found to survive on
a pasture surface for almost 4 months. About 4-15% of cases are acquired via secondary transmission
(Parry and Salmon, 1998). In addition to the shiga-toxin producing E. coll, there are a substantial number
of other pathogenic E. coll bacteria, mostly through production of other toxins. Hunter (2003) identifies
62 E. coll strains capable of causing diarrheal disease. Little data are available on the hazard associated
with waterborne transmission for most of the pathogenic E. coll other than E. coll O157:H7.
Although E. coll is monitored in PWS wells via the Total Coliform Rule, the EPA-approved
methods are not capable of identifying the shiga-toxin producing E. coll. Because E. coll are often found
in ground water supplies and because low doses can result in infection and secondary transmission is
significant, it is likely that disease due to E. coll is quite prevalent in association with PWS wells. Few
data on pathogenic E. coll in PWS wells are available. Dose response data are limited to a single point
value from a carefully documented outbreak in Japan. This point value suggests that the infectious dose
is very low compared with most pathogenic bacteria. Because few data are available on the occurrence of
pathogenic E. coll in PWS wells and dose response data are limited, the benefits of avoided illness and
chronic sequelae from the GWR can only be qualitatively discussed.
Shigella
Shigella bacteria are distinct because they are often associated with bloody diarrhea (bacillary
dysentery). The enterohemorraghic E. coll bacteria acquired the capability to produce toxins by
exchanging plasmids with Shigella. Thus, Shigella often also cause kidney failure and chronic kidney
disease. Shigella contamination only results from human fecal contamination and thus it is probably less
common than E. coll contamination, which has both human and animal sources. Shigella are one of the
most easily recognized causes of waterborne disease outbreaks because the advent of bloody diarrhea
spurs detailed investigations and often a cause is identified, even though Shigella are difficult to cultivate.
One large ground water outbreak in a sensitive aquifer occurred recently in Island Park, Idaho due,
perhaps, to an unidentified broken sewer line contaminating the well water (CDC, 1996). Because
Shigella is associated only with human feces, there is little incentive to look for Shigella as compared
with all the other bacterial pathogens that have both human and animal sources. As a result, no data are
available on occurrence although limited dose response data are available. EPA is unable to quantify the
benefits from avoided Shigella illness because no occurrence data are available.
Economic Analysis for the October 2006
Final Ground Water Rule 5-93
-------
Campylobacter and Arcobacter
Campylobacter (like Salmonella) are very common contaminants of food and water.
Campylobacter is commonly associated with animal manure, especially cow and chicken manure.
Campylobacter and Salmonella are associated with many common source (food or water) exposure
deaths, probably in large part because there are a large number of illnesses. More deaths may be
associated with Campylobacter and Salmonella than with viruses. Uniquely, Campylobacter is often
associated with Guillain-Barre paralysis that can last for weeks or months. About 1 paralysis case occurs
for every 1000 cases of campylobacteriosis (Altekruse et al, 1999). About 20% of paralysis patients are
left with some disability and approximately 5% die. Campylobacterosis is also associated with Reiter
syndrome (reactive arthritis). Approximately 1% of patients with camplybacterosis have arthritis onset in
one or more joints (especially the knee) in the 7 to 10 days after diarrheal onset (Altekruse et al. 1999).
The E. coll O157:H7 ground water outbreak in Walkerton, Ontario was also a large outbreak of
Campylobacterosis. Arcobacter (now a separate genus from Campylobacter) was responsible for a ground
water outbreak at a camp in Coeur dAlene, Idaho (McMillan, 1996). The sensitive aquifer was
contaminated by a septic tank. Campylobacter were associated with the recent outbreak in South Bass
Island, Ohio due to widespread fecal contamination in a sensitive aquifer (Ohio EPA, 2005). Like all
bacterial pathogens, special enrichment methods are needed to identify Campylobacter and Arcobacter in
environmental samples and so no data are available on the occurrence of these ubiquitous pathogens in
PWS well water. Campylobacter, like Vibrio has a viable but non-culturable environmental form which
makes it difficult to detect at times in water (Rollins and Colwell, 1986; Koenraad et al., 1997). Limited
dose-response data are available for Camplybacter (but not for Arcobacter). Because no occurrence data
are available, EPA is unable to quantify the benefits associated with avoided camplybacterosis (and
arcobacterosis) and their chronic sequelae (Guillane-Barre paralysis and reactive arthritis).
Salmonella
Salmonella causes typhoid fever, once a common and dangerous waterborne disease. Typhoid is
no longer a problem in the United States, and in recent years, Salmonella has become increasingly less
common as a common source outbreak agent while Campylobacterosis outbreaks have correspondingly
increased. The reasons for this change are unclear. Salmonella was identified in most fecally
contaminated PWS wells during the South Bass Island, OH outbreak in 2004 (Ohio EPA, 2005). The
main problems associated with Salmonella result from scenarios other than fecal source water
contamination. For example, the seven deaths that occurred due to Salmonella contamination in a ground
water PWS system in Gideon, Missouri were due to bird entry into a storage tank (Angulo et al., 1997).
Salmonella resulted in a very large outbreak in a ground water utility in Riverside, California during the
1960s (16,000 illnesses, 70 hospitalizations and 3 deaths) prior to the advent of the Total Coliform Rule
(Boring et al., 1971). This issue will be discussed further below. Limited dose-response data are
available for Salmonella but no occurrence data. The GWR EA provides nonqualified benefits resulting
from corrective actions such as disinfection that would mitigate against salmonellosis resulting from
animal entry into the distribution system or Salmonella contamination of PWS wells in general.
Legionella
Legionella are opportunistic bacterial pathogens that colonize water distributions systems. An
estimated 8,000-10,000 cases of Legionaires disease and Pontiac fever occur in the U.S. each year due to
Legionella. Twenty-one of 48 known species are able to infect humans. A study of 46 PWS wells from
Economic Analysis for the October 2006
Final Ground Water Rule 5-94
-------
16 water utilities in the United States and Canada showed that 38 wells (82%) were positive for
Legionella. Also, 33% of 114 PWS well samples were positive for Legionella. About half the identified
Legionella species were pathogenic forms. The authors conclude that this is the "first study that has
unambiguously proven that Legionella constitute a part of the microflora of ground water not known to be
under the direct influence of surface waters" (Riffard et al., 2004). EPA calculates benefits associated
with fecal contamination. However, Legionella are present in ground water in the absence of fecal
contamination and constitute a potential disease hazard if the bacterium is inhaled during showering.
Although nationally representative Legionella occurrence data are available for ground water, the
methods used are not typical or standardized. Also, no dose-response data exist so it is not possible to
quantify the disease burden from ground water exposure. If disinfection is required as a corrective
action, it will minimize the Legionella contamination of the distribution system. However, the GWR EA
only qualitatively accounts for the avoided illnesses from Legionella inhalation.
5.4.3.2 Estimate of Potentially Avoided Bacterial Caused Deaths by GWR
The quantified benefits in the EA predicts that among the 27% of the wells having sometime fecal
contamination about 34% (of the 27%) of these wells will be identified by triggered monitoring and be
required to take remedial action. The subsets of wells that fall into this category include 1) wells
predicted to have both viral and E. coli presence and identified by indicator monitoring (about 2/3 of the
27%) and 2) wells predicted to have E. coli presence but no viral presence and identified by indicator
monitoring (about 1/3 of the 27%). The quantified benefits model only addresses avoiding viral illness in
the first of these categories.
Although EPA did not formally quantify the benefits of avoided bacterial illness, EPA developed
a rough estimate of the potential deaths that might be prevented by reduced exposure to bacterial
pathogens under the GWR (to supplement the quantified benefits from virus exposure). The following
analytical steps were taken to generate such an estimate.
1) Estimate potential bacterial illnesses avoided
Analysis of outbreak data from 1991 through 2000, shows a total of 2,346 bacterial illnesses
compared to 1,806 viral illnesses and 4,523 illnesses of unknown etiology (see Exhibit 2.3) in section
2.2.1, compiled from CDC 1993, Kramer et al., 1996, Levy et al., 1998, Barwick et al., 2000, and Lee et
al., 2002). If the illnesses of unknown etiology are assumed to be viral26, then a total of 6,329 waterborne
viral illnesses would be associated with the waterborne outbreaks. The ratio of bacterial waterborne
illness to viral waterborne illness in outbreaks is therefore 0.37:1.
If we assume that this ratio of bacterial to viral waterborne illness also pertains to the relative rate
of endemic bacterial and viral illness prevented due to the GWR, then for every viral illness avoided
under the GWR we will also avoid an additional 0.37 of bacterial illness; i.e., relative to the quantified
cases of viral disease avoided, we would avoid 16,805 bacterial illnesses per year (41,868 predicted viral
illnesses avoided per year x 0.37 = 15,491).
26Because viral illnesses are generally more difficult to identify than bacterial illnesses, all illnesses with
unidentified causes were considered to be viral illnesses. This assumption most likely leads to an underestimate of
the ratio of bacterial to viral illness associated with reported outbreaks.
Economic Analysis for the October 2006
Final Ground Water Rule 5-95
-------
It is possible that the ratio of bacterial to viral endemic illness due to ground water is lower than
0.37:1 because the likelihood of detecting an outbreak due to bacterial illness may be greater than that of
detecting a viral outbreak27. However, any potential overestimate is likely to be compensated by the
assumption that all cases of outbreak illness of unknown etiology are viral. EPA recognizes that the
extrapolation from the ratio of outbreak etiology to the ratio of endemic illness etiology remains a major
uncertainty in this analysis and that the impacts of the two opposing biases on the extrapolated estimate is
unknown.
2) Estimate a mortality rate for waterborne bacterial illness
The approach taken for this step was to a) recognize the types and distribution of bacterial illness
associated with waterborne disease, b) characterize mortality rates for the different types of bacterial
illness identified in "a", and c) estimate a composite mortality rate considering "a" and "b".
For a), EPA used the same waterborne disease outbreak information from Exhibit 2.3, cited
above, to inform the types and distribution of bacterial illness in ground water systems.
For b), EPA used data from Mead et al (1999) (Table 3) to estimate mortality rates for each type
of bacterial illness identified in waterborne disease outbreaks in ground water systems (from Exhibit 2.3).
While Table 3 in Mead et al. (1999) provides estimates for illnesses, hospitalizations and deaths caused
by foodborne pathogens for specific bacterial agents, some of these agents have also been recognized as
causing waterborne disease (namely E. coli O157:H7, Salmonella non-typhoidal, Salmonella typhi,
Shigella, and Campylobacter spp). Therefore, EPA considers it appropriate to apply the calculated
mortality rates from Mead et al. (1999) to the same bacterial pathogens identified in waterborne disease
outbreaks. These mortality rates estimated from Mead ranged from 3.64 per 1,000 illnesses for
Salmonella typhi to 0.055 per 1,000 illnesses for Campylobacter spp. Fatality rates of other illnesses
caused by potentially waterborne bacterial pathogens included 0.83 per 1,000 illnesses for E. coli O157,
0.41 per 1000 for Salmonella non-typhoidal, and 0.16 per 1,000 for Shigella, and 2.54 for Vibrio the least
frequently reported ground water associated waterborne bacterial illness.
For c), EPA applied the above case fatality rates, calculated from Mead et al. (1999), to the illness
case rate reported in Exhibit 2.3, for each of these bacterial illnesses. EPA then calculated a weighted
average mortality rate for waterborne bacterial illness of 0.64 deaths per 1000 illnesses (see Exhibit 5.25).
In using mortality rates associated with foodborne illness, EPA assumes that differences in exposure
scenarios do not lead to differences in illness manifestation.
EPA thinks that the above estimate is not unreasonable because the spread among the mortality
rates of the different bacteria in Exhibit 5.25 is relatively small and therefore, giving a different weight
(e.g. based on additional information) to one particular agent over another would not significantly change
the estimate. The four most influential pathogens (Shigella, Campylobacter spp., E. coli O157,
Salmonella (non-typhoidal) in the weighted average had a relatively small difference in their case fatality
rate. The illness with the lowest case fatality rate, Vibrio, had a very small reported incidence of
waterborne disease associated with it.
27Bacterial illnesses are more likely to be diagnosed than viral illnesses because of their severity. However,
most cases of diarrhea associated with outbreaks are not identified by etiology and identifying etiology is
unnecessary in order to identify an outbreak.
Economic Analysis for the October 2006
Final Ground Water Rule 5-96
-------
With regard to E. coll, EPA only considered 0157 in the mortality rate estimate because most of
the pathogenic E. coll in reported waterborne disease outbreaks were identified as 0157. The Mead paper
reports two O157 strains with essentially the same case fatality rate and two other E. coll strains that are
not associated with mortality. Any E. coll o!57 before 1993 would probably not have been identified as
such since it was only recognized as a pathogen in 1982 and more broadly recognized only in 1993 after a
multi-state foodborne E. coll outbreak (Rangel et al., 2005). However, it is probable that all four strains
cause sporadic cases or even unidentified outbreaks of waterborne disease. If all four strains were
included in the calculation of mortality rate for pathogenic E. coll and weighted equally, the case fatality
rate for all pathogenic E. coll would have dropped by approximately 50% and the composite mortality
rate factor, calculated in Exhibit 5.25, would have dropped by approximately 15%.
3) Estimate potential annual bacterial deaths avoided by the GWR
EPA multiplied the composite mortality rate from (2) above (6.41 x 10"4) by the estimated annual
bacterial illnesses avoided in (1) (15,491) to estimate 10 potential deaths avoided per year. The
assumptions underlying these calculations are dependent on a variety of judgements. Other assumptions
are equally likely. EPA believes this calculation provides a plausible estimate of potential additional
deaths avoided from exposure to bacteria in GWSs.
Exhibit 5.25 Estimated Bacterial Illnesses and Deaths Avoided
Pathogen
Campylobacter spp
Ł. co// O1 57*
Salmonella typhi
Salmonell non-typhoidal
Shigella
Vibrio other
Illnesses
(thousands)
A
2,454.0
73.5
0.8
1,412.0
448.0
7.9
Hospitalizations
B
13,174
2,168
618
16,430
6,231
99
Hospitalizations/1 000
cases of illness
C = B/A
5.37
29.50
750.00
11.64
13.91
12.56
WBDO Total Cases:
Cases of
illness in
WBDO
D
223
807
124
625
556
11
2,346
Fraction of
WBDO illness
E = D/Total D
0.095
0.344
0.053
0.266
0.237
0.005
Composite
case
hospitalization
rate:
WBD illness
hospitalizations/
1000 cases
F = (C/1000)*E
5.10E-04
1 .01 E-02
3.96E-02
3.10E-03
3.30E-03
5.89E-05
5.68E-02
1 EPA assumed that E. coli O157 represents pathogenic E. coil identified in drinking water because it was the Ł coil strain identified as the
etiologic agent responsible all of the pathogenic Ł coli related ground water outbreaks. Other strains of pathogenic Ł. coli also cause
waterborne illness; however, they have not been identified as the etiologic agent of a ground water associated disease outbreak in the CDC
waterborne outbreak surveillance reports referenced in Exhibit 2.3.
** vibrio (other): i.e., not V. cholerae or vulnificus
Note: Estimated Total number of hospitalizations due to bacterial illness = estimated number of cases (15,491) x Bacterial
Composite Case Hospitalization Rate (15,491x 0.0568 = 880 hospitalizations/year).
Source:
(A) Data of pathogen specific illnesses from Table 3 in Mead paper.
(B) Data of pathogen specific deaths from Table 3 in Mead paper.
(D) Cases of illness from CDC outbreak surveillance reports in Exhibit 2.3. "Etiology of Waterborne Outbreaks in Ground Water Systems, 1991 -
2000," (Kramer etal., 1996; Levyetal., 1998; Barwick et al., 2000; and Lee et al., 2002).
Economic Analysis for the
Final Ground Water Rule
October 2006
5-97
-------
5.4.3.3 Estimate of a Hospitalization Rate for Waterborne Bacterial Illness
EPA used the same approach for estimating potential hospitalizations avoided by a reduction in
the incidence of bacterial illness, as it did for estimating potential deaths avoided. In order to estimate a
hospitalization rate due to waterborne bacterial illnesses, a composite hospitalization rate was developed
that represents the hospitalization rate for all waterborne bacterial pathogens adjusted by the frequency of
their occurrence as the etiologic agent in WBDOs. The data presented in Mead et al. (1999) represent
bacterial illnesses and hospitalizations due to all different exposures, e.g. food, person to person spread,
waterborne. By using information on the relative frequency of illness caused by different waterborne
pathogens under outbreak conditions we have developed a weighting scheme specific for considering
waterborne illness health effects and associated burden (hospitalization).
Exhibit 5.26 below presents the data used in the calculations and the results of the calculations
described as follows: The hospitalization rates (in column 4) were calculated from the cases of illness due
to all causes (column 2) and the number of hospitalizations (column 3) from Table 3 of Mead P.S. et al,
(1999). In order to develop the composite waterborne bacterial hospitalization rate, the etiologic fraction
was calculated by dividing cases of illness of a specific etiology from WBDOs by the total number of
WBDO cases of illness. An occurrence weighted hospitalization rate is calculated by multiplying the
etiologic fraction of WBDO cases (column 6) by the etiologic agent's hospitalization rate (column 4).
Lastly, the composite bacterial hospitalization rate is calculated by summing the weighted rates in column
7. The result is an integrated, or composite bacterial hospitalization rate of 57 hospitalizations/1000
cases of waterborne bacterial illness.
EPA estimated the potential hospitalizations avoided (from reduced bacterial illness) by
multiplying the estimate for bacterial illness avoided (15,491) by the composite bacterial hospitalization
rate of 57/1000 cases of illness to indicate 880 annual potential hospitalizations. If the average bacterial
hospitalization cost is assumed to be greater than the Type B viral illness adult hospitalization cost used in
the quantified benefits section ($5,000 - see Exhibit 5.22 E), then the benefits of preventing bacterial
illness would be $4.4 million or more.
Economic Analysis for the October 2006
Final Ground Water Rule 5-98
-------
Exhibit 5.26 Waterborne Bacterial Illness Hospitalization Rates
Pathogen
Campylobacter spp
Eco// 0157*
Salmonella typhi
Salmonell non-typhoidal
Shigella
Vibrio other
Illnesses
(thousands)
A
2,454.0
73.5
0.8
1,412.0
448.0
7.9
Hospitalizations
B
13,174
2,168
618
16,430
6,231
99
Hospitalizations/1000
cases of illness
C = B/A
5.37
29.50
750.00
11.64
13.91
12.56
WBDO Total Cases:
Cases of
illness in
WBDO
D
223
807
124
625
556
11
2,346
Fraction of
WBDO illness
E = D/Total D
0.095
0.344
0.053
0.266
0.237
0.005
Composite
case
hospitalization
rate:
WBD illness
hospitalizations/
1000 cases
F = (C/1000)*E
5.10E-04
1.01E-02
3.96E-02
3.10E-03
3.30E-03
5.89E-05
5.68E-02
* EPA assumed that E. coli O157 represents pathogenic E. coli identified in drinking water because it was the E. coli strain identified as the
etiologic agent responsible all of the pathogenic E. coli related ground water outbreaks. Other strains of pathogenic Ł. coli also cause
waterborne illness; however, they have not been identified as the etiologic agent of a ground water associated disease outbreak in the CDC
waterborne outbreak surveillance reports referenced in Exhibit 2.3.
** vibrio (other): i.e., not V. cholerae or vulnificus
Note: Estimated Total number of hospitalizations due to bacterial illness = estimated number of cases (16,805) x Bacterial
Composite Case Hospitalization Rate (15,491x 0.0568 = 880 hospitalizations/year).
Source:
(A) Data of pathogen specific illnesses from Table 3 in Mead paper.
(B) Data of pathogen specific deaths from Table 3 in Mead paper.
(D) Cases of illness from CDC outbreak surveillance reports in Exhibit 2.3. "Etiology of Waterborne Outbreaks in Ground Water Systems, 1991 -
2000," (Kramer etal., 1996; Levyetal., 1998; Barwicket al., 2000; and Lee et a I., 2002).
Economic Analysis for the
Final Ground Water Rule
October 2006
5-99
-------
Summary of Nonqualified Benefits for Bacterial Illness Avoided
EPA estimated that the total benefits could increase by a factor of five by only accounting for
additional deaths and hospitalizations caused by bacterial illness being avoided. The actual number of
hospitalizations and deaths avoided could be higher or lower. Within the context of best available science,
given all the other nonqualified benefits for chronic bacterial illnesses avoided, EPA believes that the
total nonqualified benefits for bacteria could exceed a factor of four relative to the quantified benefits
for viruses. If a value of $7.5 million is assumed for each death avoided (see Section 5.3.1.2), then the
value of bacterial deaths avoided each year would be approximately $75 million (year 2003 dollars). If a
value of $5,000 is assumed for each hospitalization avoided (see Exhibit 5.22e), then the value of
bacterial hospitalizations avoided each year would be approximately $4.4 million (year 2003 dollars).
Taken together, these additional benefits would result in benefits that are five times the quantified
benefits for viruses (four times quantified benefits plus quantified benefits) [i.e., (75 + 4.4+ 19.7)719.7 =
5].
5.4.4 Other Chronic and Acute Illness Potentially Avoided
As discussed in Section 5.1, the GWR EA quantifies only a small subset of the total benefits.
Only acute illnesses arising from Type A virus represented by rotavirus data and Type B virus
represented by enterovirus data are quantified. Acute illnesses from other viral etiologies and from
bacterial infection are not quantified. Also, chronic illnesses resulting from either viral or bacterial illness
are not quantified. This section addresses some of the chronic disease endpoints that are not quantified in
this EA and are only qualitatively assessed.
Hypertension and Reduced Kidney Function
As a result of the large number of illnesses, hospitalizations and deaths due to E. coll O157:H7
and Camplybacter contamination at Walkerton, Ontario, a long term study was undertaken to evaluate
chronic sequelae (Garg et al, 2005). This is the largest and longest study ever of chronic sequella after
ground water exposure. After a mean follow-up of 3.7 years after the outbreak, patients with moderate
and severe gastroenteritis had an adjusted relative risk of hypertension of 1.15 (0.97-1.35) and 1.28
(1.04-1.56) respectively. A similar association was seen for reduced kidney function. The authors (Garg
et al., 2005) conclude:
"Adults with symptomatic bacterial gastroenteritis from drinking contaminated water were more
likely than asymptomatic adults to have newly diagnosed hypertension and reduced renal
function during the follow-up period of almost 4 years after infection. "
This new finding documents that an acute self-limiting bacterial gastroenteritis is likely to be
followed by hypertension and reduced kidney function in a significant subset of individuals. As
discussed in Section 5.4.3.1, bacterial gastroenteritis is a frequent occurrence in the U.S. resulting from
both epidemic and endemic exposure. It is likely that a similar study of endemic acute self-limiting
bacterial gastroenteritis would also document increased likelihood of these two chronic disease sequelae.
The GWR EA quantifies only acute illnesses resulting from some virus infection. The EA does
not quantify acute bacterial disease, including acute self-limiting bacterial gastroenteritis. The EA also
does not quantify the chronic disease sequelae such as hypertension and reduced renal function that can
Economic Analysis for the October 2006
Final Ground Water Rule 5-100
-------
arise from acute illness. Because bacterial disease and their chronic sequelae are considered only
qualitatively, the GWR EA underestimates the benefits associated with bacterial illness.
Helicobacter pylori
Helicobacter pylori is often associated with ground water (Hegarty et al. 1999; Rolle-Kampzczyk
et al, 2004) and is known to cause gastric ulcers. However, Helicobacter is not culturable and so
occurrence data from ground water are fairly uncertain. Improved hygiene and water treatment have
together reduced the number of ulcers caused by this organism over the last few decades but it is
impossible to quantify that decrease. Nevertheless, corrective action and especially disinfection resulting
from the GWR will likely provide a reduction in the number of intestinal colonizations by Helicobacter
pylori and a corresponding decrease in gastric ulcers. The GWR EA does not quantify the benefits
associated with avoiding bacterial disease. Because chronic infection by Helicobacter pylori can lead to
gastric ulcers, there are significant health effects potentially avoided due to Helicobacter pylori infection
that are considered in this EA only qualitatively.
Reactive Arthritis, Irritable Bowel Syndrome and Persistent Diarrhea
A recent study (Rees, et al, 2004) evaluated the chronic sequelae resulting from enteric infection
identified but the California FoodNet Surveillance. Eight percent of respondents reported new joint pain
after infection and 35% reported new gastrointestinal symptoms including persistent diarrhea and irritable
bowel syndrome. Reiter's Disease is form of reactive arthritis which, as discussed in the section
describing the nonqualified benefits associated with Campylobacter, is often associated with
Campylobacter infection.
Primary Amoebic Meningioencephalitis (PAM)
In 2003, two five-year old boys living in the same water service area near Phoenix AZ died in the
same week from Primary Amoebic Meningioencephalitis (PAM) (Marciano-Cabral et al, 2003). Both
boys lived in homes supplied by untreated PWS wells. Atypically, the wells in that area provide water at
elevated temperatures, representing the elevated geothermal gradient in the subsurface. Seventeen
samples taken from the boys homes were positive for Naegleria fowleri and N. fowleri was also
responsible for the boy's deaths from PAM. It is likely that the heated ground water provided a suitable
habitat for TV. fowleri colonization and growth either in the aquifer, the well, the distribution system or the
household plumbing. N. fowleri is effectively treated with disinfection or chlorination. The final GWR
may require systems with elevated ground water temperature to take corrective action such as disinfection
if, for example, the elevated temperature is noted in a sanitary survey. However, EPA is unable to
quantify the number of PAM cases that would be avoided in the future, so the benefits are presented only
qualitatively.
Cryptosporidium and Giardia
Cryptosporidium and Giardia are associated with surface water rather than ground water PWS
systems. If Cryptosporidium or Giardia are recognized in well water, the system should be considered as
a surface water system (ground water under the direct influence of surface water [GWUDI]) rather than as
a ground water system. As a GWUDI system, these wells are regulated by the Long Term 2 Surface
Water Treatment Rule and not by the GWR. However, on several occasions, PWS wells were regulated
as if they were ground water rather than as surface water until a cryptosporidiosis or Giardiasis outbreak
Economic Analysis for the October 2006
Final Ground Water Rule 5-101
-------
was recognized. Examples of Cryptosporidiosis outbreaks associated with ground water wells (not
recognized as GWUDI wells) located in sensitive aquifers includes outbreaks in Braun Station, TX
(D'Antonio et al, 1985), Reading, PA (Moore et al, 1993), Brushy Creek, TX (Lee et al., 2001) and
Yakima, WA (Dworkin et al, 1996). The sanitary survey and hydrological sensitivity assessments under
the GWR rule may identify some systems as being under the direct influence of surface water and thereby
make such systems subject to surface water treatment technique requirements and avoid potential
outbreaks. Such actions would result in both increased costs and benefits for such systems.
5.4.5 Reduction in Outbreak Risk and Response Costs
Besides reducing the endemic risk of illnesses from waterborne pathogens, the GWR will reduce
the likelihood of major outbreaks from occurring. These avoided illnesses and other costs are not
estimated or included in the GWR benefits estimates and would be difficult to quantify. The economic
value of reducing the risk of outbreaks could be quite high when the magnitude of potential costs is
considered. Other types of costs associated with outbreaks include spending by local, State, and national
public health agencies; emergency corrective actions by utilities; and possible legal costs if liability is a
factor. Affected water systems and local governments may incur costs through provisions of alternative
water supplies and issuing customer water use warnings and health alerts. Commercial establishments
(e.g., restaurants) and their customers may incur costs due to interrupted and lost service. Local
businesses, institutions, and households may incur costs associated with undertaking averting and
defensive actions. Cost-benefit analyses of large engineering works typically include probabilistic failure
assessment to determine the likely benefits of avoiding catastrophic effects. Such analyses are not
included in the EA quantified or nonqualified benefits because there are too many likely failure
scenarios. Thus, to the extent that GWR reduces the likelihood of waterborne disease outbreaks, avoided
response costs are potentially numerous and significant.
During outbreaks, consumers and businesses may use alternative water sources or practice
behaviors to reduce risk, such as boiling water. If the rule reduces the need for these averting behaviors,
an economic benefit will accrue. To give a sense of the possible scale, the expenditures on averting
behaviors during an outbreak of Giardiasis, such as hauling in safe water, boiling water, and purchasing
bottled water, were estimated at between $3.10 to $9.80 per person per day (year 2003 dollars) during the
outbreak (Harrington et al., 1991). If these figures are applied to even a small drinking water system
serving 10,000 customers, total expenditures on averting behavior during a waterborne disease outbreak
could range between $31,000 and $98,000 per day. Determining the precise reduction in outbreak risk
and the resulting benefits due to reduced or avoided averting behavior is not possible given current
information, but potential benefits are expected to be substantial.
Five studies were identified that used the averting cost approach to estimate household and other
costs attributable to short-term contamination of drinking water supplies (Abdalla, 1990; Abdalla et al.,
1992; Harrington et al., 1991; Sun et al., 1992; Van Houtven et al., 1997). The most relevant of these for
the GWR analysis is a study by Harrington et al. (1991), that analyzes the costs associated with drinking
water contamination by Giardia in Luzerne County, Pennsylvania. The December 1983 outbreak resulted
in 366 confirmed Giardiasis cases resulting from sewage leaking into the unfiltered source water. The
total affected population was 75,000 individuals across Pittston Burough and 17 other municipalities.
The Harrington study also developed a theoretical and empirical example of how outbreak costs are
incurred, based on the Luzerne County example.
Economic Analysis for the October 2006
Final Ground Water Rule 5-102
-------
The four steps associated with a waterborne outbreak that may impose costs on society are
discovery, survey and testing, reaction, and aftermath (Harrington et al., 1991). These are described
below:
Discovery. Health care providers or State, local, or hospital laboratory technicians send
reports to State authorities notifying them of the need for further investigation when the
rate of new cases suddenly increases above the normal rate.
• Survey and testing. A host of epidemiological surveys may be conducted, along with
tests of the water supply, once a few cases are confirmed.
• Reaction. Local authorities and the water system may issue boil-water advisories, or
other warnings to reduce exposure once a link is made between the drinking water supply
and the disease outbreak. Businesses, as well as households, may be affected by such
action, requiring government agencies to begin surveillance and enforcement activities
and, in some cases, provide alternative water sources.
Aftermath. This final step involves discussions of any long-term solutions to the
problem, and how the costs of the outbreak and prevention of future ones may be shared.
These discussions can only take place once the outbreak is contained by actions taken
during the previous phase.
The Luzerne County outbreak resulted in losses due to actions taken by individuals to avoid the
contaminated water that are estimated to be between $36.8 million and $109.4 million (year 2003 dollars).
The predominant cost was time lost to boiling water. Losses due to averting actions for restaurants, bars,
schools and other businesses during the outbreak exceeded $1.8 million. The burden for government
agencies was $407,300 and the outbreak cost the water supply utility $3.2 million. These costs do not
include legal fees, outbreak effects on businesses that were not investigated, leisure activities, or net
losses due to substituting more expensive beverages for tap water. During a waterborne disease outbreak
in Walkerton, Ontario (population 5,000), an analysis conservatively estimated the economic impact
excluding medically related costs to be over $43 million in Canadian dollars (approximately $32 million
in U.S. dollars) (year 2003 dollars) (Livernois, 2002).
5.4.6 Reduced Disinfection Treatment Failure Rates and Associated Waterborne Disease
Implementation of the GWRmay lead to additional benefits from reductions in disinfection
failures. Such benefits would stem from the increased oversight of treatment processes as part of
compliance with regulatory requirements (e.g., during sanitary surveys and compliance monitoring, or
with upgrades in disinfection to 4 log inactivation of viruses).
Direct data on the numbers of illnesses and deaths resulting from treatment failure are not
available. However, these numbers could be substantial given that the relative rates outbreak related
illness attributed to treatment deficiencies versus outbreak related illnesses attributed to source water
deficiencies in untreated ground waters. Of the outbreak-related illnesses due to viral, bacterial, and
unknown agents reported in GWSs during 1991-2000 (see Exhibit 4.31), 4,224 were attributable to
drinking untreated ground water and 4,888 were attributable to a treatment deficiency.
Economic Analysis for the October 2006
Final Ground Water Rule 5-103
-------
EPA's baseline estimate among disinfected supplies assumes that all disinfected supplies are
providing either 2 log or 4 log inactivation of viruses and does not account for occasional disinfection
failures or upsets. Given this simplifying assumption, the baseline estimate for viral disease among
disinfected supplies is about 2,500 cases, whereas the baseline risk for the undisinfected supplies is about
179,700 cases (approximately 1.4% of illnesses we estimate to occur in currently disinfecting wells).
However, based on outbreak data, an additional 1.16 (4,888/4,224) outbreak-related illnesses occur in
systems using disinfection with treatment failures for every outbreak-related illness in systems with
untreated ground water. If this same ratio were to apply to endemic illness in disinfected versus
undisinfected supplies, the baseline incidence of disease in disinfected supplies (taking disinfection
failures into account) would be roughly 80 times higher than currently estimated, i.e., 1.16 x
179,700/2,500 additional illnesses per year. Even small percent reductions in disinfection failures or
upsets resulting from this rule (e.g., 10%) could therefore lead to substantial reductions in baseline risk
among disinfected supplies which is currently not accounted for in the quantified benefits.
As with the analysis of bacterial illnesses, there is no way of knowing whether the ratio of 1.16 is
generally higher or lower for non-outbreak illnesses. It is possible, that outbreaks represent a larger or
smaller portion of total cases for treatment deficiencies than for untreated ground water. Similarly, it is
not possible to estimate with any degree of confidence the percent reduction in treatment failures that
might result from this rule. In any case, benefits from reduced treatment are likely to result from this rule
and would be additive to benefits estimates in the main analysis.
5.4.7 Distribution System Contamination
The GWR will lead to additional systems providing disinfection at the source and this will lead to
some disinfectant residual provided in the distribution system. Opportunistic bacterial pathogens are soil
and other environmental bacteria that can colonize distribution systems and typically are harbored and
protected in distribution system biofilms where substantial bacterial population growth can occur.
Periodic biofilm sloughing can introduce these pathogens into untreated ground water. Major groups of
opportunistic bacterial include Pseudomonas, Acinetobacter, Xanthamonas, Moraxella, Mycobacterium
and Serratia (Rusin et al. 1997 a, b). Each of these bacteria is capable of causing disease, typically
pneumonia, meningitis and septicemia, but also other diseases, in sensitive subpopulations. Sensitive
subpopulations are estimated to be 20% of the U.S. population (Gerba et al., 1996). Although sensitive
subpopulations are often advised to drink only treated or bottled water, it is often difficult to know the
water source for all exposures, especially water from transient systems. Corrective action, including
disinfection can have substantial benefit in reducing exposure to opportunistic pathogens in untreated
distribution systems. The GWR only qualitatively accounts for the benefits that might accrue for reduced
exposure and illness among sensitive subpopulations due to opportunistic bacterial pathogens.
5.4.8 Benefits From the Reduction of Co-Occurring and Emerging Contaminants
While the benefits analysis for the GWR only includes reductions in illness and mortality
attributable to illnesses from Type A and Type B viruses, the GWR is expected to reduce exposure to
other pathogens. For example, some membrane technologies installed to remove viruses can reduce or
eliminate many other drinking water contaminants including arsenic and bacteria. Strengthened
regulatory requirements will translate into increased removal of additional pathogens and a resulting
reduction in risk. This may prove essential, as the impact of emerging pathogens is not well established.
Economic Analysis for the October 2006
Final Ground Water Rule 5-104
-------
Unfortunately, EPA is unable to quantify the resultant benefit associated with a reduction in risk from
emerging pathogens due to current data limitations.
5.4.9 Reduced Uncertainty/Costs to Households to Avert Infection
To the extent that the GWR decreases consumers' uncertainty about expected health outcomes
from consumption of drinking water, the rule should provide direct benefits independent of risk reduction
benefits. In other words, drinking water consumers may be willing to pay a premium for regulatory
action if it reduces their uncertainty about whether they will become ill (Moore, 1990).
Conceptually, whether consumers would be willing to pay something extra to reduce uncertainty
in the GWR context depends on several complicated factors, including consumers' degree of risk
aversion, their perceptions about drinking water quality, and the expected probability and severity of
human health effects associated with microbial contamination of drinking water. For example, risk
premiums would be expected only for consumers who are risk averse. Further, the magnitude of any
premium would be expected to be positively related to the probability and severity of expected health
outcomes, and the degree to which consumers perceive them to be affected by regulatory action.
In addition, to the extent that the GWR can be expected to reduce a household's perceptions of
the health risks associated with drinking water, regulatory action should reduce household averting
actions and costs. Any such cost savings would represent a regulatory benefit. Examples of household
averting actions include: 1) securing drinking water from alternative sources (e.g., bottled water), 2)
installation of home treatment systems (e.g., point-of-use and point-of-entry treatment), and 3) boiling tap
water used for consumption. These actions can involve significant cash outlays and implicit costs (e.g.,
time costs).
A number of factors, however, limit the relevance of this potential benefit in the GWR context.
One is the possibility that regulatory action may not affect household perceptions of health risks enough
to motivate them to forego averting actions. A related factor is that many households that undertake
averting action for health reasons may be especially risk averse (e.g., households with infants or
immuncompromised persons). These households might be expected to pursue averting actions regardless
of the level of regulatory control if they believe such actions may provide added protection against
microbial risks. However, any treatment that also improves taste and odor as well as microbial protection
may likely result in households forgoing averting behavior.
5.4.10 Summary of Nonqualified Benefits
The total benefits in this EA include both quantified and nonqualified components. The
quantified benefits are summarized in Exhibit 5.23. The total nonquantified benefits are captured by
several additional analyses and are summarized here but discussed in more detail throughout this EA.
EPA estimates that, based on avoided bacterial illnesses and deaths alone, which are discussed only in the
nonquantified benefits section, the total benefits could be underestimated by a factor of five. Other
benefits will also accrue but are not quantified.
The nonquantified benefits result from multiple factors. First, the quantified benefits are based on
limited, well-defined data and key assumptions that restrict the input parameters in the quantified benefit
Economic Analysis for the October 2006
Final Ground Water Rule 5-105
-------
calculation. Typically, these assumptions resulted in low mean values and narrow uncertainty ranges in
the benefits analysis. This EA, where applicable, discusses alternative assumptions. For example, the
enterovirus morbidity fractions are, by assumption, not determined using coxsackievirus (an enterovirus)
data although the enterovirus severity data use all enterovirus data. If coxsackievirus data were available,
the mean morbidity values would be greater. Choosing alternative values and ranges and differing key
assumptions, which might also be deemed reasonable, would increase the quantified benefits in this EA.
Second, the quantified benefits are based on data and assumptions that pertain to only partial
representation of Type A and Type B viruses potentially found in PWS wells with fecal contamination.
Due to limited available data, only rotavirus and some enterovirus data were used to calculate the
quantified benefits. As is more completely discussed in Section 5.4. other viruses as well as pathogenic
bacteria may contribute to the disease burden, both acute and chronic, associated with PWS wells with
fecal contamination. Most importantly, bacterial illnesses can result in more frequent and lengthier
hospitalization and more frequently have fatal outcomes. If bacterial diseases were considered in the
quantified benefits, the monetized benefits could be substantially greater because bacterial disease can be
more severe and can result in higher mortality rates.
Third, the quantified benefits are based on data and assumptions that limit the characterization of
acute disease. For rotavirus, only acute gastroenteritis illness and fatal dehydration associated with that
illness are monetized. Norovirus disease is not considered. For the enteroviruses, all acute disease
endpoints are considered, but the prevalence of severe endemic cases may be substantially diluted by the
large number of hand, foot, and mouth disease cases that are not likely to be waterborne. Thus, the
proportion of severe cases in the quantitative benefits is likely to be underestimated. As is discussed more
completely in Section 5.4, in neither instance, either for rotavirus or the enteroviruses, are chronic
diseases identified or monetized in the quantitative benefits calculation.
Fourth, the quantified benefits are based explicitly on what has been directly measured in PWS
wells, yet there is great difficulty in identifying and counting all infectious viral pathogens in dilute
drinking water samples. Indeed, some viral pathogens like infectious norovirus can never be identified in
any sample. Section 4.3.2 discusses these difficulties in more detail. Standard fecal indicator data such
as total coliforms and E. coll, commonly used to identify water treatment deficiencies and potential
human health hazards, are explicitly not used to determine human exposure for the purposes of
quantifying the benefits in this EA.
Fifth, the quantified benefits are assumed to be based only on one contamination scenario, fecal
contamination of source water. Other contamination scenarios are thoroughly documented in the ground
water contamination and outbreak scientific literature. However, these scenarios, such as inadequate
disinfection, are not explicitly considered in calculating the quantified benefits in this EA.
Sixth, the quantified benefits are assumed to be based only on avoidance of endemic disease. The
GWR will likely also decrease the incidence of epidemic disease (outbreaks). If epidemic illnesses and
the avoided non-health-related costs of ground waterborne disease outbreaks were included, the
quantified benefits would increase.
In summary, this EA quantifies a subset of the total health and non-health related benefits. In a
sample calculation, discussed in Section 5.4.3.2, EPA estimated that the total benefits could increase by a
factor of five by only accounting for additional deaths and hospitalizations caused by bacterial illness
being avoided. While EPA recognizes that this estimate includes substantial uncertainty, given all the
Economic Analysis for the October 2006
Final Ground Water Rule 5-106
-------
other nonquantified factors described above, EPA believes that the total benefits from the GWR are likely
to be more than five times those which have been quantified.
5.5 Alternative Analyses
To quantify the effects that differences in major modeling assumptions would have on benefits
analyses, several alternative analyses were performed. This section presents discussion and results for
four such analyses: use of an alternative baseline for viral concentration; use of two alternative dose
response functions for Type A viruses; use of an alternative dose response function for Type B viruses;
and use of an alternative set of occurrence data (Exhibit 5.27).
5.5.1 Alternative Viral Concentration Baseline
The main analysis of this EA applies data on viral concentrations derived from the Abbaszadegan
and Pennsylvania studies to estimate viral concentrations in those wells defined in this EA as "less
vulnerable" (see Chapter 4). Data from Lieberman are used to estimate viral concentrations for those
wells defined as "more vulnerable." These data are applied to each category of public water system, e.g.,
CWS, NTNCWS, and TNCWS in the main analysis.
As a sensitivity analysis, these same studies are applied differentially to the six categories (based
on three system types and their status as more or less vulnerable): Abbaszadegan's concentration data are
applied to only less vulnerable CWSs; Lieberman's concentration data are applied to more vulnerable
CWSs; and the Pennsylvania study's concentration data are applied to both the more vulnerable and less
vulnerable NTNCWSs and TNCWSs. This alternative would increase the quantified benefits of rule
implementation by a range of $1.1 million to $1.7 million (an increase of approximately 7 to 9 percent,
using a 7 percent and 3 percent discount rate, respectively) (Exhibit 5.27). Benefits would decrease for
CWSs and increase for both NTNCWSs and TNCWSs.
5.5.2 Alternative Type A Dose response
Two alternative dose response models were developed by EPA for Type A viruses using the same
study data (Ward et al., 1986) that were used for the dose response function in the main analysis. In both
alternative analyses, only the results for the 7 subjects exposed to 0.9 viruses were used since this dose
level is most similar to doses likely to be experienced by those consuming contaminated water. As
described in detail in Appendix F.5.1, an exponential model form was used having the form P = 1 - e" -
D*r. In this model, D is the expected daily dose of virus and r is the model parameter.
In the first analysis, a large sample of r values were generated using an MCMC method to capture
uncertainty in the true value of the r parameter. This alternative leads to a higher estimate of risk and
quantified benefits related to Type A viruses. The results, in terms of the monetized benefits using this
alternative dose response function (with all other aspects kept the same as in the main model), are shown
in Exhibit 5.27. This alternative dose response function would increase the annualized quantified benefits
by a range of $1.7 million to $2.1 million (approximately 10 percent) relative to the annualized benefits
for the main analysis, using a 7 percent and 3 percent discount rate, respectively.
Economic Analysis for the October 2006
Final Ground Water Rule 5-107
-------
In the second alternative dose response analysis for Type A viruses, the r value was estimated
without consideration of uncertainty. This alternative leads to a lower estimate of risk and benefits
related to Type A viruses by a range of $1.7 million to $2.0 million (approximately 10 percent in both
cases) relative to the annualized quantified benefits for the main analysis, using a discount rate of 7
percent and 3 percent, respectively.
5.5.3 Alternative Type B Dose Response
An alternative dose response model was developed by EPA for Type B viruses using the same
study data (Schiff et al., 1984) that were for the dose response function in the main analysis. In this
alternative analysis, the two highest dose groups (33,000 and 330,000 pfu) were excluded from the
analysis. As described in detail in Appendix F, an exponential model form was used having the form P =
1 - e" - D*r. In this model, D is the expected daily dose of virus and r is the model parameter. A large
sample of r values were generated using a MCMC method to capture uncertainty in the true value of the r
parameter.
This alternative leads to a lower estimate of risk and benefits related to Type B viruses. The
results, in terms of the monetized benefits using this alternative dose response function (with all other
aspects kept the same as in the main model), are shown in Exhibit 5.27. This alternative dose response
function would decrease the annualized benefits by a range of $3.7 million to $4.3 million (approximately
22 percent) relative to the annualized benefits for the main analysis, using a 7 percent and 3 percent
discount rate, respectively.
5.5.4 Alternative Occurrence (Peer Review) Data
The Agency performed an alternative benefits analysis by running the benefits model using a
subset of the studies that were the basis for the fecal indicator and viral occurrence rates in the primary
analysis. This subset comprised just those studies which had undergone peer review prior to publication28
and so consisted only of peer-reviewed data (a parallel analysis was performed for the cost model-see
section 6.4.8). Compared to the primary analysis, this alternative benefits model results in an increase in
the annualized benefits by a range of $10.4 million to $12.2 million (an approximately 62% increase),
using 7 percent and 3 percent discount rates, respectively (Exhibit 5.27).
98
Studies omitted from the alternative occurrence model are those used for the primary analysis (detailed in
Ch. 4 of the EA) that were either not published or not peer reviewed prior to publication: Missouri Alluvial Aquifer
(Vaughn, 1996), Wisconsin Migrant Worker Camp (USEPAetal., 1998), EPA Vulnerability (USEPA, 1998), New
England (Doherty et al., 1998), Three-State Study #3: Minnesota (Battigelli, 1999), Three-State Study #1:
Wisconsin (Battigelli, 1999), and the Montana Study (Miller and Meek, 1996).
Economic Analysis for the October 2006
Final Ground Water Rule 5-108
-------
Exhibit 5.27 Results of Alternative Analyses
Analysis
Primary Analysis
Alternative Viral Concentration
Baseline
First Alternative Type A Dose
Response
Second Alternative Type A Dose
Response
Alternative Type B Dose Response
Peer Review (Subset of Data Used in
Primary Analysis)
Incremental Difference in Mean Benefits from Primary
3% Discount Rate
$19.7
$1.7
$2.1
($2.0)
($4.3)
$12.2
7% Discount Rate
$16.8
$1.1
$1.7
($1.7)
($3.7)
$10.4
Notes: The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified benefits are
expected to comprise a significant portion of the overall benefits of the Rule and are presented in Section 5.4.
Source: GWR Model Output
5.6 Summary of Uncertainty
This chapter presents the data, assumptions, and methods used to estimate the baseline risks, in
terms of illnesses and deaths, from the presence of viral pathogens in ground water used as source water
for public drinking water systems. It also presents estimates of the reductions in those risks (benefits)
resulting from the implementation of the GWR, and of the monetized value of those benefits. Throughout
this chapter, an attempt has been made to address uncertainty in the inputs to and the results obtained
from the baseline risk and risk reduction modeling. Exhibit 5.28 summarizes the major uncertainties for
modeling GWR benefits as well as any anticipated influence on the cost model. Additional discussion of
the uncertainty and variability factors influencing the benefits model follows (see Chapter 6 for additional
discussion of uncertainty and variability in the cost model estimates).
Economic Analysis for the
Final Ground Water Rule
October 2006
5-109
-------
Exhibit 5.28 Summary of Uncertainties Affecting GWR Estimates
(continued on next page)
Uncertainty
Viral and indicator
occurrence
Effectiveness of
indicator monitoring
Rotavirus used to
represent Type A
infectivity
Model/data used for
rotavirus dose
response
Echovirus 12 used
to represent Type B
infectivity
Model/data used for
echovirus dose
response
Rotavirus Used to
represent Type A
morbidity
Echovirus 12 used
to represent Type B
morbidity
Population rates in
TNCWSs
Secondary spread
Infectivity risk of
sensitive subgroups
Morbidity and
mortality risk for
elderly or immuno-
compromised
Value of health
outcomes (COI)
Five repeat samples
(not modeled)
Section
Discussion
of
Uncertainty
Exhibit 4.32
Exhibit 4.32
5.2.3.6
5.2.4.1
Appendix F
5.2.4.1
Appendix F
5.2.4.1
5.2.4.1
5.2.3.4
5.2.4.2
5.2.4.3
Appendix E
5.2.5.3
5.2.5.3
5.2.4.2
5.4.10
5.2.5.6
Effect on Benefits Estimates
Under-
estimate
X
X
X
X
X
X
X
X
X
X
Over-
estimate
X
X
X
Unknown
Impact
X
X
Effect on Cost Estimates
Under-
estimate
Over-
estimate
X
Unknown
Impact
X
X
None
None
None
None
None
None
None
None
None
None
None
Economic Analysis for the
Final Ground Water Rule
5-110
October 2006
-------
Uncertainty
Treatment failure
(not modeled)
Distribution system
risk (not modeled)
Bacterial illness (not
included in main
benefits)
Section
Discussion
of
Uncertainty
5.4.7
5.4.8
5.4.3.1
Effect on Benefits Estimates
Under-
estimate
X
X
X
Over-
estimate
Unknown
Impact
Effect on Cost Estimates
Under-
estimate
X
X
Over-
estimate
Unknown
Impact
None
In the inputs to the exposure estimates for the risk model, the same elements of variability and
uncertainty that were discussed in Chapter 4 for the viral and indicator occurrence data and modeling
apply to their use here in the baseline risk and risk reduction modeling. Of particular note are the
concerns that the limited monitoring data available on viral pathogen occurrence may lead to
underestimates of the number of wells and the number of people potentially affected. In addition, the
limited data on indicator and pathogen co-occurrence may be resulting in an underestimate of the
effectiveness of indicator monitoring of source water to identify those wells that are expected to have
viral pathogens present.
Uncertainty exists in the dose response functions for infectivity developed for Type A and Type
B viruses based on rotavirus and echovirus challenge data, respectively. In the main analysis, this
uncertainty is considered to some extent by virtue of the large set of dose response function parameters
that were generated and used in the simulation model. In addition, in an effort to address the uncertainty
in the risk of infection, the alternative dose response functions were developed using data from the
challenge studies that was considered most representative of likely exposures via drinking water.
Variability and uncertainty are considered in the morbidity factors used in the risk model.
Variability is accounted for by applying different factors for young children versus the rest of the
population. Uncertainty is also considered as uniform uncertainty distributions in these morbidity factors.
Secondary spread is included in the analysis, and variability is included in secondary spread rate
for Type A viruses as a function of age of the primary infected individual, while uncertainty in the
secondary spread factor is included for Type B viruses. Beyond this, however, there is some concern that
secondary spread is being underestimated because of limited data which required use of a metric where
secondary cases are associated with primary cases of illness rather than with primary cases of infection.
In the risk and benefits analysis, there is recognition of three sensitive subgroups: young, old and
immunocompromised. There are, however, only limited quantitative considerations of the potential
increased impacts on these subgroups. There is no difference included for any of these subgroups with
respect to infectivity risk relative to the population at large. An increased morbidity factor is included for
the very young, and an increased mortality factor for neonates is included for Type B viruses. There is no
Economic Analysis for the
Final Ground Water Rule
October 2006
5-111
-------
consideration of increased morbidity or mortality risk for older persons or the immunocompromised
subgroups. (Note that the analysis does consider increased severity of illness for the
immunocompromised in terms of medical costs incurred.)
In the valuation of benefits analysis, a cost-of-illness (COI) approach has been used rather than a
willingness-to-pay (WTP) approach due to lack of available data. Generally, WTP is considered a more
appropriate approach to estimating the benefits of avoiding risks and typically leads to higher benefits
than the COI approach. Therefore, there is some concern that the value of the GWR benefits may be
understated because they are based on COI rather than WTP.
As described previously, EPA's national occurrence dataset includes information on virus and
indicator occurrence from 1309 wells among 15 studies. These wells serve community and
noncommunity public water systems of different sizes and types (e.g., disinfecting/nondisinfecting), and
the wells are situated in shallow and deep aquifers, located in sensitive or nonsensitive hydrogeologic
settings, and are geographically dispersed across the US. Although all these types of wells are included in
the dataset, their numbers in the dataset do not necessarily relate to their numbers in the population of
wells that will be affected by the GWR. In other words, these 1309 wells may not be perfectly
representative of the larger population of affected wells. Because the surveyed wells have not all been
characterized or characterized similarly for these features (system size, hydrogeologic setting,
vulnerability, etc.), EPA is not able to assess the degree to which the dataset may have overstated or
understated our national occurrence estimates.
5.7 Regulatory Alternatives
In addition to model runs to calculate benefits for the Final GWR requirements, analyses were
conducted for the other rule alternatives considered as part of the rule development process. The five
modeled regulatory conditions are: Baseline, Final GWR, Sanitary Survey and Corrective Action, Multi-
Barrier Approach, and Across-the-board Disinfection (see Chapter 3 for a full description of these
alternatives). The following exhibits present the estimated illnesses and deaths remaining for each
regulatory scenario (Exhibit 5.29), the reduction in illnesses and deaths for each of the regulatory
scenarios (Exhibit 5.30) and the annualized value of illnesses and deaths avoided for regulatory
alternatives (Exhibit 5.31). Benefits and costs of the Final GWR are compared in relation to these
alternatives in Chapter 8.
Economic Analysis for the October 2006
Final Ground Water Rule 5-112
-------
Exhibit 5.29 Remaining Number of Annual Viral Illnesses and Deaths
for Each Regulatory Alternative
Regulatory Alternative
Baseline
Final Rule
Risk Targeted Approach
Alternative 1
Sanitary Survey and
Corrective Action
Alternative 3
Multi-Barrier Approach
Alternative 4
Across-the- Board
Disinfection
Virus Type
Type A
Type B
Total
Type A
Type B
Total
Type A
TypeB
Total
Type A
TypeB
Total
Type A
TypeB
Total
Illnesses per Year
Mean
175,168
10,018
185,186
135,726
7,592
143,318
168,154
9,535
177,689
132,332
7,435
139,768
28,295
1,609
29,904
5th Percentile
32,652
501
33,153
22,559
320
22,879
31,065
470
31,535
21,207
307
21,514
5,248
80
5,328
95th Percentile
435,381
40,718
476,099
355,455
32,605
388,060
420,404
38,688
459,092
348,839
32,094
380,933
70,474
6,540
77,014
Deaths per Year
Mean
1.2
2.0
3.2
0.9
1.5
2.4
1.1
1.9
3.0
0.9
1.5
2.4
0.2
0.3
0.5
5th Percentile
0.2
0.0
0.3
0.1
0.0
0.2
0.2
0.0
0.2
0.1
0.0
0.2
0.0
0.0
0.0
95th Percentile
2.9
8.1
11.0
2.4
6.5
8.9
2.8
7.8
10.6
2.3
6.4
8.7
0.5
1.3
1.8
Notes: Details may not add to totals due to independent rounding and independent statistical analyses. Only endemic
illnesses are estimated. The figures presented in this exhibit represent only the quantifiable benefits of the GWR.
The nonquantified benefits are expected to comprise a significant portion of the overall benefits of the Rule and are
presented in Section 5.4.
Source: Appendix C
Exhibit 5.30 Comparison of Number of Annual Viral Illnesses and Deaths Avoided
for Regulatory Alternatives
Regulatory Alternative
Final Rule
Risk Targeted Approach
Alternative 1
Sanitary Survey and
Corrective Action
Alternative 3
Multi-Barrier Approach
Alternative 4
Across-the-Board
Disinfection
Illnesses per Year
Mean
41,868
7,497
45,419
155,282
5th Percentile
10,274
1,618
11,639
27,824
95th Percentile
88,039
17,007
95,166
399,085
Deaths per Year
Mean
0.7
0.1
0.8
2.7
5th Percentile
0.1
0.0
0.1
0.2
95th Percentile
2.1
0.4
2.3
9.2
Note: Details may not add to totals due to independent rounding. Only endemic illnesses are estimated.
Source: Appendix C.
Economic Analysis for the
Final Ground Water Rule
October 2006
5-113
-------
Exhibit 5.31 Annualized Value of Viral Illnesses and Deaths Avoided for
Regulatory Alternatives
Regulatory Alternative
3% Discount Rate
Mean
5th Percent! le
95th Percent! le
7% Discount Rate
Mean
5th Percent! le
95th Percent! le
Enhanced COI
Final Rule
Alternative 1
Alternative 3
Alternative 4
$19.7
$3.6
$21.3
$70.2
$6.5
$0.9
$7.1
$18.3
$45.4
$9.3
$48.7
$177.0
$16.8
$2.9
$18.2
$61.9
$5.5
$0.7
$6.0
$16.1
$38.6
$7.5
$41.6
$156.3
Traditional COI
Final Rule
Alternative 1
Alternative 3
Alternative 4
$10.0
$1.9
$10.8
$35.5
$2.2
$0.3
$2.5
$6.5
$27.0
$5.5
$28.9
$102.4
$8.6
$1.5
$9.3
$31.5
$1.9
$0.2
$2.1
$5.7
$22.9
$4.5
$24.8
$90.8
Notes: Detail may not add to totals due to independent rounding. The Traditional COI only includes valuation for medical costs and lost work time (including
some portion of unpaid household production). The Enhanced COI also factors in valuations for lost personal time (non-work time) such as childcare and
homemaking (to the extent not covered by the Traditional COI), time with family, and recreation, and lost productivity at work on days when workers are ill
but go to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified benefits are expected to comprise a significant
portion of the overall benefits of the Rule and are presented in Section 5.4.
Source: Appendix C
Economic Analysis for the
Final Ground Water Rule
October 2006
5-114
-------
6. Cost Analysis
6.1 Introduction
This chapter presents estimates for the total national and household costs for the Ground Water
Rule (GWR). To estimate the national costs of the GWR, the United States Environmental Protection
Agency (EPA or Agency) calculated the incremental cost for rule components that expand current State
practices (e.g., sanitary surveys) and the additional cost of new activities required under the rule. Cost
analyses include estimates to implement the rule, conduct sanitary surveys, perform triggered source
water monitoring, implement corrective actions (including drilling a new well, installation/operation of
treatment, etc.), and perform compliance monitoring. Assessment monitoring and hydrogeologic
sensitivity assessments (HSAs) are optional and are not included in the cost estimates.
System costs are estimated for different system types and size categories (nine size categories are
used based on population served, consistent with the Drinking Water Baseline Handbook (USEPA,
2001a)). State1 cost analyses include estimates of the labor burdens that States would face, including staff
training on GWR requirements, conducting sanitary surveys, reviewing monitoring reports, reviewing and
approving corrective action plans, and recordkeeping. EPA estimated unit costs for these various
components using cost models, equipment price lists and quotes, wage rates from government and
engineering sources (Bureau of Labor Statistics, R.S. Means, and States), stakeholder inputs, and other
relevant assumptions used in economic analyses performed for existing drinking water rules (e.g., Arsenic
Rule).
The national costs are estimated using a Monte-Carlo simulation model specifically developed for
the GWR. The GWR cost model was developed in a SAS° software platform utilizing a Monte-Carlo
simulation. The main advantage to this modeling approach is that, in addition to providing average
compliance costs, it also estimates the range of costs within each public water system (PWS) size and
type category. The GWR cost model allows for variability and uncertainty in PWS configuration, current
treatment in-place, and source water quality to be captured in the compliance cost estimates. This
information forms the basis for examining impacts to PWSs and technology affordability.
The remainder of this chapter is organized as follows:
• Section 6.2 describes the general costing and compliance assumptions used to estimate
national costs of the GWR.
• Section 6.3 describes the methodology of projecting costs over a 25-year period according to
the GWR compliance schedule, estimating the present value of each cost, and annualizing
each over a 25-year period.
Section 6.4 describes the methodology for developing costs for all rule activities.
1 The term "State" in the context of this chapter refers to any State or other primacy agency that has
oversight authority for drinking water programs.
Economic Analysis for the October 2006
Final Ground Water Rule 6-1
-------
Section 6.5 presents household cost estimates.
• Section 6.6 presents a discussion of nonqualified costs.
Section 6.7 presents a discussion of uncertainties in cost estimates.
• Section 6.8 presents the total annualized cost for the Final GWR.
Section 6.9 presents a comparison of cost estimates for the Final GWR to estimates for other
rule alternatives considered.
6.2 General Costing Assumptions and Methodology
The GWR Cost Model incorporates several baseline data elements, including the numbers, types,
and sizes of ground water systems in the United States, the percentage of ground water systems that
disinfect, and the percentage of disinfecting systems that attain 4-log treatment of viruses (using
inactivation, removal, or State-approved combination of these technologies) before or at the first
customer. Because many of the assumptions apply not to systems but to entry points, where appropriate,
exhibits in this chapter use entry point estimates. Derivations of these baselines for the GWR are
discussed in Chapter 4. In addition to those baseline elements, there are several additional baseline
costing assumptions used as inputs to the GWR Cost Model. The derivation of these inputs is discussed
in detail below.
6.2.1 Labor Rates
For costing purposes, EPA estimates the labor needs and hourly labor rates of systems and States
for two labor categories: managerial and technical. EPA recognizes that there may be significant
variation in labor rates across all PWSs. However, for purposes of this EA, and to implement national
policy, EPA uses national-level estimates fmm Labor Costs for National Drinking Water Rules (USEPA,
2003b). The technical and managerial wage rates vary with system size and include fringe benefits. The
technical and managerial wage rates (2003$) are shown in Exhibit 6.1.
Economic Analysis for the October 2006
Final Ground Water Rule 6-2
-------
Exhibit 6.1 Wage Rates by System Size
Loaded Wage Rate ($2003)
Technical Wage Rate
Managerial Wage Rate
Labor Cost (per hour)
System Size (Population Served)
25-100
$ 21.44
$ 44.36
$ 21.44
101-500
$ 23.09
$ 47.78
$ 23.09
500-3.3k
$ 24.74
$ 51.20
$ 24.74
3.3k-10k
$ 25.34
$ 51.20
$ 30.51
10k-100k
$ 26.05
$ 51.20
$ 31.08
>100k
$ 31.26
$ 51.20
$ 35.25
Notes: EPA estimates that systems with population greater than 3,300 use a combination of operators (technical)
and engineers (managerial), with an 80/20 ratio between the two, respectively. Loaded rate includes a 60
percent factor to account for the cost of fringe benefits.
Source: Labor Costs for National Drinking Water Rules (USEPA, 2003b).
To account for the general composition of staff at PWSs of smaller sizes (e.g., systems serving
3,300 people or fewer), EPA uses only the technical rate. For systems serving more than 3,300 people,
EPA uses a ratio of 80 percent technical labor to 20 percent managerial labor to arrive at a labor cost, or
weighted labor rate, of $30.51 for systems serving 3,301-10,000 people, $31.08 for systems serving
10,001-100,000 people, and $35.25 for systems serving greater than 100,000 people.
Labor costs attributable to States for administrative tasks are estimated based on an average
annual full time equivalent (FTE) labor cost, including overhead and fringe benefits, of $65,255 (2001$).
This rate was established based on data from the 2001 State Drinking Water Needs Analysis (ASDWA,
2001). For use in the GWR EA analyses, the $65,255 annual rate was updated to a year 2003 price level
($70,132) and converted to an hourly basis (1 FTE = 2,080 hours) to establish a State rate of $33.60 per
hour. For sanitary surveys and corrective action plan review, the year 1998 wage rate of $31.00 for a
field engineer is used. The field engineer rate comes from R.S. Means (1998) and includes a 60 percent
loading factor to account for the cost of fringe benefits. This wage rate is updated to 2003 dollars,
resulting in a field engineer rate of $37.34 per hour. Exhibit 6.2 displays these labor rates and their
derivations.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-3
-------
Exhibit 6.2 State Labor Rates
Cost Element
Field Engineer
State Employee
Base Hourly
Labor Cost
A
$ 31.00
$ 31.37
Ed in Year of
Data
B
$ 139.80
$ 153.20
EG 2003 Q4
C
$ 168.40
$ 164.10
2003 Labor
Cost
D=A*(C/B)
$ 37.34
$ 33.60
Sources: (A) Wage rate for a Field Engineer (1998$) from R.S. Means, 1998. Wage rate for State employee from 2001 State
Drinking Water Needs Analysis ($65,255 yearly for 2080 hours per year) (ASWDA, 2001)
(B & C) Employment Cost Index (ECI) for a Field Engineer from BLS (2003) from 1998 (Civilian; Total
compensation; Professional, specialty, and technical occupations). ECI for State Employee from BLS (2003) from
2001 (State and local government; Total compensation; Professional, specialty, and technical occupations).
www.bls.gov.
6.2.2 Laboratory Fees
A laboratory fee, or cost per sample, is associated with source water monitoring. For the purpose
of this cost analysis, EPA assumed that States will select E. coll as the indicator of fecal contamination for
source water analysis. Since States may designate alternative indicators for some sites (e.g., coliphage or
enterococci), and such analysis is more costly, this assumption may underestimate costs. EPA estimated
the cost of monitoring for both the use of an in-house and a commercial laboratory2, as shown in Exhibit
6.3. For in-house laboratories, EPA's estimate of the cost per sample includes the cost of laboratory
analysis materials ($8.95) and a total of 1.0 hour of the system staffs time to collect the sample and
conduct the analysis. For commercial laboratory analysis, EPA's estimate of the cost per sample includes
a shipping and commercial analysis fee ($74.80) and 0.5 hours of the system staffs time to collect the
sample and arrange for delivery to the laboratory. The estimated burden required to collect samples
includes travel time and reflects a national average. Individual systems may realize collection burden that
is either less than or greater than this average depending on the locations of sampling points in a
particular system. No additional costs are assumed for installation of a tap or re-piping of wells to permit
sampling, since EPA assumed all wells are equipped with existing taps for sampling.
Rates may vary due to regional variations in laboratory fees, the number of samples processed
(quantity discounts), and laboratory capacity. Although laboratory costs are often lower for multiple
samples, there are no estimates of the number of systems that may be able to take advantage of this
savings. Therefore, the rates used in this analysis may overestimate the actual costs incurred by systems.
2EPA assumed that systems serving fewer than 10,000 people would conduct 25% of the laboratory analyses in-house
and 75% would be sent to a commercial laboratory; for systems serving 10,000 to 50,000 people EPA assumed that 75% of the
laboratory analyses would be conducted in-house and 25% would be sent to a commercial laboratory; and systems serving
greater than 50,000 people were assumed to conduct all laboratory analyses in-house.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-4
-------
Exhibit 6.3 Source Water Monitoring Costs per Sample
Analysis Conducted
Sampling
Labor
Cost
(per hour)
A
Sampling
Labor Burden
(hours)
B
Total
Sampling
Cost
C=A*B
Analysis
Analysis
Labor Burden
(hours)
D
Operation &
Maintenance
E
Total
Analysis
Cost
F=(A*D)+E
Total
Burden
(hours)
G=B+D
Total
Cost
H=C+F
In-house
25-100
101-500
500-3.3K
3.3k-10k
10k-100k
>100k
$ 21.44
$ 23.09
$ 24.74
$ 30.51
$ 31.08
$ 35.25
0.5
0.5
0.5
0.5
0.5
0.5
$ 10.72
$ 11.55
$ 12.37
$ 15.26
$ 15.54
$ 17.62
0.5
0.5
0.5
0.5
0.5
0.5
$ 8.95
$ 8.95
$ 8.95
$ 8.95
$ 8.95
$ 8.95
$ 19.67
$ 20.50
$ 21.32
$ 24.21
$ 24.49
$ 26.57
1.0
1.0
1.0
1.0
1.0
1.0
$ 30.39
$ 32.04
$ 33.69
$ 39.46
$ 40.03
$ 44.20
Commercial laboratory
25-100
101-500
500-3.3k
3.3k-10k
10k-100k
>100k
$ 21.44
$ 23.09
$ 24.74
$ 30.51
$ 31.08
$ 35.25
0.5
0.5
0.5
0.5
0.5
0.5
$ 10.72
$ 11.55
$ 12.37
$ 15.26
$ 15.54
$ 17.62
0.0
0.0
0.0
0.0
0.0
0.0
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
$ 74.80
0.5
0.5
0.5
0.5
0.5
0.5
$ 85.52
$ 86.35
$ 87.17
$ 90.06
$ 90.34
$ 92.42
Sources: (A) Labor rates from Exhibit 6.1.
(B & D) Labor hours for sampling and analysis reflect EPA estimate.
(E) Operation and management costs based on best professional judgement. They are the cost of laboratory analysis materials for in-house
laboratories and the total cost for commercial laboratories, respectively.
6.2.3 Technology Unit Costs and Compliance Forecasts
EPA has assumed that systems may use a variety of existing technologies to achieve 4-log
treatment of viruses (using inactivation, removal, or State-approved combination of these technologies)
before or at the first customer. These technologies include the use of hypochlorination, chlorine gas
disinfection, ozonation and nanofiltration, chlorine dioxide, and anodic oxidants disinfection. Other
technologies or combinations of technologies (i.e., UV and chloramines, etc.) may be used to meet rule
requirements. However, technologies used in the compliance forecast (and their associated unit costs) are
based on representative use by the majority of systems. Unit cost estimates for these technologies are in
the form of "dollars per entry point" for initial capital and yearly operation and maintenance (O&M)
activities. EPA uses population-flow equations for each of the nine system size categories (see section
4.2.4) to estimate unit costs for each technology for each system type and size category. Population used
in population flow equations is based on the population served for a particular entry point. Detailed
explanations of the unit cost derivations for these technologies are presented in the Technology and Cost
Document for the Final Ground Water Rule (USEPA, 2006d).
Compliance forecasts (or technology selection forecasts) are estimates of which technologies
systems undergoing corrective action will use. Sections 6.4.6.1 and 6.4.6.2 provide detail on the
methodology used to generate compliance forecasts for sanitary survey corrective actions and source
water contamination corrective actions.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-5
-------
6.2.4 Cost Model
The GWR cost model uses the system baseline data and assumptions regarding labor hours,
laboratory costs, and labor rates to generate system costs for rule implementation, sanitary surveys,
triggered monitoring, and compliance monitoring. It also combines unit cost estimates with the predicted
number of entry points at which various corrective actions are predicted to be used to produce the cost of
correcting significant deficiencies and source water contamination. Finally, the model includes State
costs.
6.2.5 Modeled Variability and Uncertainty in National Costs
As noted throughout this EA, EPA recognizes that there is variability among many of the input
parameters to the GWR cost model (e.g., entry points per system, population served, flow per population,
labor rates, and occurrence distributions) and several rule compliance assumptions. In some cases, EPA
is able to describe this variability as distributions that are used as inputs to the cost model. In other cases,
there is insufficient information to fully characterize the distribution of variability on a national scale and
EPA uses mean values for these latter input parameters.
EPA also recognizes that there is uncertainty in the national cost estimates, and has characterized
the uncertainty around the mean unit technology costs (as described in section 6.4.9) in the GWR cost
model. There is also uncertainty built into the compliance assumptions regarding whether a system
chooses a treatment or nontreatment corrective action. To simulate the effect of this uncertainty on
national costs, the model performs a Monte-Carlo simulation. The results for the uncertainty analysis are
presented in the form of 90 percent confidence bounds around mean national cost estimates.
6.3 Projecting and Discounting National Costs
Costs must be expressed in common units so they can be added together to calculate total annual
costs and compared to benefits to compute net benefits. For this rule, some activities occur once, such as
installing new treatment technologies. Other O&M activities phase in as new technologies are installed,
then continue each year into the future. These activities do not occur instantly or simultaneously; to make
such values comparable, the year or years in which all costs are expended must be determined and the
costs must be brought back to their present value. For the purposes of this EA, one-time and yearly costs
were projected over a 25-year time period to coincide with the estimated life span of capital equipment
and a time lag of 5 to 10 years for treatment technology installation after rule promulgation. PWSs also
often finance their capital improvements over a 20-year period. The present values of costs are calculated
using discount rates of 3 and 7 percent based on EPA policy and Office of Information and Regulatory
Affairs of the Office of Management and Budget (OMB) guidance.3
3 The choice of an appropriate discount rate is a complex and controversial issue among economists and
policy makers. Therefore, the Agency compares streams of future national level costs and benefits using two
alternative discount rates, 3 and 7 percent. The underlying logic for each discount rate can be found in Guidelines
for Preparing Economic Analyses (USEPA, 2000e).
Economic Analysis for the October 2006
Final Ground Water Rule 6-6
-------
There are two adjustments made to the cost estimates in this EA. One adjustment is made when
costs are being used as part of the national cost estimate. These present value costs are then annualized
using the same discount rate so that the costs of each regulatory alternative can be directly compared with
the corresponding annual benefits. A summary of the steps in this adjustment is as follows:
Project all undiscounted costs (noncorrective action, corrective action, and State) over a 25-
year time horizon based on the rule implementation schedule.
• Calculate total present value costs using 3 and 7 percent discount rates (the same rates were
used as for the benefits calculation).
Annualize the costs over 25 years using the same discount rates.
Calculate ninety percent confidence bounds to reflect the bounds commonly used by
statisticians to assess the overall uncertainty and variability of modeled estimates.
Appendix D contains results from each step above for the final rule. Exhibits D.I through D.5
show the nominal costs projected over the rule schedule and the present value of each cost calculated to
the expected year of rule implementation for the final regulatory alternative. Exhibits D.6 through D.8
show the results for Alternatives 1,3, and 4.
A different adjustment is made when the cost estimates are used for the analysis of household-
level costs. In this case, rather than use a discount rate for determining the present value and annualized
costs, an after-tax cost-of-capital rate is used. This rate should reflect the true after-tax cost of capital
PWSs face, net of any government grants or subsidies. To annualize capital costs when determining the
costs to households, EPA uses different discount rates for private and public systems of different sizes
(annualized household costs are presented in Exhibit 6.31). The rate differences between systems
represent many factors (e.g., the different borrowing sources each type of system has available to it, bond
ratings, etc.) and vary from 5.20 to 6.27 percent depending on system size and ownership. These rates are
shown in Exhibit 6.4.
Exhibit 6.4 Discount Rates for Private and Public Systems
System Size
(Population Served)
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1,000,000
>1, 000,000
Public Rate
5.31%
5.31%
5.51%
5.51%
5.51%
5.20%
5.24%
5.24%
5.24%
Private Rate
6.22%
6.22%
6.22%
6.22%
6.22%
5.66%
6.27%
6.27%
6.27%
Source: Development of Cost of Capital Estimates for Public Water Systems, Final
Report (USEPA, 2000f).
Economic Analysis for the
Final Ground Water Rule
October 2006
6-7
-------
6.4 Derivation of Costs for Systems and States
This section presents the methodology and unit costs used to derive national costs for systems and
States to perform GWR related activities. Chapter 1 contains a summary of the GWRthat describes these
activities. The following subsections provide a brief summary of each activity and the assumptions used
to estimate the burden and costs attributable to both systems and States for each:
6.4.1 Rule Implementation and Annual Administration
6.4.2 Sanitary Surveys (SSs)
6.4.3 Triggered Monitoring (TM)
6.4.4 Corrective Actions (CAs)
6.4.5 Compliance Monitoring (CM)
This chapter uses information from the baseline analysis in Chapter 4 as a starting point for
analysis of PWSs subject to each rule requirement. Exhibits 6.5a-c present key baseline information and
intermediate model outputs that are referenced throughout this section. Because many of the assumptions
apply not to systems but to entry points, Exhibits 6.5a and 6.5b use both system and entry point estimates
where appropriate.
There are also 57 States that will incur costs as a result of the rule. As noted previously, the term
"State" in the context of this chapter refers to any State or other primacy agency that has oversight
authority for drinking water programs.
Economic Analysis for the October 2006
Final Ground Water Rule 6-8
-------
Exhibit 6.5a GWR Baselines: Number of Systems, Entry Points, and Wells
System Size
Total Number
of Systems
A
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
12,843
14,358
4,649
5,910
2,884
1,444
167
103
3
Number of
Entry Points
per System
B
1.3
1.6
2.0
2.4
3.2
5.6
11.3
12.4
11.4
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
9,456
6,758
1,894
715
73
10
1
1
-
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
64,448
18,993
1,940
585
74
19
1
1
-
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Number of
Wells per
System
c
1.5
2.0
2.3
3.1
4.6
9.8
16.1
49.9
49.9
1.5
2.0
2.3
3.1
4.6
9.8
16.1
49.9
49.9
1.5
2.0
2.3
3.1
4.6
9.8
16.1
49.9
49.9
Number of
Wells per
Entry Point
D = C/B
1.1
1.2
1.2
1.3
1.4
1.7
1.4
4.0
4.4
1.5
2.0
2.3
3.1
4.6
9.8
16.1
49.9
49.9
1.5
2.0
2.3
3.1
4.6
9.8
16.1
49.9
49.9
Entry Points
with at least 4
logs of Viral
Disinfection
E
3,996
8,873
3,547
5,378
3,547
3,856
583
545
34
850
608
170
64
7
1
0
0
-
1,160
342
35
11
1
0
0
0
-
Entry Points with
less than 4 logs
of Viral
Disinfection
F
3,689
8,191
3,274
4,964
3,274
3,559
538
503
-
1,892
1,352
379
143
15
2
0
0
-
10,441
3,077
314
95
12
3
0
0
-
Entry Points
without
Disinfection
G
9,168
6,343
2,262
4,002
2,459
698
770
227
-
6,714
4,798
1,345
508
52
7
1
1
-
52,847
15,574
1,591
480
61
16
1
1
-
Sources: (A) Exhibit 4.1, Column U
(B) Exhibit 4.3, Column A
(C) Wells per system from US EPA Drinking Water Baseline Handbook (2001).
(E) Exhibit 4.3, Column H
(F) Exhibit 4.3, Column W
(G) Exhibit 4.3, Column R
Economic Analysis for the
Final Ground Water Rule
6-9
October 2006
-------
Exhibit 6.5b Summary of Rule Implications
System Size
Systems
Receiving
Sanitary
Survey
A
Community Water Systems (CWSs
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
12,843
14,358
4,649
5,910
2,884
1,444
167
103
3
Systems with
Corrective
Actions for
Significant
Deficiencies
B
2,181
2,444
789
1,001
492
245
28
18
-
Entry Points
with
Triggered
Monitoring
c
12,797
14,819
5,578
8,910
5,638
4,357
1,295
749
-
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
9,456
6,758
1,894
715
73
10
1
1
-
1,608
1,148
322
121
12
2
0
0
-
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
64,448
18,993
1,940
585
74
19
1
1
-
10,990
3,234
329
99
13
3
0
0
-
8,609
6,149
1,724
651
66
9
1
1
-
63,295
18,648
1,905
574
73
19
1
1
-
Entry Points
with
Corrective
Actions for
Triggered
Monitoring
D
1,249
1,625
608
712
617
655
226
136
-
687
533
149
86
10
2
0
0
-
6,915
2,026
208
76
12
3
0
0
-
Entry Points with
Viral Disinfection
Increased from
less than 4 logs
to 4 logs
E
358
917
360
396
353
548
93
94
-
150
119
33
19
2
0
0
0
-
1,143
337
35
12
2
1
0
0
-
Previously Non-
disinfecting Entry Points
Taking Corrective Action
F
891
709
248
317
264
107
133
42
-
537
415
117
67
8
1
0
0
-
5,772
1,689
174
63
10
3
0
0
-
Entry Points
with
Incremental
Compliance
Monitoring
G
248
292
105
130
111
54
46
20
-
149
170
50
27
3
1
0
0
-
1,602
696
73
26
4
1
0
0
-
Sources: Cost Model Outputs
Notes:
(G) indicates number of entry points with treatment corrective actions.
(F) - (G) indicates non treatment corrective actions.
Economic Analysis for the
Final Ground Water Rule
6-10
October 2006
-------
Exhibit 6.5c Annualized Costs for Meeting Each of the GWR Provisions to
Systems and States ($Millions, 2003$)
Rule
Implementation &
Annual
Administration
A
Sanitary
Surveys
B
Corrective
Actions for
Significant
Deficiencies
c
Triggered
Monitoring
D
Corrective Actions
for Triggered
Monitoring
E
Compliance
Monitoring
F
Total Costs
G
3%
Systems
States
Total
Mean
Lower Bound
(5th %ile)
Upper Bound
(95th % Me)
Mean
Lower Bound
(5th %ile)
Upper Bound
(95th % Me)
Mean
Lower Bound
(5th %ile)
Upper Bound
(95th % Me)
$0.93
$0.93
$0.93
$9.20
$9.20
$9.20
$10.13
$10.13
$10.13
$0.21
$0.11
$0.31
$1.45
$0.66
$2.23
$1.66
$0.77
$2.54
$8.46
$5.74
$1 1 .60
$0.56
$0.52
$0.61
$9.02
$6.25
$12.21
$5.44
$5.32
$5.56
$0.09
$0.06
$0.12
$5.52
$5.38
$5.67
$25.64
$14.90
$38.39
$0.46
$0.32
$0.61
$26.10
$15.22
$39.00
$9.35
$3.02
$16.97
$0.00
$0.00
$0.01
$9.36
$3.02
$16.98
$50.02
$34.28
$68.76
$1 1 .77
$10.87
$12.64
$61 .79
$45.15
$81 .41
7%
Systems
States
Total
Mean
Lower Bound
(5th %ile)
Upper Bound
(95th % Me)
Mean
Lower Bound
(5th %ile)
Upper Bound
(95th % Me)
Mean
Lower Bound
(5th %ile)
Upper Bound
(95th % Me)
$1.33
$1.33
$1.33
$9.18
$9.18
$9.18
$10.51
$10.51
$10.51
$0.20
$0.10
$0.30
$1.39
$0.63
$2.14
$1.59
$0.74
$2.44
$8.13
$5.51
$11.14
$0.54
$0.50
$0.59
$8.67
$6.01
$1 1 .73
$5.39
$5.27
$5.51
$0.10
$0.07
$0.13
$5.48
$5.34
$5.64
$27.20
$15.89
$40.96
$0.52
$0.36
$0.69
$27.72
$16.26
$41 .65
$8.32
$2.65
$15.17
$0.00
$0.00
$0.01
$8.32
$2.65
$15.18
$50.57
$35.22
$69.00
$1 1 .74
$10.87
$12.61
$62.31
$46.09
$81.61
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
6.4.1 Rule Implementation and Annual Administration
PWSs
All systems subject to the GWR will incur one-time costs that include time for staff to read the
rule and become familiar with its provisions and to train employees on rule requirements. All systems
subject to the GWR will perform implementation activities; the number of systems performing
implementation activities is shown in column A of exhibit 6.5a. The technical and managerial labor rates
presented in section 6.2.1 are used along with estimates of labor hours to generate implementation costs
Economic Analysis for the
Final Ground Water Rule
October 2006
6-11
-------
for all systems. Technical rates apply to systems serving 3,300 or fewer people, and the 80/20 blend of
technical and managerial labor rates apply to systems serving populations greater than 3,300. Based on
previous experience with rule implementation, EPA estimates that systems will require a total of 3 hours
for the implementation activities associated with sanitary surveys, and a total of 2 - 4 hours for the
implementation of monitoring requirements. The planning and mobilization burden estimates under
monitoring implementation activities include time required to develop a sampling plan for source water
monitoring under triggered monitoring. These unit costs are presented in Exhibit 6.6.
Economic Analysis for the October 2006
Final Ground Water Rule 6-12
-------
Exhibit 6.6 PWS Unit Burden and Cost Estimates for Implementation Activities
System Size
(Population
Served)
Labor
Cost
(per hour)
A
Sanitary Surveys
Read and
Understand Rule
(hours/
system)
B
Planning
and Mobilization
(hours/
system)
C
Unit
Cost
D=A*(B+C)
Triggered and
Compliance Monitoring
Read and
Understand Rule
(hours/
system)
E
Planning
and
Mobilization
(hours/
system)
F
Unit
Cost
G=A*(E+F)
Total
Unit
Start-Up
Burden
(hours)
H=B+C+E+F
Total
Unit
Start-Up
Cost
I=A*H
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31 .08
$ 35.25
$ 35.25
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
$ 64.32
$ 69.27
$ 74.22
$ 74.22
$ 91.54
$ 93.24
$ 93.24
$ 105.74
$ 105.74
1
1
1
1
1
1
1
1
1
2
2
2
2
2
3
3
3
3
$ 64.32
$ 69.27
$ 74.22
$ 74.22
$ 91.54
$ 124.32
$ 124.32
$ 140.99
$ 140.99
6
6
6
6
6
7
7
7
7
$ 129
$ 139
$ 148
$ 148
$ 183
$ 218
$ 218
$ 247
$ 247
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
$ 64.32
$ 69.27
$ 74.22
$ 74.22
$ 91.54
$ 93.24
$ 93.24
$ 105.74
$ 105.74
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
$ 42.88
$ 46.18
$ 49.48
$ 49.48
$ 61.02
$ 62.16
$ 62.16
$ 70.50
$ 70.50
5
5
5
5
5
5
5
5
5
$ 107
$ 115
$ 124
$ 124
$ 153
$ 155
$ 155
$ 176
$ 176
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
100,001-1 Million
> 1 Million
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31 .08
$ 35.25
$ 35.25
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
$ 64.32
$ 69.27
$ 74.22
$ 74.22
$ 91.54
$ 93.24
$ 93.24
$ 105.74
$ 105.74
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
$ 42.88
$ 46.18
$ 49.48
$ 49.48
$ 61.02
$ 62.16
$ 62.16
$ 70.50
$ 70.50
5
5
5
5
5
5
5
5
5
$ 107
$ 115
$ 124
$ 124
$ 153
$ 155
$ 155
$ 176
$ 176
Notes: Detail may not add to totals due to independent rounding.
Sources: (A) Labor rates from Exhibit 6.1.
(B,C,E, & F) Labor hours for reading and understanding rule and planning and mobilization reflect EPA estimates.
Economic Analysis for the
Final Ground Water Rule
6-13
October 2006
-------
States
States will incur administrative costs while implementing the GWR. These implementation costs
are not directly required by specific provisions of GWR alternatives, but are necessary for States to ensure
the provisions of the GWR are properly carried out. States will need to allocate time for their staff to
establish and then maintain the programs necessary to comply with the GWR, including developing and
adopting State regulations and modifying data management systems to track new required system reports
to the States. For those GWR requirements that include monitoring with a laboratory method not
currently required by the State, the State must devote a portion of its staff time to certifying laboratories
for the new method. Time requirements for a variety of State agency activities and responses are
estimated in this EA and Exhibit 6.7a lists the activities required to start the program following
promulgation of the GWR along with their respective costs and burden.
In addition to these one-time costs, States will use resources to continue administrative activities.
On an annual basis, States must coordinate with their particular EPA Region to be certain that the State's
program is consistent with federal requirements. States will also continue to train State and PWS staffs,
maintain laboratories' certifications, and report system compliance information to the Safe Drinking
Water Information System (SDWIS). Exhibit 6.7b lists these annual activities with their respective costs
and burden.
States will also be required to spend time responding to PWSs with fecally contaminated ground
water sources or significant deficiencies. These costs are beyond any items specifically described and
costed as part of the cost model and include items such as time to provide additional consultation to
systems, prepare violation letters, and conduct data entry, etc. Because time requirements for
implementation and annual administration activities vary among State agencies, EPA recognizes that the
burden and cost estimates presented in Exhibits 6.7a and 6.7b may be an over- or under-estimate for some
States.
Economic Analysis for the October 2006
Final Ground Water Rule 6-14
-------
Exhibit 6.7a State Burden and Cost Estimates for Implementation Activities
Compliance Activity
Read and Understand Rule
Regulation Adoption and Program Development
Initial Laboratory Certification
Modify Data Management Systems
System Training and Technical Assistance
Staff Training
Labor Cost
(per hour)
A
$ 33.60
$ 33.60
$ 33.60
$ 33.60
$ 33.60
$ 33.60
Per State Total
National Totals (57 States/Primacy Agencies)
Hours
B
60
1,040
800
2,080
2,080
520
6,580
375,060
FTEs
C=B/2,080
0.03
0.50
0.38
1.00
1.00
0.25
Cost
D=A*B
$ 2,016
$ 34,946
$ 26,882
$ 69,892
$ 69,892
$ 17,473
$ 221,101
$ 12,602,743
Notes: Detail may not add due to independent rounding.
Sources: (A) Labor rate for state employee from Exhibit 6.2.
(B) Labor hours for start-up activities reflect EPA estimate.
(C) Full-time equivalent (FTE) assumes individual working 40 hours per week, 52 weeks per year.
Exhibit 6.7b State Burden and Cost Estimates for Annual Administration
Annual Administrative Activities
Compliance Activity
Coordination with EPA
Lab Certification
Ongoing Technical Assistance
SDWIS Reporting
Recordkeeping
Staff Training
Labor Cost
(per hour)
A
$ 33.60
$ 33.60
$ 33.60
$ 33.60
$ 33.60
$ 33.60
Per State Total
National Totals (57 States/Primacy Agencies)
Hours
B
1,040
1,040
1,040
1,040
880
104
5,144
293,208
FTEs
C=B/2,080
0.50
0.50
0.50
0.50
0.42
0.05
Cost
D=A*B
$ 34,946
$ 34,946
$ 34,946
$ 34,946
$ 29,570
$ 3,495
$ 172,848
$ 9,852,357
Notes: Detail may not add due to independent rounding.
Sources: (A) Labor rate for state employee from Exhibit 6.2.
(B) Labor hours for start-up activities reflect EPA estimate.
(C) Full-time equivalent (FTE) assumes individual working 40 hours per week, 52 weeks per year.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-15
-------
Annualized cost estimates for systems and States to perform implementation activities for the
preferred regulatory alternative are presented in Exhibit 6.8.
Exhibit 6.8 PWS and State Cost Estimates for Implementation and Annual
Administration Activities ($Millions, 2003$)
Annualized Costs for Implementation and Annual Administration Activities
3%
7%
Systems
Mean
$0.93
$1.33
Lower
Bound
(5th %ile)
$0.93
$1.33
Upper
Bound
(95th
%ile)
$0.93
$1.33
States
Mean
$9.20
$9.18
Lower
Bound
(5th %ile)
$9.20
$9.18
Upper
Bound
(95th
%ile)
$9.20
$9.18
Total
Mean
$10.13
$10.51
Lower
Bound
(5th %ile)
$10.13
$10.51
Upper
Bound
(95th
%ile)
$10.13
$10.51
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
6.4.2 Sanitary Surveys
PWSs
Under the sanitary survey provision of the GWR, community water systems (CWSs) and
noncommunity water systems (NCWSs) will undergo sanitary surveys once every three and five years,
respectively, to address eight specific components of a PWS. The exception to these frequencies is some
CWSs. Those CWSs that provide 4-log treatment of viruses (using inactivation, removal, or State-
approved combination of these technologies) before or at the first customer, or have an outstanding
performance record (i.e., no significant deficiencies) and no history of total coliform maximum
contaminant level (MCL) or monitoring violations under the Total Coliform Rule (TCR) may undergo
sanitary surveys every 5 years. To account for this EPA assumes, based on best professional judgement,
that the number of CWSs receiving sanitary surveys every 5 years is calculated by summing 50% of the
non-disinfecting and disinfecting to less than 4-log systems with the number of 4-log disinfecting
systems.
Sanitary surveys may increase either in scope or in frequency or in both scope and frequency
under the GWR for some systems. Systems not currently performing surveys at the frequencies specified
under the GWR will incur costs for performing additional full surveys (i.e., full sanitary survey costs).
The scope of sanitary surveys may also increase to completely address eight specific components of a
PWS. The sanitary surveys in this case are called incremental surveys (only having an incremental
increase in survey effort). Incremental surveys may also be applicable to some systems under the same
survey frequencies. Although States or designated agents perform the surveys, systems will incur costs to
accompany State inspectors during a review of the treatment plant and the distribution system, as well as
to review and discuss the sanitary survey report. The primary increase in costs that systems will incur as
a result of this rule provision is the additional number of surveys undertaken during the period of analysis
(i.e., full survey costs).
Economic Analysis for the
Final Ground Water Rule
October 2006
6-16
-------
EPA estimates that surveys for all PWSs currently average once every 5 years. Exceptions to this
frequency are those NCWSs that may be on a 10 year schedule as allowed under TCR, because they have
reliable disinfection and are deemed not vulnerable by the State. Thus, CWSs now required by the GWR
to receive sanitary surveys every 3 years will either undergo an increase in full surveys or incremental
surveys. For CWSs now on the three year schedule, the number of additional full surveys required to be
conducted is calculated as the difference between the frequency of conducting sanitary surveys on a three
year schedule and a five year schedule: 22 years/3 years - 22 years/5 years = 2.9 (assuming that the first
survey cycle starts as soon as compliance begins at the beginning of fourth year after the rule
promulgation and the rule stays effective through the 25th year). For systems on the five year schedule
that have to do incremental surveys, the number is 4.4 incremental surveys (22years/5 years).
NCWSs are either on a 5 year or a 10 year schedule as allowed for under the TCR. EPA assumes
that all NCWSs serving greater than 4,400 people are on a 10 year schedule. All other NCWSs are
assumed to be on a 5 year schedule. For those NCWSs on a 10 year schedule, the number of additional
full surveys required to be conducted is calculated as the difference between the frequency of conducting
sanitary surveys on a five year schedule and a ten year schedule: 22 years/5years - 22 years/10 years =
2.2. For systems on the five year schedule that have to do incremental surveys, the number is 4.4
incremental surveys (22years/5 years).
For any system incurring costs for incremental effort to comply with GWR requirements, EPA
assumes that the incremental effort will be 50% of current efforts. This assumption is based on the
expectation that many elements of a sanitary survey are similar to the requirements (compliance with
eight elements) of the GWR. In addition, based on examination of State sanitary survey requirements,
90% of systems are already required to comply with eight elements required under the GWR. Therefore,
only 10% of systems are anticipated to incur any incremental costs. These 50% and 10% factors are
applied to the unit cost estimates presented in Exhibits 6.11 and 6.12 to derive a weighted unit cost which
is then used in the cost model.
Exhibit 6.9 provides a schematic of the sanitary survey process for one system size and type to
show the process used in the cost model. Exhibit 6.10 details the number of full and incremental sanitary
surveys for systems on both the 3 and 5 year schedule. PWS unit costs to perform sanitary surveys are
provided in Exhibit 6.1 la for systems with treatment and in Exhibit 6.1 Ib for systems with no treatment.
Systems performing either full or incremental sanitary surveys are divided out by whether or not they
perform treatment because treatment status impacts unit costs applied in the cost model (i.e., additional
burden is estimated to inspect the system with treatment)4. All burden estimates used in the cost model
are based on consultations with EPA, State, and industry professionals with significant experience
conducting sanitary surveys on ground water systems. Lower unit costs are applied to NCWSs because
they are normally simpler than CWSs.
4 No differentiation in inspection burden is made between systems providing greater than or equal to 4-log
treatment or less than 4-log treatment.
Economic Analysis for the October 2006
Final Ground Water Rule 6-17
-------
Exhibit 6.9 Schematic of Sanitary Survey Process
(Numbers based on 3,301 -10,000 population category for CWSs)
SS Conducted by State
2,884 Systems
(1,102 systems disinfecting 4-log or
greater* 1,018 systems disinfecting
less than 4-log + 764 nondisinfecting
systems)
CWS [50% (systems not disinfecting H
systems disinfection < 4-log)]
891 Systems
Systems disinfecting to 4-log +
CWS [50% (systems not disinfecting + systems
disinfection < 4-log)]
1,993 Systems
New Full Survey due to a Higher
Frequency
891 Systems
891 Systems * 2.9 Surveys / System
= 2,584 Surveys
Incremental Survey for complete
coverage
891 Systems * 10% = 89 Systems
89 Systems * 4.4 Incremental Surveys /
System = 392 Incremental Surveys
-No-
1
Incremental Survey for complete
coverage
1 ,993 Systems * 10% = 199 Systems
199 Systems * 4.4 Incremental
Surveys /System
= 877 Incremental Surveys
Treatment
86.8% Disinfecting * 877
Systems = 761 Systems
Economic Analysis for the
Final Ground Water Rule
6-18
October 2006
-------
States
As required by this rule, systems must provide all of necessary information and assistance
requested from states for conducting surveys. The difference in costs between systems and States stems
from differing unit costs for performing sanitary surveys. State unit costs are significantly higher than the
system unit costs because the States are actually performing the surveys while systems are only required
to provide information and assistance to support the surveys. State unit costs to perform sanitary surveys
are provided in Exhibits 6.12a and 6.12b. As with the system unit costs estimates, all burden estimates
used in the cost model are based on consultations with EPA, State, and industry professionals with
significant experience conducting sanitary surveys on ground water systems. Lower unit costs are applied
to NCWSs because they are normally simpler than CWSs.
The total annualized costs estimates for both systems and States to perform sanitary surveys are
presented in Exhibit 6.13.
Economic Analysis for the October 2006
Final Ground Water Rule 6-19
-------
Exhibit 6.10 Number of Full and Incremental Sanitary Surveys for Systems
System Size
(Population Served)
Number of Systems
Receiving Sanitary
Survey
A
Number of
Additional Full
Surveys
B
Number of Incremental
Surveys for Systems
Performing Additional Full
Surveys
C
Number of Incremental
Surveys for Systems
Already on GWR
Schedule
D
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
1 00,000-1 M
>1, 000,000
12,843
14,358
4,649
5,910
2,884
1,444
167
103
3
2.9
2.9
2.9
2.9
2.9
2.9
2.9
2.9
2.9
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
1 00,000-1 M
>1, 000,000
9,456
6,758
1,894
715
73
10
1
1
0
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
NA
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
NA
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
1 00,000-1 M
>1, 000,000
64,448
18,993
1,940
585
74
19
1
1
0
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
NA
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
NA
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
NA
Note: Incremental surveys reflect CWSs moving from a 5 year to a 3 year schedule and NCWSs moving from a 10 year to
a 5 year schedule.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-20
-------
Exhibit 6.11 a PWS Unit Burden and Cost Estimates for Performing Full and
Incremental Sanitary Surveys (Treatment)
System Size
(Population
Served)
Labor
Cost
(per hour)
A
Review/
Inspect
Wells
B
Review/
Inspect
Treatment
C
Review/
Inspect
Distribution
System
D
Report
Review and
Discussion
w/ State
E
Total
Unit Burden
(hours)
F=sum(B-E)
Unit Cost
(Full Survey)
G=A*F
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
1 00,000-1 M
>1, 000,000
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
1.1
1.2
1.5
2.2
2.7
3.7
9.0
15.0
24.0
0.8
0.8
1.1
1.3
1.6
2.0
3.0
8.0
10.0
1.2
1.2
1.7
2.9
3.6
4.3
12.0
24.0
36.0
1.1
1.1
1.2
1.4
1.8
1.9
3.0
3.0
4.0
4.3
4.3
5.4
7.7
9.6
11.8
27.0
50.0
74.0
$ 92
$ 99
$ 135
$ 191
$ 291
$ 368
$ 839
$ 1 ,762
$ 1 ,762
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
1 00,000-1 M
>1, 000,000
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
1.0
1.0
1.1
1.1
1.5
1.3
1.5
8.0
NA
0.8
0.8
0.9
1.1
1.5
0.8
0.8
1.0
NA
1.0
1.1
1.3
1.2
1.7
1.8
2.3
10.0
NA
1.3
1.3
1.2
1.3
1.5
1.3
1.3
1.5
NA
4.0
4.2
4.5
4.7
6.2
5.0
5.8
20.5
NA
$ 87
$ 96
$ 110
$ 116
$ 188
$ 155
$ 179
$ 723
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
1 00,000-1 M
>1, 000,000
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
0.7
0.7
1.0
0.9
1.2
0.8
1.3
8.0
NA
0.6
0.6
0.8
1.0
1.3
0.5
0.5
1.0
NA
0.6
0.6
1.0
0.9
1.2
1.3
1.3
10.0
NA
0.9
0.9
0.9
1.1
1.2
0.8
0.8
1.0
NA
2.7
2.7
3.7
3.9
4.8
3.3
3.8
20.0
NA
$ 59
$ 63
$ 92
$ 96
$ 147
$ 101
$ 117
$ 705
NA
Weighted
Unit Cost
(Incremental Survey)
H=0.05*G
$ 5
$ 5
$ 7
$ 10
$ 15
$ 18
$ 42
$ 88
$ 88
$ 4
$ 5
$ 6
$ 6
$ 9
$ 8
$ 9
$ 36
NA
$ 3
$ 3
$ 5
$ 5
$ 7
$ 5
$ 6
$ 35
NA
Notes:
Weighted unit costs equal 5% of the unit costs.
10% of systems that do not already comply with
This factor accounts for 50% effort for an incremental survey and
rule requirements (see text discussion).
Economic Analysis for the
Final Ground Water Rule
6-21
October 2006
-------
Exhibit 6.11 b PWS Unit Burden and Cost Estimates for Performing Full and
Incremental Sanitary Surveys (No Treatment)
System Size
(Population
Seryed^^^
Labor
Cost
(per hour)
A
Review/
Inspect
Wells
B
Review/
Inspect
Distribution
System
C
Report
Review and
Discussion
w/ State
D
Total
Unit Burden
(hours)
E=sum(B-D)
Unit Cost
(Full Survey)
F=A*E
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50.001-100K
1 00,000-1 M
>1, 000,000
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
1.1
1.2
1.5
2.2
2.7
3.7
9.0
15.0
24.0
1.2
1.2
1.7
2.9
3.6
4.3
12.0
24.0
36.0
1.1
1.1
1.2
1.4
1.8
1.9
3.0
3.0
4.0
3.5
3.5
4.4
6.4
8.0
9.8
24.0
42.0
64.0
$ 75
$ 81
$ 108
$ 159
$ 243
$ 305
$ 746
$ 1 ,480
$ 1 ,480
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50.001-100K
1 00,000-1 M
>1, 000,000
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
1.0
1.0
1.1
1.1
1.5
1.3
1.5
8.0
NA
1.0
1.1
1.3
1.2
1.7
1.8
2.3
10.0
NA
1.3
1.3
1.2
1.3
1.5
1.3
1.3
1.5
NA
3.3
3.4
3.6
3.6
4.7
4.3
5.0
19.5
NA
$ 70
$ 79
$ 89
$ 89
$ 142
$ 132
$ 155
$ 687
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50.001-100K
1 00,000-1 M
>1, 000,000
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
0.7
0.7
1.0
0.9
1.2
0.8
1.3
8.0
NA
0.6
0.6
1.0
0.9
1.2
1.3
1.3
10.0
NA
0.9
0.9
0.9
1.1
1.2
0.8
0.8
1.0
NA
2.2
2.2
2.9
2.9
3.5
2.8
3.3
19.0
NA
$ 46
$ 50
$ 72
$ 72
$ 107
$ 85
$ 101
$ 670
NA
Weighted
Unit Cost
(Incremental Survey)
G=0.05*F
$ 4
$ 4
$ 5
$ 8
$ 12
$ 15
$ 37
$ 74
$ 74
$ 4
$ 4
$ 4
$ 4
$ 7
$ 7
$ 8
$ 34
NA
$ 2
$ 2
$ 4
$ 4
$ 5
$ 4
$C
O
$ 33
NA
Notes:
Weighted unit costs equal 5% of the unit costs. This factor accounts for 50% effort for an incremental survey and 10% of
systems that do not already comply with rule requirements (see text discussion).
Economic Analysis for the
Final Ground Water Rule
October 2006
6-22
-------
Exhibit 6.12a State Unit Burden and Cost Estimates
for Performing Full and Incremental Sanitary Surveys (Treatment)
System Size
(Population
Served)
Labor
Cost
(per hour)
A
Review/
Inspect
Wells
B
Review/
Inspect
Treatment
C
Review/
Inspect
Distribution
System
D
Report
Documenta
tion/
File Review
E
Report
Develop
ment
F
Data Entry
G
Report
Review and
Discussion
w/PWS
H
Travel
I
Total
Unit Burden
(hours)
J=sum(B-l)
Unit Cost
(Full Survey)
K=A*J
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
1 00,000-1 M
>1, 000,000
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
1.1
1.2
1.5
2.2
2.7
3.7
9.0
15.0
24.0
0.8
0.8
1.1
1.3
1.6
2.0
3.0
8.0
10.0
1.2
1.2
1.7
2.9
3.6
4.3
12.0
24.0
36.0
2.3
2.3
2.6
3.4
3.7
5.3
12.0
18.0
18.0
5.7
5.8
7.4
8.8
9.6
10.1
12.0
18.0
18.0
0.8
0.8
0.8
1.2
1.3
1.4
2.0
3.0
4.0
1.1
1.1
1.2
1.4
1.8
1.9
3.0
3.0
4.0
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
14.8
14.9
18.0
22.8
25.9
30.3
54.8
90.8
115.8
$ 551
$ 557
$ 671
$ 851
$ 967
$ 1,132
$ 2,044
$ 3,389
$ 3,389
Nontransient Noncomm unity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
1 00,000-1 M
>1, 000,000
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
NA
1.0
1.0
1.1
1.1
1.5
1.3
1.5
8.0
NA
0.8
0.8
0.9
1.1
1.5
0.8
0.8
1.0
NA
1.0
1.1
1.3
1.2
1.7
1.8
2.3
10.0
NA
1.9
2.0
2.1
2.1
2.2
2.5
2.5
8.0
NA
5.1
5.3
6.5
6.2
6.7
5.0
5.0
10.0
NA
1.0
1.0
0.8
0.8
0.8
0.8
0.8
1.0
NA
1.3
1.3
1.2
1.3
1.5
1.3
1.3
1.5
NA
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
13.8
14.2
15.6
15.6
17.6
15.0
15.8
41.3
NA
$ 515
$ 531
$ 583
$ 581
$ 657
$ 560
$ 588
$ 1 ,540
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10,001-SOK
50.001-100K
100.000-1M
>1, 000,000
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
NA
0.7
0.7
1.0
0.9
1.2
0.8
1.3
8.0
NA
0.6
0.6
0.8
1.0
1.3
0.5
0.5
1.0
NA
0.6
0.6
1.0
0.9
1.2
1.3
1.3
10.0
NA
1.5
1.5
1.8
1.7
1.5
1.3
1.3
3.0
NA
5.1
5.3
5.8
4.7
5.2
3.8
3.8
8.0
NA
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.5
NA
0.9
0.9
0.9
1.1
1.2
0.8
0.8
1.0
NA
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
11.9
12.1
13.9
12.9
14.1
10.8
11.3
33.3
NA
$ 443
$ 452
$ 518
$ 480
$ 526
$ 401
$ 420
$ 1 ,242
NA
Weighted Unit Cost
(Incremental Survey)
L=0.05*K
$ 28
$ 28
$ 34
$ 43
$ 48
$ 57
$ 102
$ 169
$ 169
$ 26
$ 27
$ 29
$ 29
$ 33
$ 28
$ 29
$ 77
NA
$ 22
$ 23
$ 26
$ 24
$ 26
$ 20
$ 21
$ 62
NA
Weighted unit costs equal 5% of the unit costs.
text discussion).
This factor accounts for 50% effort for an incremental survey and 10% of systems that do not already comply with rule requirements (see
Economic Analysis for the
Final Ground Water Rule
October 2006
6-23
-------
Exhibit 6.12b State Unit Burden and Cost Estimates
for Performing Full and Incremental Sanitary Surveys (No Treatment)
System Size
(Population
Served)
Labor
Cost
(per hour)
A
Review/
Inspect
Wells
B
Review/
Inspect
Distribution
System
C
Report
Documenta
tion/
File Review
D
Report
Develop
ment
E
Data Entry
F
Report
Review and
Discussion
w/PWS
G
Travel
H
Total
Unit Burden
(hours)
l=sum(B-H)
Unit Cost
(Full Survey)
J=A*I
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10.001-50K
50.001-100K
100,000-1 M
>1, 000,000
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
1.1
1.2
1.5
2.2
2.7
3.7
9.0
15.0
24.0
1.2
1.2
1.7
2.9
3.6
4.3
12.0
24.0
36.0
2.3
2.3
2.6
3.4
3.7
5.3
12.0
18.0
18.0
5.7
5.8
7.4
8.8
9.6
10.1
12.0
18.0
18.0
0.8
0.8
0.8
1.2
1.3
1.4
2.0
3.0
4.0
1.1
1.1
1.2
1.4
1.8
1.9
3.0
3.0
4.0
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
13.9
14.1
16.9
21.5
24.3
28.3
51.8
82.8
105.8
$ 521
$ 526
$ 631
$ 803
$ 909
$ 1,058
$ 1,932
$ 3,090
$ 3,090
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10.001-50K
50.001-100K
100,000-1 M
>1, 000,000
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
NA
1.0
1.0
1.1
1.1
1.5
1.3
1.5
8.0
NA
1.0
1.1
1.3
1.2
1.7
1.8
2.3
10.0
NA
1.9
2.0
2.1
2.1
2.2
2.5
2.5
8.0
NA
5.1
5.3
6.5
6.2
6.7
5.0
5.0
10.0
NA
1.0
1.0
0.8
0.8
0.8
0.8
0.8
1.0
NA
1.3
1.3
1.2
1.3
1.5
1.3
1.3
1.5
NA
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
NA
13.0
13.5
14.8
14.5
16.1
14.3
15.0
40.3
NA
$ 487
$ 503
$ 551
$ 540
$ 601
$ 532
$ 560
$ 1,503
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
10.001-50K
50.001-100K
100,000-1 M
>1, 000,000
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
NA
0.7
0.7
1.0
0.9
1.2
0.8
1.3
8.0
NA
0.6
0.6
1.0
0.9
1.2
1.3
1.3
10.0
NA
1.5
1.5
1.8
1.7
1.5
1.3
1.3
3.0
NA
5.1
5.3
5.8
4.7
5.2
3.8
3.8
8.0
NA
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.5
NA
0.9
0.9
0.9
1.1
1.2
0.8
0.8
1.0
NA
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
NA
11.3
11.5
13.0
11.9
12.8
10.3
10.8
32.3
NA
$ 421
$ 431
$ 487
$ 443
$ 476
$ 383
$ 401
$ 1,204
NA
Weighted Unit Cost
(Incremental Survey)
K=0.05*J
$ 26
$ 26
$ 32
$ 40
$ 45
$ 53
$ 97
$ 155
$ 155
$ 24
$ 25
$ 28
$ 27
$ 30
$ 27
$ 28
$ 75
NA
$ 21
$ 22
$ 24
$ 22
$ 24
$ 19
$ 20
$ 60
NA
Notes: Weighted unit costs equal 5% of the unit costs. This factor accounts for 50% effort for an incremental survey and 10% of systems that do not already comply with rule
requirements (see text discussion).
Economic Analysis for the
Final Ground Water Rule
October 2006
6-24
-------
Exhibit 6.13 PWS and State Cost Estimates for Sanitary Survey Performance
(SMillions, 2003$)
Annualized Costs for Sanitary Survey Activities
3%
7%
Systems
Mean
$0.21
$0.20
Lower
Bound
(5th %ile)
$0.11
$0.10
Upper
Bound
(95th
%ile)
$0.31
$0.30
States
Mean
$1.45
$1.39
Lower
Bound
(5th %ile)
$0.66
$0.63
Upper
Bound
(95th
%ile)
$2.23
$2.14
Total
Mean
$1.66
$1.59
Lower
Bound
(5th %ile)
$0.77
$0.74
Upper
Bound
(95th
%ile)
$2.54
$2.44
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
1 6.4.3 Triggered Source Water Monitoring
2 PWSs
3 Systems that do not achieve 4-log treatment of viruses (using inactivation, removal, or State-
4 approved combination of these technologies) before or at the first customer will be subject to triggered
5 source water monitoring. Under this provision, systems are required to collect and analyze samples at
6 the ground water source following the detection of total coliform (TC) in one or more samples collected
7 for compliance with the TCR.5 Systems are not required to conduct triggered source water monitoring if,
8 according to State criteria or a State determination, the cause of the total coliform-positive sample
9 collected under the TCR directly relates to the distribution system. While States have the option of
10 requiring the triggered monitoring samples to be tested for the presence of a State-specified fecal
11 indicator, for the purpose of this cost analysis, EPA assumed that States will select E. coll as the indicator
12 for analysis.
13 If a system detects the State-specified fecal indicator at its source, then the system must take five
14 additional samples within 24 hours unless the State determines that corrective action must be taken
15 immediately. If any of the additional samples is positive, the system must implement a corrective action.
16 Corrective actions fall into two major categories: non-treatment and treatment (see Section 6.4.4 for
17 detailed discussion). Several compliance estimates were considered to develop estimates of the cost
18 associated with triggered monitoring: the frequency with which systems will have to perform triggered
19 monitoring; the frequency that a TC positive is deemed to be related to the distribution system; the year in
20 which the triggered monitoring occurs; and the number of systems that are expected to test positive for
21 the fecal indicator. These factors are discussed in greater detail below. Exhibit 6.14 presents a schematic
22 of the triggered monitoring process as applied in the cost model.
5 If TC is detected in more than one sample collected at a single location during the same sampling event,
only one triggered monitoring sample is required to be taken from a water source directly related to that location.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-25
-------
1
2
Exhibit 6.14 Schematic of Triggered Monitoring Process
(Numbers based on 3,301 -10,000 population category for CWSs)
EPs not achieving 4-log virus treatment
5,638 EPs
(cost model output, see Exhibit 6.5b)
Triggered Monitoring
-Collect source water sample within 24 hours of receiving
initial notification of TC+
-If purchased system, notify wholesaler of TC+ within 24
hours. Wholesaler must must sample source water within
24 hours of receiving notification of TC+ sample result.
5,638 EPs * 0.58 samples / EP/year =
3,267 samples/year (approx.)
Yes
Has system tested
positive for the
fecal indicator?
617 EPs
(cost model output, see Exhibit 6.5b)
617 EPs/22 years = 28 EPs/year
No
5,021 EPs
See section 6.4.4 on
Corrective Action
Continue TM at
specified intervals
3 * Each positive sample corresponds to either one entry point predicted to have treatment corrective action or one well per entry
4 point to have a nontreatment corrective action. Systems must provide corrective actions if a source water positive sample is also
5 followed by a positive additional source water sample. The State, at its discretion, may require corrective action based on the
6 initial positive source water sample (See sections 6.4.3 and 6.6 for discussion).
7 Frequency of Performing Triggered Monitoring: EPA estimated the number of times per year
8 that a ground water system's total coliform sampling produced a positive result, which would trigger the
9 GWR source water monitoring requirements. The GWR allows States to determine which well within a
10 system may be related to each TC-positive. EPA makes the simplifying assumption that monitoring at
11 one EP within a system will approximate this rule provision. Due to the uncertainty in DV data, this
12 assumption may over- or under-estimate the number of EPs subject to triggered monitoring (see Appendix
13 I for further discussion). A summary of the methodology used to estimate the annual number of TC-
14 positive samples per system is presented in Section 4.2.7 and additional detail on the analysis is presented
Economic Analysis for the
Final Ground Water Rule
October 2006
6-26
-------
1
2
3
4
5
6
in Appendix I. The results presented in that Section 4.2.7 (Exhibit 4.11) are repeated in Exhibit 6.15
below for convenience. These estimates reflect the annual number of samples taken by systems of
different sizes and types. Sampling costs for triggered monitoring are calculated by multiplying the
annual number of triggered monitoring samples taken by the sampling unit costs presented previously in
Exhibit 6.3.
Exhibit 6.15 Estimated Number of Triggered Samples Per Year Per Entry Point
System
Type
cws
NTNCWS
TNCWS
System Size (Population Served)
<100
0.38
0.22
0.47
101-
500
0.41
0.23
0.48
501 -1K
0.49
0.28
0.60
1,001-
3,300
0.22
0.70
1.1
3,301-
10K
0.58
1.8
2.9
10,001-
50K
2.2
7.0
11.0
50,001-
100K
6.6
20.8
32.6
>100K
10.6
33.7
52.8
9
10
11
12
13
Source: Exhibit 4.11
14 Total Coliform Positive Samples Related to the Distribution System: Systems are not required to
15 conduct triggered source water monitoring if, according to State criteria or a State determination, the
16 cause of the total coliform-positive sample collected under the TCR directly relates to the distribution
17 system. Since it is difficult to predict the number of TC positives that are due to source water versus
18 distribution system contamination, EPA assumes each TC positive will result in a triggered source water
19 monitoring sample. The assumption likely overestimates triggered monitoring sample costs.
20 The Five Additional Samples Following an Initial Indicator Positive Sample: Neither the
21 schematic nor this cost analysis includes additional sampling (the cost model assumes that systems take
22 corrective action based on the first positive indicator sample result). EPA expects about 3% of the initial
23 triggered monitoring assays will be positive and therefore require additional sampling. The total number
24 of assays (triggered plus additional samples) would therefore be about 15% greater than the total number
25 of initial triggered monitoring samples that are included in the cost analysis. EPA believes that this factor
26 (understating the cost of triggered plus additional monitoring by $200,000 per year) is more than offset by
27 including costs for samples that are already collected and assayed under the Total Coliform Rule
28 (estimated value of more than $3 million per year).6 The GWR allows selected samples that are collected
29 under the TCR to be used in satisfying both the TCR repeat sample requirements and (per State approval)
30 the initial source water fecal indicator under this GWR. The costs of these assays can therefore be
31 attributed to the TCR rather than the GWR. By attributing all of these costs to the GWR, this analysis has
6 Estimated costs for additional triggered monitoring ($200,000) are calculated as the product of all entry
points (by system size and type) taking corrective action for triggered monitoring (Exhibit 6.5b, column D) and
sampling unit costs (Exhibit 6.3). Estimated costs savings from samples already taken under the TCR are calculated
as the product of all entry points serving 1,000 or fewer people (by system size and type) subject to triggered
monitoring (Exhibit 6.5b, column C), the average number of TC positive samples per year under TCR routine
monitoring (Exhibit 6.15), and sampling unit costs (Exhibit 6.3). Under both scenarios, costs are apportioned on an
annual basis and final costs are calculated as annualized present values using a three percent discount rate.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-27
-------
1 probably overestimated the triggered monitoring costs by more than $3 million. In addition, although
2 additional sampling will add to the total cost of monitoring, it will cause a decrease in the cost estimate
3 for providing corrective actions (i.e., some systems with an initial positive sample will not have any
4 positive additional samples). Since the costs of corrective actions are higher than the costs of additional
5 sampling, the net cost may be an overestimate. This overestimation coupled with the TCR-related
6 overestimation more than offsets the underestimation due to not including additional samples
7 (approximately $200,000) and the net effect is an overestimation of costs.
8 Percent of EPs Testing Positive: The triggered monitoring component of the model applies only
9 to the nondisinfecting subset of EPs (and any EPs that are applying disinfectant but not achieving 4-log
10 treatment of viruses using inactivation, removal, or State-approved combination of these technologies,
11 before or at the first customer). It is assumed that no triggered monitoring samples are taken during the
12 first three years following rule promulgations, but may be taken during any of the 22 years from year 4
13 through year 25.
14 For the triggered monitoring component of the cost analysis, each EP not achieving 4-log
15 treatment of viruses using inactivation, removal, or State-approved combination of these technologies,
16 before or at the first customer, goes through a 2-step process similar to that described in Chapter 5 for the
17 risk reduction model. In the first step of the process, estimates are made of the number of TC positives -
18 and therefore the number of source water indicator samples - that occur during the 22 years of the
19 compliance period between year 4 and year 25. The number of TC positives expected per year for each
20 EP of a given type and size are obtained from the DV data as described in chapter 4. The number of TC
21 positives per year, and therefore the expected number of triggered indicator samples taken per year, are
22 summarized in Exhibit 6.15. The total number of TC positives expected through year 25 is then
23 calculated as the number of TC positives per year times 22. If, for example, an EP is in the CWS size
24 10,000-50,000 category, the DV data indicate that these EPs average 2.21 TC positives per year; over 22
25 years, then, it is expected that each EP in this system size category will have 48.6 TC positives and
26 therefore take up to 49 source water indicator samples between years 4 and 25.
27 In the second step of the process, a simulation is performed to determine which, if any, of the
28 indicator samples taken through year 25 is the first fecal positive indicator result. In each uncertainty
29 loop of the cost model a set of values is selected reflecting the probability that the first positive of an
30 indicator will occur on a given assay number. Exhibit 4.21 showed the probability of the first fecal
31 indicator positive occurring on a given assay for the median and the 5th and 95th percentiles of a sample of
32 1,000 of these uncertainty sets of occurrence values.
33 The data for each uncertainty set provide the cumulative probabilities of observing the first fecal
34 positive on or before each specific assay number. In the cost model, a random value between 0 and 1 is
35 generated for each EP. That value is used as a look-up value to determine what assay number would
36 produce the first fecal positive.
37
38 For example, if the curve shown as the median data set in Exhibit 4.26 were the set of values
39 being used for a particular uncertainty loop, and the random number between 0 and 1 generated for an EP
40 in the CWS size 10,000-50,000 category were 0.095, the look-up function would indicate that the first
41 fecal indicator positive would occur on assay number 8. Since these EPs are expected to take 48.6 fecal
42 indicator assays over the 22 year period, the 8th assay would occur in the 4th year [(22 years/48.6
43 assays)*8 assays ~ 3.6]. Since there are no triggered monitoring samples taken in years 1 through 3
Economic Analysis for the October 2006
Final Ground Water Rule 6-28
-------
1 following promulgation, the 4th year of sampling corresponds to year 7 of the 25 year modeling period.
2 Therefore, this well would be "caught" by triggered monitoring in year 7.
3 EPs that test positive (both with initial and additional samples) for an indicator of fecal
4 contamination must perform corrective action. Costs for performing corrective actions are discussed in
5 Section 6.4.4. In addition, a report must be submitted to the State notifying them of the problem. EPA
6 estimates that this report will require, on average, 2.5 hours to complete and submit. EPA has developed
7 and systems will have access to automated forms that will minimize the burden to systems in complying
8 with this reporting requirement. Exhibit 6.16 below presents system unit costs for triggered monitoring
9 reporting requirements.
10
11 Based on the number of TC positives expected per EP across all EP types and sizes, together with
12 the expected values of fecal indicator positives as a function of assay number across all of the uncertainty
13 sets available to draw from for the simulation model, it can be estimated that approximately 10.2% (90%
14 confidence bounds of 7.4% - 13.4%) of all nondisinfecting EPs will have a fecal indicator positive from
15 the triggered monitoring provision of the rule by the 25th year of the modeling timeframe. The 10.2% of
16 all nondisinfecting EPs that are identified over the 25-year timeframe as having fecal indicator occurrence
17 comprise nearly 40% of the subset of those EPs that are expected to have fecal contamination at some
18 time. This is because the occurrence analysis shows that about 26.2% of all EPs are believed to have
19 some fecal contaminant occurrence. The estimate of 26.2% of EPs having some fecal contaminant
20 occurrence is based on the mean values of P2 and P3 from a sample of 10,000 estimates, where P2
21 (-16.5%) is the fraction of EPs having both viral and fecal indicator occurrence and P3 (-9.7%) is the
22 fraction having fecal indicator, but no viral occurrence. Therefore the fraction of all fecally contaminated
23 wells that are identified over the 25-year period can be estimated as 10.2% / 26.2% = 38.9% = -40%. See
24 Chapter 4, Section 4.3.4.1 for the derivation of P2, P3, and other viral and fecal indicator hit rate
25 parameters.
26 Invalidation of Samples: The GWR allows a State to invalidate a positive source water sample if
27 it believes that the positive is due to improper analysis. States may also invalidate positive source water
28 samples that are due to circumstances not reflecting source water quality. Systems must resample after a
29 sample is invalidated. For costing purposes, EPA, based on best professional judgement, estimates that
30 States will invalidate a minimal number of samples resulting in a negligible cost and burden.
Economic Analysis for the October 2006
Final Ground Water Rule 6-29
-------
Exhibit 6.16 PWS Unit Costs for Triggered Monitoring
System Size
(Population Served)
Entry Points
Triggered
Monitoring
A
Mean
Number
of Samples
per EP per
Year
B
Triggered
Monitoring
Samples/year
C=A*B
Average Fecal
Positive
Triggered
Monitoring
Entry
Points/year
D
Reporting/Positive Triggered Monitoring
Sample
Report
Prep
(hours)
E
Labor
Cost
(per
hour)
F
Unit
Cost
G=E*F
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
Totals
12,797
14,819
5,578
8,910
5,638
4,357
1,295
749
-
54,142
0.38
0.41
0.49
0.22
0.58
2.21
6.56
10.63
10.63
4,842
6,076
2,760
1,975
3,267
9,609
8,492
7,958
-
44,979
57
74
28
32
28
30
10
6
-
265
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 53.60
$ 57.73
$ 61 .85
$ 61 .85
$ 76.28
$ 77.70
$ 77.70
$ 88.12
$ 88.12
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
Totals
8,609
6,149
1,724
651
66
9
1
1
-
17,209
0.22
0.23
0.28
0.70
1.84
6.99
20.78
33.67
33.67
1,887
1,384
485
457
122
64
19
31
NA
4,447
31
24
7
4
0
0
0
0
NA
67
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 53.60
$ 57.73
$ 61 .85
$ 61 .85
$ 76.28
$ 77.70
$ 77.70
$ 88.12
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
Totals
Grand Total
63,295
18,648
1,905
574
73
19
1
1
-
84,515
155,867
0.47
0.48
0.60
1.10
2.88
10.97
32.60
52.83
52.83
29,605
8,955
1,144
633
209
205
32
52
NA
40,835
90,261
314
92
9
3
1
0
0
0
NA
420
752
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 53.60
$ 57.73
$ 61 .85
$ 61 .85
$ 76.28
$ 77.70
$ 77.70
$ 88.12
NA
Notes: Detail may not add to totals due to independent rounding.
NA Not applicable (no NCWSs of this size category).
Costs of repeat samples are not included. See Section 6.6 for discussion.
Source: (A) Number of entry points from Exhibit 6.5b.
(B) Mean triggered samples per system calculated from Exhibit 6.15.
(D) Values in Column D, Exhibit 6.5b are divided by 22 years to obtain postive triggered monitoring EPs/year.
(E) Labor hours for report preparation based on EPA experience with similar rule requirements.
(F) Labor rates from Exhibit 6.1.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-30
-------
1 States
2 State costs for triggered monitoring are assumed to be solely administrative. States incur costs to
3 review several paperwork requirements: reports of positive samples (including the results of additional
4 samples) and sample invalidation documentation. Each of these reports and associated costs is explained
5 below.
6 Review Reports of Source Water Positives: The GWR requires systems that find a source water
7 sample positive for fecal contamination to report this to the State. Based on its experience with similar
8 reporting requirements, EPA estimates that States will require 3.5 hours to review the report. Using a
9 labor cost of $33.60 per hour, EPA estimates a unit cost of $117.61 to review the report. This estimate is
10 greater than the burden estimated for systems to prepare and submit the report because it is anticipated
11 that the State will have less familiarity with any particular system and will be required to look up
12 additional historical information to make any assessments/determinations regarding the report.
13 Invalidation of Samples: As noted above, the GWR allows a State to invalidate a fecal positive
14 source water sample result that is due to improper analysis. States may also invalidate a fecal positive
15 source water sample results that are due to circumstances not reflecting source water quality. For costing
16 purposes, EPA estimates that States will invalidate a minimal number of samples, resulting in a negligible
17 cost and burden.
18 Annualized cost estimates for systems and States to perform triggered monitoring are presented in
19 Exhibit 6.17.
20 Exhibit 6.17 PWS and State Cost Estimates for Performing Triggered Monitoring
21 ($Millions, 2003$)
22
23
Annualized Costs for Triggered Monitoring Performance
3%
7%
Systems
Mean
$5.44
$5.39
Lower
Bound
(5th %ile)
$5.32
$5.27
Upper
Bound
(95th
%ile)
$5.56
$5.51
States
Mean
$0.09
$0.10
Lower
Bound
(5th %ile)
$0.06
$0.07
Upper
Bound
(95th
%ile)
$0.12
$0.13
Total
Mean
$5.52
$5.48
Lower
Bound
(5th %ile)
$5.38
$5.34
Upper
Bound
(95th
%ile)
$5.67
$5.64
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
Economic Analysis for the
Final Ground Water Rule
October 2006
6-31
-------
6.4.4 Corrective Actions
The purpose of conducting sanitary surveys is to identify significant deficiencies in PWSs that
need correction. Systems also conduct source water monitoring to detect fecal contamination in source
water. Identification of significant deficiencies as well as the detection of fecal indicators requires
systems to take corrective action under the GWR. National costs of these corrective actions comprise the
majority of the costs associated with the GWR. This section explains the methodology used for
estimating the national costs of implementing corrective actions under the GWR. This section includes
discussions of the compliance forecast used to determine the corrective actions undertaken by PWSs,
estimation of the burden and costs of corrective action plans, and the derivation of capital and O&M costs
associated with the various corrective actions. Where available, existing data on treatment practices were
used to determine compliance forecasts, burden, and cost estimates. In the absence of data, EPA used
best professional judgement based on consultations with Agency, industry, and State experts and
representative organizations (e.g., ASDWA) to make estimates.
6.4.4.1 Sanitary Survey Corrective Actions
Compliance Forecast
The GWR requires each PWS to correct any significant deficiencies found during a sanitary
survey. Because States have the authority to define significant deficiencies under the GWR, EPA
predicted the types of deficiencies that will be found and corrected as a result of the rule. EPA consulted
with experts from within the Agency and from States to develop a list of corrective actions to address
deficiencies that are likely to be identified in sanitary surveys of ground water systems (USEPA, 1996b).
Potential significant deficiencies identified can occur at the source of water, in a treatment plant, or in the
distribution system.
EPA lacks adequate data to quantify the number of significant deficiencies that will be detected
and corrected in the distribution system as well as in the treatment processes. Therefore, costs associated
with these deficiencies are not included in the cost model (see section 6.6 for further discussion of the
nonqualified costs). Similarly, the associated benefits are not included in the benefits model. To
estimate costs for significant deficiencies detected at or near the source, the following corrective actions
are used in the cost model.
Replace a sanitary well seal
• Rehabilitate an existing well
By limiting the corrective actions considered to the two options listed above, EPA has created a
simplified, but representative estimate of actions that may be taken by systems. In general, EPA believes
the costs for these significant deficiencies represent the range of costs systems would be expected to incur
for providing correction of significant deficiencies. Many other specific corrective actions may be taken
to adequately address significant deficiencies identified at or near the source. Some actions, such as
drilling a new well or purchasing water from another supplier, would be more expensive than these
options. However, based on discussions with experts, EPA believes that a majority of corrective actions
(e.g., fencing off or providing other limited access to infrastructure to protect wells) may actually be less
expensive than the two used in the cost model. Other well repairs that could correct source contamination
Economic Analysis for the October 2006
Final Ground Water Rule 6-32
-------
such as installation of pump block seals, pump block/well pad repair, or correcting runoff and drainage
problems would be less expensive, and as such, be the first consideration for any system looking to most
cost-effectively correct a problem. This is especially true for small systems which often have to address
such issues with limited resources.
To further account for uncertainty in the corrective actions that will be taken, PWSs are assigned
one of the two potential significant deficiencies listed above according to one of two probability
distributions. Exhibit 6.18 presents these distributions. Because the corrections of significant
deficiencies are dependent upon the deficiencies defined as significant by States and the conditions of
specific systems, both of which are highly variable, EPA used a high scenario/low scenario estimating
procedure to bound the cost estimates. The low-cost scenario assumes a greater percentage of the systems
with significant deficiencies will have deficiencies that are less expensive to correct (e.g., more systems
will have to replace their sanitary well seal than will have to perform a complete rehabilitation of their
well). This high/low bounding provides an estimate of the uncertainty with respect to the percentages of
each type of defect to be corrected.
Exhibit 6.18 Estimated Distribution of Significant Deficiency Corrective Actions
Corrective Action
Replace a sanitary well seal
Rehabilitate an existing well
Low Cost
Distribution
Percentage
A
60
40
Number
B
15,028
10,018
High Cost
Distribution
Percentage
c
40
60
Number
D
10,018
15,028
Source: (A,C) Distribution of corrective actions based on best professional judgement.
(B,D) based on (A,C), 147,330 total number of systems subject to the GWR (Exhibit 4.1), and 17% percent
of systems not constructed according to applicable State regulations (ASDWA survey).
The number of PWSs identifying a significant deficiency during a sanitary survey is determined
based on survey data from the Association of State Drinking Water Administrators (ASDWA) (1997).
Based on responses to the ASDWA survey, it was determined that 17% of systems had wells that were
not constructed according to applicable State regulations. This percentage is used as an estimate of the
number of systems that will find significant deficiencies at or near the source over the 25-year cost model
analysis period. Within the cost model, the assignment of significant deficiencies is applied equally in
years 4-25 of the analysis, resulting in approximately 0.77% of systems (17% / 22 years) being assigned
a corrective action in each of those years.
Unit Cost Estimates
Once a corrective action has been selected, a unit cost for that change is applied based on system
size. The costs for correction of significant deficiencies identified during sanitary surveys are dependent
upon the nature of the deficiency. Costs were developed based on the Technology and Cost Document for
the Final Ground Water Rule (USEPA, 2006d) to correct each of the identified significant deficiencies
Economic Analysis for the
Final Ground Water Rule
October 2006
6-33
-------
and are presented in Exhibit 6.19 below. These costs are considered one-time expenditures that occur in
the year the significant deficiency is found.
Exhibit 6.19 Estimated Unit Costs of Significant Deficiency Corrective Actions
Corrective Action
Replace a Sanitary
Well Seal
Rehabilitate an
Existing Well
Size Category (Population Served)
<100
$3,627
$11,986
101-500
$3,627
$11,986
501-
1,000
$3,627
$11,986
1,001-
3,300
$3,627
$11,986
3,301-
10,000
$3,627
$11,986
10,001-
50,000
$3,627
$11,986
50,001-
100,000
$3,627
$11,986
100,001-
1 Million
$3,627
$11,986
>1 Million
$3,627
$11,986
Source: GWR Technology and Cost Document
Corrective Action Plans
All systems are required to take corrective action for a significant deficiency that is identified
during sanitary surveys. Systems must consult with States and develop a corrective action plan, submit it
to the State for approval, and implement the corrective action (or combination of actions) approved by the
State. Exhibit 6.20a details the burden and cost to systems to prepare corrective action plans, as well as
the burden and cost to States to review them. EPA has developed and systems and States will have access
to automated forms that will minimize the burden to systems in complying with this reporting
requirement. These costs also apply for corrective actions due to source water contamination, discussed
below. It is assumed that large systems will have more technical and managerial staff compared to small
systems, and will therefore have less interaction with States than small systems. This assumption is
reflected in column D of Exhibit 6.20a; State burden is incurred on a 1:1 ratio to system burden for small
systems, and on a 0.5:1 ratio for large systems.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-34
-------
Exhibit 6.20a PWS and State Unit Costs for Corrective Action Plans
System Size
(Population Served)^
PWSs
Corrective
Action
Plan
(hours)
A
Plan
Labor
Cost
(per hour)
B
Unit
Plan
Cost
C=A*B
States
Review
Plan
(hours)
D
Review
Labor
Cost
(per hour)
E
Unit
Plan
Review
Cost
F=D*E
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
12
13
19
29
58
60
70
74
74
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 257.28
$ 300.17
$ 470.06
$ 717.46
$ 1,769.70
$ 1,864.80
$ 2,175.60
$ 2,608.35
$ 2,608.35
12
13
19
29
58
30
35
37
37
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 448.10
$ 485.44
$ 709.50
$ 1 ,082.92
$ 2,165.83
$ 1,120.26
$ 1,306.97
$ 1,381.65
$ 1,381.65
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
12
13
19
29
58
60
70
74
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 257.28
$ 300.17
$ 470.06
$ 717.46
$ 1,769.70
$ 1,864.80
$ 2,175.60
$ 2,608.35
NA
12
13
19
29
58
30
35
37
NA
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
NA
$ 448.10
$ 485.44
$ 709.50
$ 1 ,082.92
$ 2,165.83
$ 1,120.26
$ 1,306.97
$ 1,381.65
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
12
13
19
29
58
60
70
74
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 257.28
$ 300.17
$ 470.06
$ 717.46
$ 1,769.70
$ 1,864.80
$ 2,175.60
$ 2,608.35
NA
12
13
19
29
58
30
35
37
NA
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
$ 37.34
NA
$ 448.10
$ 485.44
$ 709.50
$ 1 ,082.92
$ 2,165.83
$ 1,120.26
$ 1,306.97
$ 1,381.65
NA
Notes: Detail may not add due to independent rounding.
NA= Not applicable (no NCWSs of this size category).
Sources: (A) Labor hours for preparing plan reflect EPA estimate based on best professional judgement.
(B) Labor rates from Exhibit 6.1.
(D) Labor hours for corrective action plan review reflect EPA estimate.
(E) Labor rate for field engineer from Exhibit 6.2.
Annualized cost estimates for systems and States to perform sanitary survey corrective actions
(including capital, O&M, and corrective action plan costs) are presented in Exhibit 6.20b.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-35
-------
Exhibit 6.20b PWS and State Cost Estimates for Sanitary Survey Corrective
Action Activities ($Millions, 2003$)
Annualized Costs for Sanitary Survey Corrective Action Activities
3%
7%
Systems
Mean
$8.46
$8.13
Lower
Bound
(5th %ile)
$5.74
$5.51
Upper
Bound
(95th
%ile)
$11.60
$11.14
States
Mean
$0.56
$0.54
Lower
Bound
(5th %ile)
$0.52
$0.50
Upper
Bound
(95th
%ile)
$0.61
$0.59
Total
Mean
$9.02
$8.67
Lower
Bound
(5th %ile)
$6.25
$6.01
Upper
Bound
(95th
%ile)
$12.21
$11.73
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
6.4.4.2 Source Water Contamination Corrective Actions
Compliance Forecast
As discussed in section 6.4.3, unless a system that detects fecal contamination in its source water
from triggered monitoring is directed by the State to take immediate corrective action, the systems must
collect and test an additional five source water samples for the presence of the same State-specified fecal
indicator within 24 hours. If any one of the five additional source water samples tests positive for the
State-specified fecal indicator (E. coli, enterococci, or coliphage), this rule requires the system to take
corrective action. For costing purposes, EPA assumes that the initial indicator positive sample will
correspond to either one entry point predicted to have corrective treatment or one well per entry point to
have a nontreatment corrective action. EPA believes this may underestimate treatment and nontreatment
corrective action costs if more than one entry point or well in a given system requires corrective action.
However, corrective action costs may be overestimated because if none of the five additional samples are
positive, no corrective action is required under the rule.
Similar to sanitary survey corrective actions, systems must consult with States and develop a
corrective action plan, submit it to the State for approval, and implement the corrective action (or
combination of actions) approved by the state (see Section 6.4.4.1 for estimates of corrective action plan
costs).
EPA assumed that corrective actions fall into four categories: fixing the deficiency that may lead
to contamination; eliminating the contamination from the source; obtaining an alternative source of water;
or providing disinfection treatment that achieves 4-log treatment of viruses (using inactivation, removal,
or State-approved combination of these technologies) before or at the first customer.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-36
-------
EPA considered four nontreatment alternatives in the cost estimate for eliminating contamination
identified during source water monitoring:
• Rehabilitate an existing well
• Drill a new well
Purchase water
Eliminate contamination source
For systems that employ nontreatment corrective actions, the GWR model assumes interim
disinfection will be installed until the nontreatment corrective action is completed and the cost model
includes the interim disinfection costs for this scenario. While interim disinfection would also apply to
systems installing disinfection as a permanent strategy prior to completion of the permanent strategy,
EPA assumes that the additional costs of interim disinfection for the systems projected to install
permanent disinfection as their corrective actions are insignificant because the long-term disinfection is
assumed in the model to occur immediately after the corrective action is required. Costs for setting up a
temporary chlorine disinfection process depend on the duration of interim disinfection, which is
determined by the nontreatment corrective action chosen. Based on best professional judgement, EPA
estimates that interim disinfection will be performed for a duration of one year for systems rehabilitating
an existing well, for two years for systems either drilling a new well or purchasing water, and for six
months for systems eliminating the source of contamination.
For systems that employ treatment as a corrective action, the Agency developed costs for the
following six disinfection technologies:
Hypochlorination (with and without additional storage)
Chlorine gas (with and without additional storage)
Chlorine dioxide (with and without additional storage)
Anodic oxidants7 (with and without additional storage)
• Ozone
• Nanofiltration
To estimate the corrective action chosen by each system, the cost model uses a five-step
compliance forecast. Systems may incur costs that differ from these if the State specifies an alternative
treatment technology (e.g., ultraviolet light). Each step is explained below, and summarized in Exhibits
6.21aand6.21b.
7 EPA estimated the cost of adding Anodic oxidation in the Technology and Cost Document for the Final
Ground Water Rule (USEPA, 2006d) as a representative technology for mixed oxidants.
Economic Analysis for the October 2006
Final Ground Water Rule 6-37
-------
Step 1—Current Treatment Practices: The first step divides entry points requiring corrective
action for source water contamination into those that already have some treatment in place (i.e., <4-log)
and those without treatment. EPA assumes that the ratio of correcting entry points that have disinfection
treatment in place to nondisinfecting entry points performing a corrective action is identical to the ratio of
entry points that disinfect but do not achieve 4-log treatment of viruses (using inactivation, removal, or
State-approved combination of these technologies) before or at the first customer to nondisinfecting entry
points, by system type and size category. Entry points that have some disinfection treatment but do not
achieve 4-log are estimated to choose their corrective action based on the distribution in step five.
Step 2—Current Implementation of Treatment Types: For entry points which do not currently
apply disinfection, EPA assumes that the percentage of entry points that will choose disinfection as a
corrective action for source water contamination is based on a range. The high end of the range is based
on the percentage of CWS entry points currently employing disinfection by system size using information
from the Community Water Systems Survey (CWSS) (USEPA, 1997a). The low end of the range is
assumed to be 10% based on discussions with State representatives, who indicated that the use of the
CWSS data would overestimate the percentage of entry points using disinfection in response to
fecally-contaminated source water. The remaining percentage of non-disinfecting entry points are
predicted to employ nontreatment corrective actions in response to fecal contamination of the source
water. Because of the uncertainty inherent in projecting the number of entry points that would employ
each nontreatment corrective action, EPA assigns equal proportions of the nontreatment corrective actions
into the high- and low-cost scenarios for significant deficiencies.
Step 3—Distribution of Nontreatment Corrective Actions: Each non-disinfecting entry point that
is predicted to require a nontreatment corrective action (from step two), having been assigned to either the
high- or the low-cost scenario, is then assigned a corrective action according to the corresponding
percentages in that distribution, as shown in Exhibit 6.2Ib.
Step 4 - Distribution of Treatment Corrective Actions: The compliance forecast assigns each non-
disinfecting entry point predicted to require a treatment corrective action (from step two) to one of the ten
possible treatment scenarios based on the percentage of CWSs currently engaged in those treatment
practices, which are estimated with the CWSS results (USEPA 1997a).
Step 5 - Distribution of Corrective Actions for Disinfecting Entry Points: Finally, for those entry
points with disinfection that do not achieve 4-log treatment of viruses (using inactivation, removal, or
State-approved combination of these technologies) before or at the first customer, the compliance forecast
assigns them a corrective action (i.e., add storage or increase disinfectant dose). EPA bases this
probability distribution on AWWA survey data (AWWA, 1998).
Economic Analysis for the October 2006
Final Ground Water Rule 6-38
-------
Exhibit 6.21 a Summary Flow Chart
Estimated Distribution of Source Water Contamination Corrective Actions
(Numbers based on 3,301 -10,000 population category for CWSs)
Step #2
Step #1
Water System Entry Points (EP's)
Taking Corrective Action Due to
Source Water Contamination
617 EP
Sources: SDWIS
CWSS
AWWA
Nondisinfecting
42.9%*617EP=265EP
Treatment
Action
41.8%*265EP
=111EP
Hypochlori nation
27.5%*111EP=30EP
plus storage
21.6%*111EP=24EP
Chlorine gas
27.5%*111EP=30EP
plus storage
21.6%*111EP=24EP
Chlorine dioxide
0%*111EP=OEP
plus storage
0%*111EP=OEP
Anodic oxidants
0.6%111EP=1EP
plus storage
0.5%*111EP=1EP
Ozone
0%*111EP
=OEP
Nanofiltration
0.6%*111EP
=1EP
Nontreatment
Action
58.2%*265EP=154EP
Step #3
Rehabilitate an
existing well
35.0%*154EP
=53 EP
Drill a new well
25.0%*154EP
=39 EP
Purchase water
15.0%*154EP
=23EP
Eliminate
contamination
source
25.0%*154EP
=39EP
Rehabilitate an
existing well
50.0%*154EP
=77EP
Drill a
new well
15.0%*154EP
=23EP
Purchase
Water
5.0%*154EP
=8EP
Eliminate
contamination
Source
30.0%*154EP
=46 EP
Step #4 I
! Step #5 \
Economic Analysis for the
Final Ground Water Rule
6-39
October 2006
-------
Exhibit 6.21 b Estimated Distribution of Source Water Contamination Corrective
Actions (continued on next page)
Step 1: Current Treatment Practice
System Type/
Disinfection Practice
CWSs
Disinfecting less than 4 log
Nondisinfecting
NTNCWSs
Disinfecting less than 4 log
Nondisinfecting
TNCWSs
Disinfecting less than 4 log
Nondisinfecting
System Size (Population Served)
<100
28.7%
71 .3%
22.0%
78.0%
16.5%
83.5%
101-500
56.4%
43.6%
22.0%
78.0%
16.5%
83.5%
501-1,000
59.1%
40.9%
22.0%
78.0%
16.5%
83.5%
1,001-
3,300
55.4%
44.6%
22.0%
78.0%
16.5%
83.5%
3,301-
10,000
57.1%
42.9%
22.0%
78.0%
16.5%
83.5%
10,001-
50,000
83.6%
16.4%
22.0%
78.0%
16.5%
83.5%
50,001-
100,000
41 .2%
58.8%
22.0%
78.0%
16.5%
83.5%
>1 00,000
68.9%
31.1%
22.0%
78.0%
16.5%
83.5%
Source: Derived from Exhibit 4.3. Taken as percentages of all entry points not achieving 4-log disinfection.
Step 2: Current Implementation of Treatment Types
Category of
Corrective Action
Nontreatment action
High cost distribution
Low cost distribution
Treatment action
System Size (Population Served)
<100
54.4% - 90%
36.1%
36.1%
10% -45.6%
101-500
27.1% -90%
29.3%
29.3%
10% -72. 9%
501-1,000
24.9% - 90%
28.7%
28.7%
10% -75.1%
1,001-
3,300
27.9% - 90%
29.5%
29.5%
10% -72.1%
3,301-
10,000
26.5% - 90%
29.1%
29.1%
10% -73.5%
10,001-
50,000
8.6% - 90%
24.7%
24.7%
10% -91 .4%
50,001-
100,000
40.7% - 90%
32.7%
32.7%
10% -59.3%
>1 00,000
1 1 .8% - 90%
25.5%
25.5%
10% -88.2%
Note: EPA assumes that systems points will make treatment corrective action in proportion to entry points' current disinfection practices.
Source: Derived from Exhibit 4.3.
Step 3: Distribution of Nontreatment Corrective Actions
Corrective Action
High Cost Distribution
Rehabilitate an existing well
Drill a new well
Purchase water
Eliminate contamination source
Low Cost Distribution
Rehabilitate an existing well
Drill a new well
Purchase water
Eliminate contamination source
System Size (Population Served
<100
46.3%
27.8%
5.6%
20.4%
55.6%
18.5%
1 .9%
24.1%
101-500
33.3%
40.0%
10.0%
16.7%
50.0%
20.0%
3.3%
26.7%
501-1,000
33.3%
25.0%
12.5%
29.2%
50.0%
12.5%
4.2%
33.3%
1,001-
3,300
33.3%
25.0%
12.5%
29.2%
50.0%
12.5%
4.2%
33.3%
3,301-
10,000
35.0%
25.0%
15.0%
25.0%
50.0%
15.0%
5.0%
30.0%
10,001-
50,000
46.2%
30.8%
0.0%
23.1%
61 .5%
15.4%
0.0%
23.1%
50,001-
100,000
50.0%
26.2%
0.0%
23.8%
59.5%
19.0%
0.0%
21.4%
>1 00,000
50.0%
37.5%
0.0%
12.5%
62.5%
25.0%
0.0%
12.5%
Note: Percentage of those entry points performing nontreatment corrective action.
Economic Analysis for the
Final Ground Water Rule
6-40
October 2006
-------
Exhibit 6.21 b Estimated Distribution of Source Water Contamination Corrective
Actions (continued)
Step 4: Distribution of Treatment Corrective Actions
Corrective Action
Hypochlorination
Hypochlorination plus storage
Chlorine gas
Chlorine gas plus storage
Chlorine dioxide
Chlorine dioxide plus storage
Anodic oxidants
Anodic oxidants plus storage
Ozone
Nanofiltration
System Size (Population Served)
<100
55.1%
43.3%
0.0%
0.0%
0.7%
0.6%
0.2%
0.1%
0.0%
0.0%
101-500
54.8%
43.0%
0.0%
0.0%
0.6%
0.4%
0.3%
0.2%
0.0%
0.7%
501-1,000
56.0%
44.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
1,001-
3,300
55.3%
43.4%
0.0%
0.0%
0.0%
0.0%
0.4%
0.3%
0.0%
0.6%
3,301-
10,000
27.5%
21.6%
27.5%
21.6%
0.0%
0.0%
0.6%
0.5%
0.0%
0.6%
10,001-
50,000
0.0%
0.0%
53.8%
42.3%
0.3%
0.3%
1 .5%
1 .2%
0.0%
0.6%
50,001-
100,000
0.0%
0.0%
54.1%
42.5%
1 .7%
1 .4%
0.0%
0.0%
0.0%
0.3%
>1 00,000
0.0%
0.0%
55.7%
43.7%
0.0%
0.0%
0.0%
0.0%
0.6%
0.0%
Note: EPA assumes that systems will choose treatment in proportion to current treatment practices. Estimates for hypochlorination and chlorine gas based on
remainder of entry points not performing other treatment practices.
Source: CWSS, Table 1-23 (1997) data. Nanofiltration substituted for all membranes based on professional engineering judgement.
Step 5: Distribution of Corrective Actions for Disinfecting Entry Points
Corrective Action
Add storage
Increase dose - hypochlorination
Increase dose - chlorine gas
System Size (Population Served)
<100
39%
61%
0%
101-500
39%
61%
0%
501-1,000
39%
61%
0%
1,001-
3,300
39%
61%
0%
3,301-
10,000
39%
61%
0%
10,001-
50,000
39%
0%
61%
50,001-
100,000
39%
0%
61%
>1 00,000
39%
0%
61%
Note: EPA assumes that systems add storage or increase dose based on AWWA study.
Unit Cost Estimates
Unit cost estimates for the four nontreatment corrective actions and interim disinfection as
described above are presented in Exhibit 6.22a. Unit costs for installing and operating each of the six
treatment scenarios (and variations) in the compliance forecast are presented in Exhibit 6.22b. For each
corrective action, costs generally increase corresponding to system size. However, some corrective
actions are assumed to be constant (on average) over the range of system sizes (e.g., costs for actions such
as rehabilitating a well or drilling a new well are heavily influenced by factors such as hydrogeologic
setting and well depth that are independent of system size). For further description of the assumptions
and methodologies used to develop all corrective action unit costs see the Technology and Cost Document
for the Final Ground Water Rule (USEPA, 2006d).
Economic Analysis for the
Final Ground Water Rule
October 2006
6-41
-------
Exhibit 6.22a Estimated Unit Costs of Nontreatment Corrective Actions
for Source Water Contamination
Corrective Action
Size Category (Population Served)
<100
101-500
501-
1,000
1,001-
3,300
3,301-
10,000
10,001-
50,000
50,001-
100,000
100,001-
1 Million
>1 Million
Nontreatment Corrective Actions
Rehabilitate an
Existing Well
Drill a New Well
Purchase Water
Capita
O&M ($ per kgal)
Eliminate Source of
Contamination
$11,986
$30,172
$1 1 ,986
$30,172
$11,986
$30,172
$11,986
$30,172
$11,986
$30,172
$1 1 ,986
$30,172
$1 1 ,986
$30,172
$11,986
$30,172
$1 1 ,986
$30,172
$173,180
$1.12
$16,533
$173,180
$1.18
$16,533
$198,599
$0.63
$16,533
$198,599
$1.44
$16,533
$242,618
$2.09
$16,533
$242,618
$1.35
$16,533
$353,697
$1.39
$16,533
$390,999
$0.91
$16,533
$390,999
$0.91
$16,533
Interim Disinfection
Rehabilitate an
Existing Well
Capital
Total O&M
Drill a New Well
Capita
Total O&M
Purchase Water
Capital
Total O&M
Eliminate Source of
Contamination
Capita
Total O&M
$1,874
$2,636
$1,874
$5,272
$1,874
$5,272
$1,874
$1,318
$1,874
$2,887
$1,874
$5,774
$1,874
$5,774
$1,874
$1,444
$1 ,874
$3,356
$1 ,874
$6,712
$1 ,874
$6,712
$1 ,874
$1 ,678
$1 ,874
$3,936
$1 ,874
$7,871
$1 ,874
$7,871
$1 ,874
$1 ,968
$1,874
$4,259
$1,874
$8,517
$1,874
$8,517
$1,874
$2,129
$1 ,874
$5,210
$1 ,874
$10,420
$1 ,874
$10,420
$1 ,874
$2,605
$1,971
$7,579
$1,971
$15,158
$1,971
$15,158
$1,971
$3,790
$2,302
$19,177
$2,302
$38,354
$2,302
$38,354
$2,302
$9,589
$20,774
$135,513
$20,774
$271 ,026
$20,774
$271 ,026
$20,774
$67,757
Note: Based on best professional judgement, EPA estimates that interim disinfection will be performed for a duration of one year for
systems rehabilitating an existing well, for two years for systems either drilling a new well or purchasing water, and for six months for
systems eliminating the source of contamination.
Source: Technology and Cost Document for the Final Ground Water Rule (USEPA, 2005d)
Economic Analysis for the
Final Ground Water Rule
October 2006
6-42
-------
Exhibit 6.22b Estimated Unit Costs of Treatment Corrective Actions for Source
Water Contamination
Corrective Action
System Size (Population Served)
<100
101-500
501-
1,000
1,001-
3,300
3,301-
10,000
10,001-
50,000
50,001-
100,000
100,001-
1 Million
> 1 Million
Systems Adding Treatment
Chlorine gas feed
capital cost
Chlorine gas feed
annual O&M cost
Chlorine gas feed & storage
capital cost
Chlorine gas feed & storage
annual O&M cost
Hypochlorite feed
capital cost
Hypochlorite
annual O&M cost
Hypochlorite feed & storage
capital cost
Hypochlorite & storage
annual O&M cost
Chlorine Dioxide System
capital cost
Chlorine Dioxide
annual O&M cost
Chlorine Dioxide System
& storage capital cost
Chlorine Dioxide
& storage annual O&M cost
Anodic Oxidant
capital cost
Anodic Oxidant
annual O&M cost
Anodic Oxidant & storage
capital cost
Anodic Oxidant & storage
annual O&M cost
Ozonation
capital cost
Ozonation
annual O&M cost
Nanofiltration
capital cost
Nanofiltration
annual O&M cost
$ 29,868
$ 6,192
$ 31,216
$ 6,192
$ 8,970
$ 1,585
$ 10,318
$ 1,585
N/A
N/A
N/A
N/A
$ 47,219
$ 2,911
$ 48,568
$ 2,911
N/A
N/A
$ 62,691
$ 7,520
$ 29,868
$ 6,227
$ 33,354
$ 6,227
$ 8,970
$ 2,076
$ 12,456
$ 2,076
N/A
N/A
N/A
N/A
$ 65,151
$ 5,471
$ 68,637
$ 5,471
N/A
N/A
$ 104,856
$ 10,253
$ 29,868
$ 6,307
$ 37,960
$ 6,307
$ 15,072
$ 4,180
$ 23,164
$ 4,180
$ 35,011
$ 15,261
$ 46,196
$ 16,251
$ 87,450
$ 7,480
$ 95,543
$ 7,480
$ 347,027
$ 55,668
$ 182,768
$ 20,140
$ 29,868
$ 6,456
$ 46,039
$ 6,456
$ 24,402
$ 6,582
$ 40,573
$ 6,582
$ 39,299
$ 16,897
$ 61,792
$ 17,720
$ 110,256
$ 9,791
$ 126,427
$ 9,791
$ 431,809
$ 59,028
$ 304,122
$ 37,037
$ 29,868
$ 6,857
$ 66,058
$ 6,857
$ 24,402
$ 7,326
$ 60,593
$ 7,326
$ 42,363
$ 17,901
$ 89,439
$ 18,733
$ 151,129
$ 12,855
$ 187,320
$ 12,855
$ 622,023
$ 60,789
$ 573,460
$ 63,670
$ 58,781
$ 16,951
$ 152,883
$ 16,951
$ 72,631
$ 7,558
$ 166,733
$ 7,558
$ 80,836
$ 19,878
$ 191,678
$ 20,392
$ 255,055
$ 17,479
$ 349,157
$ 17,479
$ 903,927
$ 63,718
$ 1,086,398
$ 133,397
$ 65,006
$ 18,197
$ 266,275
$ 18,197
$ 79,658
$ 7,909
$ 280,927
$ 7,909
$ 82,091
$ 21,705
$ 307,085
$ 22,257
$ 354,880
$ 22,181
$ 556,149
$ 22,181
$ 1,175,442
$ 67,004
$ 1,872,457
$ 194,361
$ 96,958
$ 21,854
$ 541,906
$ 21,854
$ 96,180
$ 19,177
$ 541,128
$ 19,177
$ 202,017
$ 25,983
$
$
$ 745,098
$ 38,439
$ 1,190,046
$ 38,439
$ 1,991,127
$ 87,225
$ 5,140,179
$ 541,543
$ 337,511
$ 61,772
$ 2,192,279
$ 61,772
$ 187,445
$ 135,513
$ 2,042,213
$ 135,513
$ 371,828
$ 59,412
$
$
$ 2,188,039
$ 179,932
$ 4,042,807
$ 179,932
$ 6,518,099
$ 253,317
$ 29,028,479
$ 3,873,384
Systems upgrading from less than 4-log to 4-log or greater
Add storage
capital cost
Increase dose - hypochlorination
annual O&M cost
Increase dose - chlorine gas
annual O&M cost
$ 1,349
$ 72
NA
$ 3,486
$ 179
NA
$ 8,093
$ 179
NA
$ 16,171
$ 195
NA
$ 36,191
$ 470
NA
$ 94,102
NA
$ 1,342
$ 201,269
NA
$ 2,108
$ 444,947
NA
$ 3,846
$ 1,854,768
NA
$ 17,838
Source: Technology and Cost Document for the Final Ground Water Rule (USEPA, 2005d)
Economic Analysis for the
Final Ground Water Rule
6-43
October 2006
-------
Annualized cost estimates for systems and States to perform source water contamination
corrective actions are presented in Exhibit 6.23.
Exhibit 6.23 PWS and State Cost Estimates for Triggered Monitoring Corrective
Action Activities ($Millions, 2003$)
Annualized Costs for Triggered Monitoring Corrective Action Activities
3%
7%
Systems
Mean
$25.64
$27.20
Lower
Bound
(5th %ile)
$14.90
$15.89
Upper
Bound
(95th
%ile)
$38.39
$40.96
States
Mean
$0.46
$0.52
Lower
Bound
(5th %ile)
$0.32
$0.36
Upper
Bound
(95th
%ile)
$0.61
$0.69
Total
Mean
$26.10
$27.72
Lower
Bound
(5th %ile)
$15.22
$16.26
Upper
Bound
(95th
%ile)
$39.00
$41.65
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
6.4.5 Compliance Monitoring
PWSs that provide 4-log treatment of viruses (using inactivation, removal, or State-approved
combination of these technologies) before or at the first customer to meet requirements of the GWR must
monitor the effectiveness and reliability of their treatment. This section presents the assumptions used to
estimate the burden and costs attributable to compliance monitoring. Exhibit 6.24 provides a schematic
of the compliance monitoring process.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-44
-------
Exhibit 6.24 Schematic of Compliance Monitoring Process
(Numbers based on 3,301 -10,000 population category for CWSs)
4,164 EP = EP achieving
4-log prior to GWR+ EP
corrected to 4-log
Yes
Did the system provide 4-
log treatment of viruses
prior to the GWR?
Yes
Does the system provide 4-log treatment of
viruses?
Total number of systems in size category
(2,884 systems = (approx.) 9,281 EP)
No
No
617 EP = EP corrected to 4-log
(previously partially disinfected,
353 EP) + (previously
nondisinfected, 264 EP)
See section 6.4.3
on Source Water
Monitoring
264 EP
System must continue
to perform monitoring
and notify State of
level of treatment
provided
No additional costs
of compliance
monitoring
With additional
costs of
compliance
monitoring
No additional costs
of compliance
monitoring
Economic Analysis for the
Final Ground Water Rule
October 2006
6-45
-------
PWSs
Different cost assumptions were needed for systems engaged in compliance monitoring of
disinfection treatment in place prior to promulgation of the GWR versus systems having technology
installed as a result of corrective action required by the GWR. The numbers of entry points subject to
each situation are presented in Exhibit 6.25.
Systems Disinfecting Prior to GWR: EPA assumed that systems already employing chemical
disinfection treatment for purposes other than compliance with the GWR, or prior to the Rule's
promulgation, would already have the monitoring program and monitoring equipment in place to monitor
the disinfection. These systems include systems currently disinfecting to 4-log as well as systems
currently disinfecting to less than 4-log. Therefore, systems which are already disinfecting would incur
no costs for adding disinfection monitoring equipment. EPA bases this assumption on the language
obtained from the pre-existing State disinfection requirements. Most States specify in the State
regulations that systems which disinfect must maintain the prescribed level of disinfection in no less than
5% of monthly samples (suggests at least 20 samples per month), or the system's disinfectant level may
not be below the prescribed disinfectant level for more than four hours. Furthermore, some States specify
that disinfecting systems must monitor the disinfection at a specified frequency, such as daily or
continuously. This type of language suggests monitoring at a frequency consistent with that specified in
the GWR, resulting in no monitoring equipment purchases for these systems.
For systems using nanofiltration technology, the monitoring capability is built into the
technology's core process. Therefore, EPA assumes that systems using nanofiltration technology prior to
implementation of the GWR will incur no treatment monitoring costs to comply with the Rule by using
nanofiltration.
Before beginning compliance monitoring, however, systems must inform the State that they
achieve 4-log treatment of viruses (using inactivation, removal, or State-approved combination of these
technologies) before or at the first customer. For costing purposes, EPA assumes that preparing and
submitting notification for the State requires 0.5 hours. Systems must also notify the State each time the
treatment technology used fails to maintain 4-log treatment for more than 4 hours (see Chapter 3 for more
detail on requirements for each treatment technology). For costing purposes, EPA assumes a disinfection
failure rate of 5 percent of entry points engaged in 4-log source water virus treatment. EPA assumes that
preparing and submitting a treatment failure report to the State requires 2.5 hours. EPA has developed
and systems will have access to automated forms that will minimize the burden to systems in complying
with these reporting requirements. Exhibit 6.26 presents the unit burdens and costs for these reporting
requirements.
Economic Analysis for the October 2006
Final Ground Water Rule 6-46
-------
Exhibit 6.25 Assumptions for Entry Points Subject to Compliance Monitoring
System Size
(Population
Served)
Systems
Subject to
Compliance
Monitoring
A
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
Totals
3,998
6,440
2,127
2,509
1,294
803
72
55
3
17,300
Entry Points
Subject to
Compliance
Monitoring
B=C+D+E
5,246
10,499
4,155
6,090
4,164
4,511
810
681
34
36,189
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
Totals
1,537
1,141
320
150
17
2
0
0
-
3,167
1,537
1,141
320
150
17
2
0
0
-
3,167
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
Totals
Grand Total
8,075
2,368
243
86
13
4
0
0
-
10,790
31,257
8,075
2,368
243
86
13
4
0
0
-
10,790
50,146
Entry Points
Achieving
4-log Prior
toGWR
C
3,996
8,873
3,547
5,378
3,547
3,856
583
545
34
30,359
850
608
170
64
7
1
0
0
-
1,700
1,160
342
35
11
1
0
0
0
-
1,549
33,608
Entry Points
Corrected
to 4-log;
partially
disinfected
D
358
917
360
396
353
548
93
94
-
3,118
150
119
33
19
2
0
0
0
-
323
1,143
337
35
12
2
1
0
0
-
1,530
4,971
Entry Points
Corrected
to 4-log;
previously
nondisinfected
E
891
709
248
317
264
107
133
42
-
2,712
537
415
117
67
8
1
0
0
-
1,144
5,772
1,689
174
63
10
3
0
0
-
7,711
11,567
Notes: Detail may not add to totals due to independent rounding.
Source: (C) Number of entry points from Exhibit 4.3.
(D) Exhibit 6.5b, Column E
(E) Exhibit 6.5b, Column, F
Economic Analysis for the
Final Ground Water Rule
6-47
October 2006
-------
Systems Treating As a Corrective Action: Systems that adopt treatment as a corrective action for
source water fecal contamination must also install monitoring equipment to perform compliance
monitoring. The unit capital and O&M costs8 are presented in Exhibit 6.27a (for systems serving 3,300
and fewer people) and Exhibit 6.27b (for systems serving more than 3,300 people). These system size
categories are costed differently due to different compliance monitoring requirements (see Chapter 1 for a
full description of compliance monitoring requirements).
EPA assumes that all systems serving 3,300 or fewer people will conduct daily grab samples for
chlorine residual measurement and incur 0.5 hours labor burden per day. Because nanofiltration
technologies have monitoring capability incorporated into their core processes, systems that adopt these
treatment techniques are assumed to incur no capital or O&M costs for compliance monitoring.
As noted above, systems must also notify the State each time the treatment technology used fails
to maintain 4-log treatment of virus for more than 4 hours. EPA assumes that preparing and submitting a
treatment failure report to the State requires 2.5 hours. This cost is the same as reported for systems
disinfecting prior to the GWR as listed in Exhibit 6.26.
8 Systems may choose to install monitoring systems that are more complicated and costly (i.e., SCADA
systems) than those presented here. However, this level of monitoring is not required under the rule and therefore is
not included as part of the cost analysis.
Economic Analysis for the October 2006
Final Ground Water Rule 6-48
-------
Exhibit 6.26 PWS Unit Costs for Compliance Monitoring for Initial State
Notification and Disinfection Failure Reports
System Size
(Population Served)
Initial State Notification for Systems Treating
Prior to GWR
Notification
Preparation
(hours)
A
Labor Cost
(per hour)
B
Unit
Report
Cost
C=A*B
Treatment Failure Report for All Systems
Treating
Report
Preparation
(hours)
D
Labor Cost
(per hour)
E
Unit
Report
Cost
F=D*E
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 10.72
$ 11.55
$ 12.37
$ 12.37
$ 15.26
$ 15.54
$ 15.54
$ 17.62
$ 17.62
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 53.60
$ 57.73
$ 61.85
$ 61.85
$ 76.28
$ 77.70
$ 77.70
$ 88.12
$ 88.12
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 10.72
$ 11.55
$ 12.37
$ 12.37
$ 15.26
$ 15.54
$ 15.54
$ 17.62
NA
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 53.60
$ 57.73
$ 61.85
$ 61.85
$ 76.28
$ 77.70
$ 77.70
$ 88.12
NA
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3,301-1 OK
1 0,001 -50K
50,001 -100K
100,001-1 Million
> 1 Million
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 10.72
$ 11.55
$ 12.37
$ 12.37
$ 15.26
$ 15.54
$ 15.54
$ 17.62
NA
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
NA
$ 21.44
$ 23.09
$ 24.74
$ 24.74
$ 30.51
$ 31.08
$ 31.08
$ 35.25
$ 35.25
$ 53.60
$ 57.73
$ 61.85
$ 61.85
$ 76.28
$ 77.70
$ 77.70
$ 88.12
NA
Notes: Detail may not add to totals due to independent rounding.
NA Not applicable (no NCWSs of this size category).
Sources: (A, D) Labor hours for initial notification and treatment failure report reflect EPA estimate.
(B, E) Labor rate from Exhibit 6.1.
Economic Analysis for the
Final Ground Water Rule
6-49
October 2006
-------
Exhibit 6.27a PWS Compliance Monitoring Unit Costs for Systems
Serving 3,300 or Fewer People
Component
Unit
Cost
(1998)
A
PPI
(1998)
B
PPI
(2003)
C
Unit
Cost
(2003)
D=A*(C/B)
Labor
burden
(per day)
E
Annual
Cost
Frequency
F
Annual
Labor
Burden
G=E*F
Annual
Total
Cost
H=D*E*F
Compliance Monitoring Labor
25-100
101-500
500-3.3k
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
$ 21.44
$ 23.09
$ 24.74
0.50
0.50
0.50
365
365
365
183
183
183
$ 3,913
$ 4,214
$ 4,515
Chlorine Test Kits
25-100
101-500
500-3.3k
$ 34.00
$ 34.00
$ 34.00
143.7
143.7
143.7
150.1
150.1
150.1
$ 35.50
$ 35.50
$ 35.50
N/A
N/A
N/A
3.65
3.65
3.65
N/A
N/A
N/A
$ 130
$ 130
$ 130
Totals
25-100
101-500
500-3.3k
183
183
183
$ 4,042
$ 4,344
$ 4,645
Notes: Detail may not add to totals due to independent rounding.
Sources: (A) Unit cost for test kit from Products for Analysis, 1998 Hach Co. Model 2231-02. Unit cost derivation for system operator
in section 6.2.1.
(B & C) Producer Price Index (PPI) Commodity Code 3500 (Finished goods less food and energy) from BLS (www.bls.gov).
(D) Labor rate from section 6.2.1 .
(E) Labor hours for compliance monitoring reflect EPA estimate.
(F) Monitoring performed daily. New test kit needed every 100 days.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-50
-------
Exhibit 6.27b PWS Compliance Monitoring Capital Unit and O&M Costs
for Systems Serving More than 3,300 People
Component
Unit
Cost
(1998)
A
PPI
(1998)
B
PPI
(2003)
C
Unit
Cost
(2003)
D=A*(C/B)
Quantity
Purchased
E
Total
Cost
F=D*E
Capital Costs
Chlorine analyzer (Hach CL17)
Power cord
Chart recorder (Honeywell 10" round)
Installation labor (System Operator)
3.3k-10k
10k-100k
>100k
Total Capital Cost
3.3k-10k
10k-100k
>100k
$ 2,375
$ 10
$ 665
N/A
N/A
N/A
121.0
121.0
121.0
N/A
N/A
N/A
114.7
114.7
114.7
N/A
N/A
N/A
$ 2,251
$ 9
$ 630
$ 25.34
$ 26.05
$ 31 .26
1
1
1
8
8
8
$ 2,251
$ 9
$ 630
$ 203
$ 208
$ 250
$ 3,094
$ 3,100
$ 3,141
Annual Operation and Maintenance
Compliance monitoring
3.3k-10k
10k-100k
>100k
Maintenance kit
Monthly reagents
Charts
Recorder pens
N/A
N/A
N/A
$ 140
$ 18
$ 15
$ 52
N/A
N/A
N/A
143.7
143.7
143.7
143.7
N/A
N/A
N/A
150.5
150.5
150.5
150.5
$ 25.34
$ 26.05
$ 31.26
$ 147
$ 19
$ 16
$ 54
80
80
80
1
12
1
1
Total Annual Operation and Maintenance Costs
3.3k-10k
10k-100k
>100k
$ 2,027
$ 2,084
$ 2,501
$ 147
$ 226
$ 16
$ 54
$ 2,470
$ 2,527
$ 2,944
Notes: Detail may not add to totals due to indpendent rounding.
Sources: (A) Unit costs for equipment (both capital and O&M) from Products for Analysis, 1998 Hach Co.
(B & C) Producer Price Index Commodity Code 1 1 7 (Electrical machinery and equipment) for Capital Costs and Commodity Code
3500 (Finished goods less food and energy) for Annual O&M, BLS (www.bls.gov). December values.
(D) Labor rate from Exhibit 6.1 .
(E) 80 hours per year for O&M compliance monitoring.
Economic Analysis for the
Final Ground Water Rule
October 2006
6-51
-------
States
States incur costs to document the initial notification from systems that achieve 4-log treatment of
viruses (using inactivation, removal, or State-approved combination of these technologies) before or at
the first customer, as well as to review the systems' disinfection failure reports, presented in Exhibit 6.28.
For costing purposes, EPA assumes that documenting the PWSs' notifications requires 0.5 hours and that
reviewing disinfection failure reports requires 3.5 hours. The burden estimate for States to review a
disinfection failure reports is greater than the burden estimated for systems to prepare and submit the
report because it is anticipated that the State will have less familiarity with any particular system and will
be required to look up additional historical information to make any assessments/determinations regarding
the report. EPA has developed and States will have access to automated forms that will minimize the
burden to systems in complying with this reporting requirement.
Exhibit 6.28 State Unit Costs for Compliance Monitoring
Cost Component
Document Initial Notification
Review Disinfection Failure Report
Labor
Hours
A
0.5
3.5
Labor
Cost
(per hour)
B
$ 33.60
$ 33.60
Unit
Cost
C=A*B
$ 16.80
$ 117.61
Notes: Detail may not add to totals due to independent rounding.
Sources: (A) Labor hours for rule activities reflect EPA estimate.
(B) Labor rate from Exhibit 6.2.
Annualized costs for systems and States to perform compliance monitoring is presented in
Exhibit 6.29.
Exhibit 6.29 PWS and State Cost Estimates for Compliance Monitoring Activities
(SMillions, 2003$)
Annualized Costs for Compliance Monitoring Activities
3%
7%
Systems
Mean
$9.35
$8.32
Lower
Bound
(5th %ile)
$3.02
$2.65
Upper
Bound
(95th
%ile)
$16.97
$15.17
States
Mean
$0.00
$0.00
Lower
Bound
(5th %ile)
$0.00
$0.00
Upper
Bound
(95th
%ile)
$0.01
$0.01
Total
Mean
$9.36
$8.32
Lower
Bound
(5th %ile)
$3.02
$2.65
Upper
Bound
(95th
%ile)
$16.98
$15.18
Notes: Detail may not add to totals due to independent rounding and independent cost model runs.
Source: Cost Model Outputs
Economic Analysis for the
Final Ground Water Rule
October 2006
6-52
-------
6.4.6 Total Capital and One-Time Costs
Exhibit 6.30 presents total capital and one-time costs of the GWR. Note, the bulk of capital costs
are incurred by transient noncommunity water systems (TNCWSs) serving 10,000 or fewer people. One-
time costs for PWSs and States are limited to rule implementation.
Exhibit 6.30 Total Initial Capital and One-Time Costs
(SMillions, 2003$)
CWS Total Initial Capital
NTNCWS Total Initial Capital
TNCWS Total Initial Capital
Total Initial PWS Capital Costs
CWS Implementation Costs
NTNCWS Implementation Costs
TNCWS Implementation Costs
Total One-Time PWS Costs
State Start-Up Cost
Total State One-Time Costs
PWSs Serving <10,000
Mean
Value
$ 79
$ 31
$ 173
$ 283
$ 5
$ 2
$ 9
$ 16
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 29
$ 12
$ 64
$ 105
$ 5
$ 2
$ 9
$ 16
Upper
(95th %ile)
$ 158
$ 60
$ 339
$ 556
$ 5
$ 2
$ 9
$ 16
PWSs Serving > 10,000
Mean
Value
$ 62
$ 0
$ 1
$ 64
$ 0
$ 0
$ 0
$ 0
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 24
$ 0
$ 0
$ 24
$ 0
$ 0
$ 0
$ 0
Upper
(95th %ile)
$ 117
$ 1
$ 2
$ 120
$ 0
$ 0
$ 0
$ 0
Total
Mean
Value
$ 141
$ 31
$ 174
$ 346
$ 5
$ 2
$ 9
$ 17
$ 13
$ 13
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 53
$ 12
$ 64
$ 129
$ 5
$ 2
$ 9
$ 17
$ 13
$ 13
Upper
(95th %ile)
$ 275
$ 61
$ 341
$ 676
$ 5
$ 2
$ 9
$ 17
$ 13
$ 13
Notes: Detail may not add to totals due to independent rounding.
The mean and confidence bounds are equal for both systems and state implementation costs because EPA derived these costs from point estimates.
Source: Appendix D, Exhibits D. 1, D.2, and D.3.
6.4.7 Uncertainty in Unit Costs
As stated in section 6.2.6, EPA recognizes that there are both variability and uncertainty in unit
cost estimates for treatment. Variability is expected in the actual costs that will be experienced by
different water systems with similar flows installing the same treatment technology. Otherwise similar
systems may experience different capital and/or O&M costs due to site-specific factors. Inputs to unit
costs such as water quality conditions, labor rates, and land costs can be highly variable and increase the
system-to-system variability in unit costs. In developing the unit cost estimates, there is insufficient
information to fully characterize what the distribution of this variability will be on a national scale for all
of the treatments and all possible conditions.
The unit costs for this EA are developed as average or representative estimates of what these unit
costs will be nationally. That is, in developing unit costs, design criteria for the technologies were
selected to represent typical, or average, conditions for the universe of systems. As a result, there is
uncertainty inherent in these unit cost estimates as they are based on independent assumptions with
supporting data and vendor quotes, where available, rather than on a detailed aggregation of State,
regional, or local estimates based on actual field conditions. In this EA, uncertainty in these national
Economic Analysis for the
Final Ground Water Rule
October 2006
6-53
-------
average unit cost factors for specific technologies is characterized as a triangular distribution with
minimum and maximum values set at the following percentages relative to the best estimate:
Capital costs: +/- 30 percent
O&M costs: +/-15 percent
These percentages were developed by EPA based on input from engineering professionals and
reflect recommendations from the National Drinking Water Advisory Council (NDWAC) (2001) in their
review of the national cost estimation methodology for the Arsenic Rule. EPA believes that the
uncertainties in capital and O&M costs for a given treatment technology are independent of one another
and that uncertainties across all technologies are independent.
The uncertainty in unit costs is reflected in the 90 percent confidence bounds shown in the
national cost summary exhibit in section 6.4.6 and later in this chapter.
6.4.8 Alternative Cost Analysis
The Agency performed an alternative cost analysis by running the cost model using a subset of
the studies that were the basis for the fecal indicator and viral occurrence rates in the primary analysis.
This subset comprised just those studies which had undergone peer review prior to publication9 and so
consisted only of peer-reviewed data. This cost model run parallels the alternative benefits model run
described in section 5.5.4. Compared to the primary analysis, this alternative cost model run results in an
increase in the annualized costs by a range of $29.1 million to $29.4 million (an approximately 47%
increase), using 3 percent and 7 percent discount rates, respectively.
6.5 Household Costs
EPA assumes that systems may pass some or all costs of a new regulation on to their consumers
in the form of rate increases. Household costs, which are in units of $ per household per year, are
estimated in this section to provide a measure of the increase in water bills that is expected to result from
the GWR. These cost increases incorporate the costs of rule implementation, sanitary surveys, triggered
monitoring, corrective actions, and compliance monitoring. Exhibit 6.31 presents the mean expected
increases in annual household costs for all CWSs, including those systems that do not have to take
corrective action for significant deficiencies or source water contamination. Exhibit 6.31 also presents the
same information for CWSs that must take corrective action. Household costs tend to decrease as system
size increases, due mainly to the economies of scale for the corrective actions.
9 Studies omitted from the alternative occurrence model are those used for the primary analysis (detailed in
Ch. 4 of the EA) that were either not published or not peer reviewed prior to publication: Missouri Alluvial Aquifer
(Vaughn, 1996), Wisconsin Migrant Worker Camp (USEPAetal., 1998), EPA Vulnerability (USEPA, 1998),
New England (Doherty et al., 1998), Three-State Study #3: Minnesota (Battigelli, 1999), Three-State Study #1:
Wisconsin (Battigelli, 1999), and the Montana Study (Miller and Meek, 1996).
Economic Analysis for the October 2006
Final Ground Water Rule 6-54
-------
To calculate household costs, the CWS population subject to each rule requirement is divided by
2.59 people per household (U.S. Bureau of the Census, 2001a) to calculate a number of households
subject to each requirement. The cost of the rule, by size category, is then divided by that number of
households to determine a per-household cost. To annualize capital costs when determining the costs to
households, EPA uses different discount rates for private and public systems and for systems of different
sizes. The rate differences between systems represent the different borrowing sources each type of
system has available to it, differences in risk, and expectations regarding inflation. The rates vary from
5.20 to 6.27 percent depending on system size and ownership, and are summarized in Exhibit 6.4.
As shown in Exhibit 6.31, annual household costs for all CWSs (including both those that do and
those that do not add treatment) range from $0.21 to $16.54, depending on system size. Household costs
for the subset of systems that undertake corrective actions range from $0.45 to $52.38, depending on
system size. EPA estimates that, as a whole, households subject to the GWR face minimal increases in
their annual costs. Approximately 66 percent of the households potentially subject to the rule are
customers of systems serving at least 10,000 people; these systems experience the lowest increases in
costs due to significant economies of scale. Households served by small systems that undertake
corrective actions will face the greatest increases in annual costs. Only CWSs are included in this
analysis because they are the only systems that serve households directly.
Exhibit 6.31 Summary of Annual Per-Household Costs for the GWR (2003$/Year)
Systems Size
(Population
Served)
Households
Mean
Median
90th
Percent! le
All Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
>1 00,000
Total
289,222
1 ,303,890
1,278,081
4,196,105
6,271,380
11,468,813
4,204,584
9,755,817
38,767,890
$ 16.54
$ 3.51
$ 0.97
$ 0.37
$ 0.27
$ 0.21
$ 0.34
$ 0.21
$ 0.51
$ 2.81
$ 0.64
$ 0.16
$ 0.04
$ 0.03
$ 0.04
$ 0.10
$ 0.04
$ 0.09
$ 9.31
$ 6.11
$ 1.70
$ 0.61
$ 0.43
$ 0.49
$ 1.02
$ 0.62
$ 0.88
Corrective Action Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50,001 -100K
>1 00,000
Total
70,563
312,484
302,557
919,133
1,487,159
2,871,250
1,215,544
2,283,144
9,461 ,833
$ 52.38
$ 12.00
$ 3.23
$ 1.33
$ 0.80
$ 0.45
$ 0.53
$ 0.68
$ 1.51
$ 18.99
$ 4.52
$ 1.33
$ 0.47
$ 0.25
$ 0.18
$ 0.26
$ 0.39
$ 0.60
$ 82.21
$ 25.76
$ 6.56
$ 2.59
$ 2.18
$ 1.18
$ 1.36
$ 1.65
$ 3.20
Source: GWR model output.
Economic Analysis for the
Final Ground Water Rule
6-55
October 2006
-------
6.6 Nonquantified Costs
Although EPA has quantified the significant costs of the GWR, there are some costs that the
Agency did not quantify. Overall, EPA believes that these nonqualified costs are much smaller than the
nonqualified benefits. These nonqualified costs result from uncertainties surrounding rule assumptions
and from modeling assumptions. For example, EPA estimated that some systems may need to acquire
land if they need to build a treatment facility or drill a new well. This was not considered for most
systems because EPA expects that the majority of the technologies that systems will use to comply with
this rule will fit within the existing plant footprint. In addition, if the cost of land is prohibitive, a system
may choose another lower cost alternative such as connecting to another source. EPA has also not
quantified costs for systems already using disinfection to conduct compliance monitoring because EPA
believes such systems are already incurring these costs.
EPA did not include the costs for taking five additional samples following a positive source water
sample. However, EPA overestimated the cost of triggered monitoring because it assumed all systems
would take an additional sample beyond the current TCR requirements. However, many small systems
(and most ground water systems are small) will be able to use one of their TCR samples to also comply
with the GWR. Overall, the impact of not including five additional sample cost (approximately $200,000
per year) is much smaller compared to the overestimate of a few million dollars associated with the initial
fecal indicator sampling cost already conducted for TCR monitoring.
EPA did not include compliance monitoring costs for systems improving disinfection from less
than 4-log to greater than 4-log inactivation because EPA believes that essentially all of these systems are
already measuring their residuals and recording such information on a daily basis as standard operating
procedures. However, there may be some systems that are not doing this, and if so, additional costs
would need to be incurred.
For some systems, further investigation into problems identified through implementation of GWR
requirements could lead to a determination that the source is more appropriately classified as GWUDI. In
such instances, systems will need to work closely with the State to clarify the source classification. If a
source is reclassified as GWUDI, the system may incur significant additional costs to comply with the
requirements (e.g., filtration) under the various regulations applicable to GWUDI/surface water systems.
Although a reclassification may be prompted by GWR requirements, the actual costs for complying with
GWUDI requirements fall under the applicable regulations governing GWUDI supplies and are not
included in this EA.
The optional assessment monitoring provision was not included in the quantitative cost analysis.
However, EPA was not able to quantify either the benefits or costs of this program.
Due to lack of information, EPA was unable to quantify the costs (as well as benefits) from the
correction of sanitary survey deficiencies in distribution systems and treatment plants.
Economic Analysis for the October 2006
Final Ground Water Rule 6-56
-------
6.7 Uncertainty Analysis
Many uncertain values are used to derive estimates of costs of this rule. Most, but not all, of
these are mathematically modeled so that a "realization" is selected for them in each "uncertainty
iteration" of this EA. These uncertainties then propagate through the derivation of final estimates so the
total uncertainty of those final estimates can be understood. The paragraphs that follow discuss the most
important of these uncertain quantities.
The Baseline Numbers of Ground Water Systems, Populations Served, and Associated Disinfection
Practice
The baseline number of systems is uncertain because of data limitations in the Safe Drinking
Water Information System (SDWIS). For example, some systems use both ground and surface water, but
because of other regulatory requirements, they are labeled in SDWIS as surface water systems. In
addition, the SDWIS data on non community water systems do not reflect a consistent reporting
convention for population served. Some States may report the population served by TNCWSs over the
course of a year, while others may report the population served on an average day. For example, a State
park may report the population served yearly instead of daily. Thus, SDWIS data may, in some cases,
overestimate the daily population served. Also, SDWIS does not require States to provide information on
current disinfection practices, resulting in uncertainty in the percentage of disinfecting systems providing
4-log or greater virus treatment. Although these different factors influencing the baseline estimates are
uncertain, EPA believes their relative degree of uncertainty in influencing the estimates within the GWR
EA is small compared to other uncertain components of the EA, so these are not treated probabilistically
in the analysis.
The Baseline Occurrence of Viruses and E. coll In Ground Water Wells
EPA's occurrence analysis is based on monitoring data from over 1,200 public drinking water
supply wells that were tested for culturable viruses, E. coll, or both. Compiled from 15 ground water
surveys that were designed for different purposes, these wells were used to represent the universe of
ground water wells. Although the number of wells is large, the number of assays per well is small, and
most wells were sampled only once for either viruses or E. coll. Because of the limited amount of data,
these data do not provide precise occurrence estimates. EPA's analysis recognizes the limitations of the
data, producing a large "uncertainty sample" of estimates that are consistent with the data. This
uncertainty sample is an input to the probabilistic economic analysis, where these uncertainties are
combined with the uncertainties of other inputs to portray total uncertainty in the GWR cost and benefit
estimates. EPA's occurrence model includes concentration differences between more and less vulnerable
wells, but applies the same hit rate model to both types of wells. Also, because of data limitations, EPA
was unable to make an assessment of aquifer sensitivity as part of the final rule and, therefore, no
difference in hit rates or concentration levels between sensitive and nonsensitive wells is assumed.
For the Sanitary Survey Provisions, the Percent of Systems Identified as Having Significant Deficiencies,
the Percent of These Deficiencies That Are Corrected, and Associated Costs and Benefits
For the sanitary survey provisions, EPA estimated the impacts associated with well deficiencies.
EPA used data from the 1998 ASDWA survey to estimate the percent of wells with deficiencies
(ASDWA, 1997). To estimate benefits, EPA assumed that if a correction of a well defect occurred at a
Economic Analysis for the October 2006
Final Ground Water Rule 6-57
-------
virally contaminated well, some, but not all of these virally contaminated wells would no longer have
viral contamination. EPA used an uncertainty distribution for this estimate.
To estimate costs for significant deficiencies detected at or near the source, EPA chose two
representative corrective actions to use in the cost model: replacement of a sanitary well seal or
rehabilitation of an existing well. Because the corrections of significant deficiencies are dependent upon
the deficiencies defined as significant by States and the conditions of specific systems, both of which are
highly variable, EPA used a high and low scenario to bound the cost estimates. The low-cost scenario
assumes a greater percentage of the systems with significant deficiencies will have deficiencies that are
less expensive to correct (e.g., more systems will have to replace their sanitary well seal than will have to
perform a complete rehabilitation of their well). This high/low bounding provides an estimate of the
uncertainty with respect to the percentages of each type of defect to be corrected.
While the sanitary survey provisions will also result in identification and correction for
deficiencies associated with treatment or distribution system deficiencies, due to insufficient data, EPA
did not quantify either costs or benefits for these types of deficiencies.
The Predicted Rates at Which Virally Contaminated (and Non-Contaminated) Wells Will Be Required to
Take Action After Finding E. coll Ground Water Sources
EPA's occurrence model estimates the percentage of wells that have only virus present, both E.
coll and virus present, or only E. coll present. The occurrence model also includes parameters that
describe how often contaminated wells actually have the contaminant present. For example, some
contaminated wells have E. coll present less than one percent of the time, while others have E. coll
present more than 10 percent of the time (some of which will also have sometime viral presence). When
E. coll contaminated wells are tested for the first time, those with frequent E. coll occurrence are the most
likely to be identified as contaminated. As these problems are addressed and corrected, there should be
fewer and fewer wells with frequent E. coll occurrence (as well as viral occurrence since a fraction of E.
coll wells will also have sometime viral presence. This diminishing rate of fecal contamination
identification is included in this EA. Uncertainty about the diminishing rate is due to uncertainty about
the EPA's estimates of how often E. coll occurs in contaminated wells. As with other key uncertain
inputs, this uncertainty is represented by an uncertainty sample of the relevant parameters. Again, EPA
assumes no difference based on vulnerability or sensitivity.
Undisinfected wells are subjected to triggered monitoring. The rate at which triggered
monitoring identifies a well as fecally-contaminated depends on both the fraction of time that E. coll is
present in the well and the frequency at which the well is sampled. Data verification (DV) data on total
coliform occurrence in distribution systems provide the basis for estimates of sampling frequency in
different types and sizes of systems. Although the data are limited, EPA has not modeled these as
uncertain estimates. Compared to other uncertain parameters, these have relatively little uncertainty and
are expected to make only minor contributions to the total uncertainty in this EA.
EPA also did not consider the cost impacts of additional sampling on corrective action costs. The
analysis assumes that for every triggered monitoring positive, at least one additional sample will also be
positive, resulting in corrective action. However, it is possible that some systems will not have a positive
additional sample and will therefore not incur costs for corrective action. Accounting for this would
reduce the costs of the rule associated with corrective actions and, to the extent that these systems actually
do have viral or bacterial pathogens present, would reduce the benefits of the rule as well.
Economic Analysis for the October 2006
Final Ground Water Rule 6-58
-------
EPA assumes that the occurrence of fecal contamination will remain constant throughout the
implementation of the rule. However, this might not be the case if increased development results in fecal
contamination of a larger number of aquifers in areas served by ground water systems or if other rules,
such as Concentrated Animal Feeding Operations (CAFO), and Class V Underground Injections Control
(UIC) Well regulations result in decreased fecal contamination. This uncertainty is not mathematically
modeled in this EA.
The Costs of Taking Action After Finding E. coli in Ground Water Sources
EPA recognizes that there are both variability and uncertainty in unit cost estimates for treatment.
Variability is expected in the actual costs that will be experienced by different water systems with similar
flows installing the same treatment technology. Otherwise similar systems may experience different
capital and/or O&M costs due to site-specific factors. Inputs to unit costs such as water quality
conditions, labor rates, and land costs can be highly variable and increase the system-to-system variability
in unit costs. In developing the unit cost estimates, there is insufficient information to fully characterize
what the distribution of this variability will be on a national scale for all of the treatments and all possible
conditions.
The unit costs for this EA are developed as average or representative estimates of what these unit
costs will be nationally. That is, in developing unit costs, design criteria for the technologies were
selected to represent typical, or average, conditions for the universe of systems. As a result, there is
uncertainty inherent in these unit cost estimates since they are based on independent assumptions with
supporting data and vendor quotes, where available, rather than on a detailed aggregation of State,
regional, or local estimates based on actual field conditions. EPA quantifies the uncertainty in these
national average unit cost factors for specific technologies. The percent uncertainty bounds used to
characterize unit costs were developed based on input from engineering professionals and reflect
recommendations from the National Drinking Water Advisory Council (NDWAC, 2001) in their review
of the national cost estimation methodology for the Arsenic Rule. EPA believes that the uncertainties in
capital and O&M costs for a given treatment technology are independent of one another and that
uncertainties across all technologies are independent.
Optional Assessment Monitoring
The Agency was not able to estimate the benefits or costs resulting from the optional assessment
monitoring program. States can determine what systems they deem most vulnerable to fecal
contamination and require these systems to conduct assessment monitoring. Systems would incur
additional costs from monitoring and reporting results as well as any corrective action associated with
fecal indicator positives. States would incur additional costs for determining what systems would be
required to monitor, assisting systems with corrective actions decisions, and recordkeeping.
Corrective Actions and Significant Deficiencies
The Agency also did not develop costs for corrective actions for all conceivable significant
deficiencies that a system may encounter. Instead, representative actions that span the range of low cost
to expensive actions were used. The corrective actions that are a result of significant deficiencies
identified during sanitary surveys do not include the ones performed within the treatment plant or in the
distribution system due to lack of adequate data. Exclusion of these costs from the cost analysis results in
an underestimate of potential rule costs, though the magnitude of the underestimate is unknown.
Economic Analysis for the October 2006
Final Ground Water Rule 6-59
-------
In addition, EPA also recognizes that some costs (as well as benefits) from the correction of
sanitary survey deficiencies identified in the distribution systems and treatment plant have not been
quantified.
Uncertainty Summary
Overall, EPA recognizes that there is uncertainty in various parts of its estimates that could result
in either an over- or underestimate of the costs as presented in this chapter. Exhibit 6.32 presents a
summary of these issues, references the section or appendix where the information is introduced, and
estimates the effects that each may have on national costs. The Agency has been careful to use the best
available data, to account for uncertainty quantitatively when possible, and to avoid any consistent biases
in assumptions and the use of data. The primary known bias is that some benefits and costs have not been
quantified, and therefore are not included in the quantitative comparison of regulatory alternatives.
Exhibit 6.32 Cost Uncertainty Summary
Uncertainty
Percentage of systems
finding TC positive samples
Number of triggered
monitoring samples
attributable to TC positives
under the GWR
Percentage of triggered
monitoring systems
receiving indicator-positive
source water
Uncertainty in compliance
forecasts for corrective
actions
Percentage of disinfection
failures
Unit Costs
Sanitary Survey corrective
actions
Compliance with multiple
rules
Section With
Full Discussion
of Uncertainty
6.4.3
6.4.3
6.4.3
6.4.4
6.4.5
6.4.7
6.6
6.7
Most Likely Effect of Current Assumptions on
Estimate of National Costs
Underestimate
X
Overestimate
X
Unknown
Impact
X
X
X
X
X
X
Economic Analysis for the
Final Ground Water Rule
October 2006
6-60
-------
6.8 Total Annualized Cost for Final GWR Regulatory Alternative
Based on information presented previously in this chapter, EPA developed national cost estimates
for the Final GWR. Exhibit 6.33 presents the total annualized costs to PWSs for the final GWR at 3 and 7
percent discount rates. Exhibit 6.34 presents the total annualized cost for the final GWR by system size
and type at 3 and 7 percent discount rates.
Exhibit 6.33 Total Annualized Present Value Costs ($Millions, 2003$)
Discount
Rate
3 percent
7 percent
Systems
Mean
Value
$ 50.0
$ 50.6
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 34.3
$ 35.2
Upper
(95th %ile)
$ 68.8
$ 69.0
States
Mean
Value
$ 11.8
$ 11.7
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 10.9
$ 10.9
Upper
(95th %ile)
$ 12.6
$ 12.6
Total
Mean
Value
$ 61.8
$ 62.3
90 Percent
Confidence Bound
Lower
(5th %ile)
$ 45.2
$ 46.1
Upper
(95th %ile)
$ 81.4
$ 81.6
Notes: Detail may not add to totals due to independent rounding.
Source: Cost Model Outputs
Economic Analysis for the
Final Ground Water Rule
October 2006
6-61
-------
Exhibit 6.34 Total Annualized Costs to Systems by System Size and Type
(SMillions, 2003$)
System Size
(Population Served)
At 3%
Mean
Value
90 Percent
Confidence Bound
Lower
(5th %ile)
Upper
(95th %ile)
At 7%
Mean
Value
90 Percent
Confidence Bound
Lower
(5th %ile)
Upper
(95th %ile)
Community Water Systems (CWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
1 00,001 -1M
>1 Million
All Sizes
$ 2.91
$ 3.61
$ 1.45
$ 1.96
$ 2.23
$ 2.85
$ 1.78
$ 1.89
$ 0.00
$ 18.67
$ 2.13
$ 2.54
$ 0.99
$ 1.29
$ 1.40
$ 1.87
$ 1.01
$ 1.14
$ 0.00
$ 12.39
$ 3.69
$ 4.94
$ 1.98
$ 2.76
$ 3.26
$ 4.06
$ 2.73
$ 2.89
$ 0.00
$ 26.31
$ 2.91
$ 3.59
$ 1.44
$ 1.94
$ 2.26
$ 3.05
$ 1.88
$ 2.12
$ 0.00
$ 19.19
$ 2.16
$ 2.60
$ 1.00
$ 1.30
$ 1.49
$ 2.05
$ 1.11
$ 1.31
$ 0.00
$ 13.03
$ 3.72
$ 4.83
$ 1.94
$ 2.69
$ 3.25
$ 4.41
$ 2.84
$ 3.18
$ 0.00
$ 26.87
Nontransient Noncommunity Water Systems (NTNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
1 00,001 -1M
>1 Million
All Sizes
$ 1.77
$ 1.86
$ 0.67
$ 0.46
$ 0.10
$ 0.04
$ 0.01
$ 0.01
$
$ 4.91
$ 1.34
$ 1.21
$ 0.43
$ 0.28
$ 0.05
$ 0.02
$ 0.00
$ 0.00
$
$ 3.33
$ 2.27
$ 2.60
$ 0.97
$ 0.68
$ 0.16
$ 0.07
$ 0.01
$ 0.01
$
$ 6.77
$ 1.78
$ 1.83
$ 0.66
$ 0.46
$ 0.10
$ 0.04
$ 0.01
$ 0.01
$
$ 4.89
$ 1.34
$ 1.23
$ 0.42
$ 0.28
$ 0.06
$ 0.02
$ 0.00
$ 0.00
$
$ 3.36
$ 2.26
$ 2.53
$ 0.94
$ 0.67
$ 0.17
$ 0.07
$ 0.01
$ 0.02
$
$ 6.68
Transient Noncommunity Water Systems (TNCWSs)
<100
101-500
501-1,000
1,001-3,300
3.301-10K
1 0,001 -50K
50.001-100K
1 00,001 -1M
>1 Million
All Sizes
TOTAL
$ 17.74
$ 7.07
$ 0.98
$ 0.42
$ 0.13
$ 0.08
$ 0.01
$ 0.01
$
$ 26.44
$ 50.02
$ 12.87
$ 4.75
$ 0.58
$ 0.25
$ 0.07
$ 0.04
$ 0.00
$ 0.01
$
$ 18.56
$ 34.28
$ 23.27
$ 9.95
$ 1.45
$ 0.63
$ 0.21
$ 0.14
$ 0.01
$ 0.02
$
$ 35.68
$ 68.76
$ 17.86
$ 7.00
$ 0.96
$ 0.42
$ 0.14
$ 0.09
$ 0.01
$ 0.01
$
$ 26.49
$ 50.57
$ 13.11
$ 4.76
$ 0.59
$ 0.26
$ 0.07
$ 0.04
$ 0.00
$ 0.01
$
$ 18.84
$ 35.22
$ 23.44
$ 9.61
$ 1.38
$ 0.63
$ 0.22
$ 0.15
$ 0.01
$ 0.02
$
$ 35.46
$ 69.00
Notes: Detail may not be consistent with summary presentations due to independent statistical analysis.
Source: Cost Model Outputs, Exhibit 6.6.
6.9 Comparison of Regulatory Alternatives
During the development of the GWR, the Agency considered several regulatory alternatives.
Exhibit 6.35 provides a comparison of the total annual cost of compliance across the four regulatory
alternatives evaluated for the GWR. The cost of the final rule lies between the least costly and most costly
alternatives. The costs increases, however, by more than a factor often from the final rule to the across-
the-board disinfection alternative. This increase in costs results from the fact that the final rule and
Alternatives 1 and 3 first are tailored to focus on PWSs that have a demonstrated risk of providing their
customers fecally contaminated drinking water. Across-the-board disinfection, as the name implies,
requires all PWSs to treat their source water, even if there is no demonstrated potential or actual fecal
Economic Analysis for the
Final Ground Water Rule
October 2006
6-62
-------
contamination. This means that costs are being incurred by many more PWSs under this alternative than
the others.
The burden associated with State oversight and administration varies among the different
regulatory alternatives. State burden and cost estimates for implementation and annual administration for
the final regulatory alternative are presented in Exhibits 6.7a and 6.7b. These costs are estimated to be
similar for Alternatives 2 and 3. States incur less burden from oversight and administration of Alternatives
1 and 4 because these alternatives do not have monitoring components of the rule. Additionally,
laboratory certification is eliminated in Alternatives 1 and 4, and due to limited numbers and complexity of
rule components, the burden for the remaining activities is reduced by 50% compared to Alternatives 2 and
3. For Alternative 4, technology selection is allocated in accordance with the compliance forecast for all
systems choosing treatment technologies.
EPA used the same process for developing costs for the final rule to develop costs for the other
alternatives. Unit costs were multiplied by the number of systems performing various components of each
alternative, and results were summed for all components.
Exhibit 6.35 Comparison of National Annual Costs
by Regulatory Alternative ($Millions, 2003$)
Rule Alternative
Alternative 1
Sanitary Survey Only
Alternative 2 - Risk-Targeted Approach (Final Rule)
Sanitary Survey and Triggered Monitoring
Alternative 3 - Multi-Barrier Approach
Sanitary Survey, Hydrogeologic Sensitivity Assessment,
Assessment Monitoring, and Triggered Monitoring
Alternative 4
Across-the-Board Disinfection
Total Annualized Cost ($Millions)
3 Percent Discount Rate
Mean
Estimate
$ 15.3
$ 61.8
$ 67.9
$ 686.4
90 Percent
Lower
(5th %tile)
$ 11.8
$ 45.2
$ 49.4
$ 636.8
Upper
(95th %tile)
$ 19.2
$ 81.4
$ 89.5
$ 735.4
7 Percent Discount Rate
Mean
Estimate
$ 15.3
$ 62.3
$ 69.4
$ 665.3
90 Percent
Lower
(5th %tile)
$ 11.9
$ 46.1
$ 51.0
$ 612.3
Upper
(95th %tile)
$ 19.0
$ 81.6
$ 90.6
$ 717.0
Source: Cost Model Outputs
Economic Analysis for the
Final Ground Water Rule
October 2006
6-63
-------
7. Economic Impact Analysis
7.1 Introduction
As part of the rulemaking process, EPA is required to address the direct and indirect burdens that
the GWR may place on certain types of governments, businesses, and populations. This chapter presents
the analyses performed by EPA in accordance with the following 12 Federal mandates.
1) The Regulatory Flexibility Act (RFA) of 1980, as amended by the Small Business Regulatory
Enforcement Fairness Act (SBREFA) of 1996.
2) An analysis of small system affordability to determine variance technologies in accordance with
Section 1415(e)(l) of the SDWA Amendments.
3) Feasible technologies available to all systems as required by Section 1412(b)(4)(E) of the SDWA
Amendments.
4) A Technical, Financial, and Managerial Capacity Assessment as required by Section 1420(d)(3)
of the 1996 Amendments to the SDWA.
5) Paperwork Reduction Act (a separate Information Collection Request document contains the
complete analysis).
6) Unfunded Mandates Reform Act (UMRA) of 1995.
7) Executive Order 13175 (Consultation and Coordination with Indian Tribal Governments).
8) Impacts on sensitive subpopulations as required by Section 1412(b)(3)(c)(i) of the Safe Drinking
Water Act (SDWA) Amendments.
9) Executive Order 13045 (Protection of Children from Environmental Health Risks and Safety
Risks).
10) Executive Order 12898 (Federal Actions to Address Environmental Justice in Minority
Populations and Low-Income Populations).
11) Executive Order 13132 (Federalism).
12) Executive Order 13211 (Actions Concerning Regulations That Significantly Affect Energy
Supply, Distribution, or Use).
Many of the requirements and executive orders listed above call for an explanation of why the
rule is necessary, the statutory authority for the rule, and the primary objectives that the rule is intended to
achieve (refer to Chapter 2 for more information regarding the objectives of the rule). More specifically,
they are designed to assess the financial and health effects of the rule on sensitive, low-income, and Tribal
populations as well as on small systems. The chapter also examines how much additional capacity
systems will need to meet GWR requirements and whether there are existing, feasible technologies and
treatment techniques available to meet rule requirements.
7.2 Regulatory Flexibility Act and Small Business Regulatory Enforcement Fairness
Act
The Regulatory Flexibility Act (RFA) generally requires an agency to prepare a regulatory
flexibility analysis for any rule subject to notice and comment rulemaking requirements under the
Administrative Procedure Act or other statute, unless the Agency certifies that the rule will not have a
significant economic impact on a substantial number of small entities (5 U.S.C. 602(a)). Small entities
include small businesses, small organizations, and small governmental jurisdictions.
Economic Analysis for the 7-1 October 2006
Final Ground Water Rule
-------
Defining Small Entities Affected by the Rule
The RFA provides default definitions for each type of small entity. Small entities are defined as:
(1) a small business as defined by the Small Business Administration's (SBA) regulations at 13 CFR
121.201; (2) a small governmental jurisdiction that is a government of city, county, town, school district,
or special district with a population of less than 50,000; and (3) a small organization that is any "not-for-
profit enterprise that is independently owned and operated and is not dominant in its field." However, the
RFA also authorizes an agency to use alternative definitions that "are appropriate to the activities of the
agency after proposing the alternative definition(s) in the Federal Register and taking comment" (5
U.S.C. 601(3)-(5)) for each category of small entity. In addition, to establish an alternative small
business definition, agencies must consult with SBA's Chief Council for Advocacy.
The RFA references the definition of "small business" found in the Small Business Act, which
authorizes the SBA to define "small business" further by regulation. The SBA defines small businesses
by category using the North American Industry Classification System (NAICS). The NAICS code for
public water supplies (PWSs) is 22131 (Water Supply and Irrigation Systems), and State agencies that
include drinking water programs are classified as 92411 (Administration of Air and Water Resource and
Solid Waste Management Programs) or 923120 (Administration of Public Health Programs). Ancillary
systems (i.e., those that supplement the function of other establishments like factories, power plants,
mobile home parks, etc.) cannot be categorized in a single NAICS code. For ancillary systems, the
NAICS code is that of the primary establishment or industry. Examples of small businesses include
small, privately-owned PWSs and for-profit businesses where provision of water may be ancillary, such
as mobile home parks or day-care centers. Examples of small organizations include churches, schools,
and homeowner associations.
The GWR applies to all PWSs that use ground water. Although the SBA and the RFA provide
clear definitions for small businesses, organizations, and governmental jurisdictions, small entities are not
necessarily small water systems. The size of the entity has no relation to the number of people it serves as
a water supply. Furthermore, data are not collected on businesses, organizations, and governmental
jurisdictions in terms of the number of water customers they serve. Therefore, EPA chose to use an
alternative definition for small entities.
For purposes of assessing the impacts of the GWR on small entities, EPA considered small
entities to be PWSs serving fewer than 10,000 people. This is the cut-off level specified by Congress in
the 1996 Amendments to the Safe Drinking Water Act for small system flexibility provisions. As
required by the RFA, EPA proposed using this alternative definition in the Federal Register (63 FR 7620;
February 13, 1998), requested public comment, consulted with the SBA, and finalized the alternative
definition for all future drinking water regulations in the Consumer Confidence Reports regulation (63 FR
44511; August 19, 1998). As stated in that final rule, the alternative definition would be applied to this
regulation as well.
Measuring Significant Impacts
EPA has not revised its determination at proposal that the GWR will have a substantial impact
upon a significant number of small water systems. EPA assessed the potential impact of today's rule on
small entities. There are 147,330 CWSs, NTNCWSs, and TNCWSs providing potable ground water to
the public, 145,580 (99 percent) are classified by EPA as small entities. EPA has determined that all
small ground water systems are impacted by the sanitary survey requirement and a substantial number
these systems will be impacted by additional requirements of today's final rule including the source water
monitoring requirements and the corrective action requirements. Exhibit 6.5b provides a detailed
summary of the numbers of systems and entry points impacted by each GWR requirement.
Economic Analysis for the 7-2 October 2006
Final Ground Water Rule
-------
As required by section 604 for the RFA, EPA also prepared a final regulatory flexibility analysis
(FRFA) for today's final rule. The FRFA addresses the issues raised by public comments on the IRFA,
which was part of the proposal of this rule. The FRFA is available for review in the docket.
Small Entity Outreach and Small Business Advocacy Review Panel
Section 609(b) of the RFA, as amended by SBREFA, and Section 203 of UMRA require EPA to
provide small governments with an opportunity for timely and meaningful participation in the regulatory
development process. In addition, the Agency must consult with small entity stakeholders and convene
an Small Business Advocacy Review (SBAR) Panel prior to publication of a proposed rule.
EPA convened an SBAR Panel for the proposed rule. The SBAR Panel members for the GWR
included the Small Business Advocacy Chair of EPA, the Director of the Standards and Risk
Management Division in the Office of Ground Water and Drinking Water (OGWDW) within EPA's
Office of Water, the Administrator for the Office of Information and Regulatory Affairs of OMB, and the
Chief Counsel for Advocacy of the SBA. The SBAR Panel convened on April 10, 1998 and met seven
times before the end of the 60-day period on June 8, 1998. The culmination of these meetings was the
SBAR Panel's report, Final Report of the SBREFA Small Business Advocacy Review Panel on EPA 's
Planned Proposed Rule for National Primary Drinking Water Regulations: Ground Water. The small
entity stakeholder comments on components of the GWR and the background information provided to the
SBAR Panel and the small entity stakeholders are available for review in the water docket. This
information and the Agency's response to the SBAR Panel's recommendations in developing the GWR
are summarized below.
Prior to convening the SBAR Panel, OGWDW consulted with a group of 22 small entity
stakeholders likely to be impacted by a GWR. The small entity stakeholders small system operators, local
government officials, small business owners (e.g., a bed and breakfast with its own water supply), and
small nonprofit organization (e.g., a church with its own water supply). The small entity stakeholders
were provided with background information on the rule, on the need for the rule, and the potential
requirements. The small entity stakeholders were asked to provide input on the potential impacts of the
rule from their perspective. All 22 small entity stakeholders commented on the information provided in
the IRFA. These comments were provided to the SBAR Panel when it convened. After a teleconference
between the small entity stakeholders and the SBAR Panel, the small entity stakeholders were invited to
provide additional comments on the information provided. Three small entity stakeholders provided
additional comments on the rule components after the teleconference.
In general, the small entity stakeholders consulted on the GWR were concerned about the impact
of the rule on small water systems (because of their small staff and limited budgets), the additional
monitoring that might be required, and the data and resources necessary to conduct a hydrogeologic
sensitivity assessment (HSA) or sanitary survey. There was also considerable discussion about whether
the source data was nationally representative. Small entity stakeholders suggested providing flexibility to
the States/Primacy Agencies implementing these provisions and opposed mandatory disinfection across-
the-board. Small entity stakeholders expressed support for existing monitoring requirements as a means
of determining compliance, and some supported increased requirements for total coliform monitoring.
Consistent with the RFA/SBREFA requirements, the SBAR Panel evaluated the assembled
materials and small entity comments related to the elements of the IRFA. A copy of the SBAR Panel
report is available in the Office of Water docket for the GWR. The SBAR Panel suggested that, given the
number of systems that could be affected by the rule, EPA consider focusing compliance requirements on
those systems most at risk of fecal contamination. From this perspective, the SBAR Panel suggested that
EPA evaluate whether it would be appropriate to establish different rule requirements for systems based
Economic Analysis for the 7-3 October 2006
Final Ground Water Rule
-------
on system type, size, or location. The SBAR Panel also suggested providing States/Primacy Agencies
with maximum flexibility, consistent with ensuring an appropriate minimum level of public health
protection, to tailor specific requirements to individual system needs and resources. The SBAR Panel's
recommendations to address the small entity stakeholders' concerns about the GWR were considered in
developing the regulatory alternatives analyzed in this final rulemaking.
7.3 Small Drinking Water System Variances
Section 1415(e) of SDWA allows States/Primacy Agencies to grant variances to small water
systems that cannot afford to comply with a National Primary Drinking Water Regulation.
Section 1415(e)(6)(B) of SDWA, however, states that a variance shall not be available for a "national
primary drinking water regulation for a microbial contaminant (including a bacterium, virus, or other
organism) or an indicator or treatment technique for a microbial contaminant." Therefore, this provision
does not apply, because the GWR is a regulation to control a microbial contaminant.
7.4 Feasible Treatment Technologies for All Systems
In accordance with Section 1412(b)(4)(E) of the 1996 SDWA Amendments, EPA examined
whether there were existing, feasible technologies and treatment techniques available that would allow
systems to meet the GWR requirements. EPA determined that ground water systems of all sizes can meet
the requirements of 4-log virus inactivation by using chlorine (hypochlorination or gas disinfection),
which is also relatively inexpensive and simple for systems to install and operate.
7.5 Effect of Compliance with the GWR on the Technical, Financial, and Managerial
Capacity of Public Water Systems
Section 1420(d)(3) of the SDWA, as amended, requires that, in promulgating aNPDWR, the
Administrator shall include an analysis of the likely effect of compliance with the regulation on the
technical, managerial, and financial (TMF) capacity of PWSs. The following analysis fulfills this
statutory obligation by identifying the incremental impact that the GWR will have on the TMF of
regulated water systems. Analyses presented in this document reflect only the impact of new or revised
requirements, as established by the GWR; the impacts of previously established requirements on system
capacity are not considered.
Overall water system capacity is defined in Guidance on Implementing the Capacity Development
Provisions of the Safe Drinking Water Act Amendments of 1996 (USEPA 1998a) as the ability to plan for,
achieve, and maintain compliance with applicable drinking water standards. Capacity encompasses three
components: technical, managerial, and financial. Technical capacity is the operational ability of a water
system to meet those SDWA requirements. Key issues of technical capacity include the following:
1) Source Water Adequacy—Does the system have a reliable source of water with adequate
quantity? Is the source generally of good quality and adequately protected?
2) Infrastructure Adequacy—Can the system provide water that meets SDWA standards?
What is the condition of its infrastructure, including wells or source water intakes,
treatment, storage, and distribution? What is the infrastructure's life expectancy? Does
the system have a capital improvement plan?
Economic Analysis for the 7-4 October 2006
Final Ground Water Rule
-------
3) Technical Knowledge and Implementation—Are the system's operators certified? Do the
operators have sufficient knowledge of applicable standards? Can the operators
effectively implement this technical knowledge? Do the operators understand the
system's technical and operational characteristics? Does the system have an effective
operation and maintenance (O&M) program?
Managerial capacity is the ability of a water system's managers to make financial, operating, and
staffing decisions that enable the system to achieve and maintain compliance with SDWA requirements.
Key issues include:
Ownership Accountability—Are the owners clearly identified? Can they be held
accountable for the system?
• Staffing and Organization—Are the operators and managers clearly identified? Is the
system properly organized and staffed? Do personnel understand the management
aspects of regulatory requirements and system operations? Do they have adequate
expertise to manage water system operations? Do personnel have the necessary licenses
and certifications?
• Effective External Linkages—Does the system interact well with customers, regulators,
and other entities? Is the system aware of available external resources, such as technical
and financial assistance?
Financial capacity is a water system's ability to acquire and manage sufficient financial resources
to allow the system to achieve and maintain compliance with SDWA requirements. Key issues include:
Revenue Sufficiency—Do revenues cover costs?
Creditworthiness—Is the system financially healthy? Does it have access to capital
through public or private sources?
• Fiscal Management and Controls—Are adequate books and records maintained? Are
appropriate budgeting, accounting, and financial planning methods used? Does the
system manage its revenues effectively?
7.5.1 Requirements of the Final GWR
This capacity analysis is presented only for the Final Rule, although EPA took similar
considerations into account in the selection of the Final Rule over the other alternatives. This process led
to the incorporation of less expensive rule features for systems having fewer capabilities. For example,
EPA allowed flexibility in the triggered monitoring requirements for certain small systems. Very small
systems that must take four repeat samples following a total coliform (TC) positive sample under the TCR
may designate one of the repeat samples as a source water sample, which covers both TCR repeat
sampling and GWR triggered monitoring requirements. In addition, the requirement for performing
assessment monitoring, which was part of the preferred regulatory alternative at proposal, is now a State
option as part of the Final Rule. Flexibility was also built into the GWR through extended compliance
time frames for NCWSs, which are primarily small systems.
Economic Analysis for the 7-5 October 2006
Final Ground Water Rule
-------
The GWR establishes four major requirements that may affect the TMF capacity of affected
PWSs:
1) Sanitary surveys
2) Triggered source water monitoring
3) Corrective actions
4) Compliance monitoring
In addition, personnel from systems regulated under the GWR will need to familiarize themselves
with the rule and its requirements.
7.5.2 Systems Subject to the GWR
The GWR will apply to all PWSs that use ground water and may affect 42,361 CWSs, 18,908
NTNCWSs, and 86,061 TNCWSs—147,330 systems in all. While most will not, some systems may
require increased TMF capacity to comply with the new requirements, or will need to tailor their
compliance approaches to match their capacities. Refer to section 7.5.4 for a detailed discussion of the
changes in TMF capacity for small and large systems.
7.5.3 Impact of the GWR on System Capacity
The estimates presented in Exhibits 7. la and 7. Ib reflect the anticipated impact of the GWR on
system capacity based on the expected measures that systems will be required to adopt. The extent of the
expected impact of a particular requirement on system capacity is estimated using a scale of 0-5, where 0
represents a requirement that is not expected to have any impact, 1 represents a requirement that is
expected to have a minimal impact, and 5 represents a requirement that is expected to have a very
significant impact on system capacity. Criteria used to develop the scores and associated impacts are
discussed further in section 7.5.4.
These impacts are assessed separately for small systems (Exhibits 7. la) and for large systems
(Exhibit 7.1b). This distinction is necessary because most large systems will face fewer challenges in
implementing the rule than smaller systems. For both large and small systems, EPA evaluated the
capacity impact of each requirement on those systems affected by that particular requirement. Because in
many cases the requirements only affect a small percentage of systems/entry points, the exhibits also
display the number of systems and percentage of systems/entry points (of the subset of small or large
systems/entry points) estimated to be affected by each specific requirement.
Economic Analysis for the 7-6 October 2006
Final Ground Water Rule
-------
Exhibit 7.1a Estimated Impact of the GWR on Small System's Technical, Managerial, and Financial Capacity
(0 = no impact, 1 = minimal impact, and 5 = very significant impact)
Requirement
Familiarization with rule
requirements
Sanitary surveys
Triggered monitoring
Corrective actions for
significant deficiencies
Compliance monitoring
Number and Percent of
Small Systems
145,580(100%)
145,580(100%)
128,711 (88%)
24,749(17%)
3,685 (3%)
Technical Capacity
>,
§ 0
01 §
=3 O>
O T3
0
0
0
4
2
0
ro
cr
73
o>
"§
CO
M—
0
0
0
4
2
o3
o>
D)
T3
"5
Ł•Ł
^ (0
— c
<0 Oi
•1 1
.i®
0 Q.
0
0
0
4
2
Managerial Capacity
!Q
•Ł
o
o
o
Q.
U)
c
0
1
0
1
4
2
o
ro
N
ro
D)
O
D)
it
OT
2
0
1
4
2
ro
L_
LJJ
^ O1
o ro
Oi -^
LJJ _l
1
0
1
4
2
Financial Capacity
[T
Oi
o
it
»
c
QJ
QJ
cc
1
0
1
4
2
u)
c
1
>
Oi
ii
O
1
0
1
4
2
o
o
O
o3
D)
^
(0
0
LL
1
0
1
4
2
Notes: (1) To analyze the impact of these requirements on system capacity, the requirements believed to have the most and the least impact on
affected systems were analyzed first. These initial analyses were then used as the bases against which the relative impact of the remaining
requirements were assessed. The impact estimates developed for each requirement were also compared to those developed for the Long Term
2 Enhanced Surface Water Treatment Rule (LT2ESWTR) and the Stage 2 Disinfectants and Disinfection Byproducts Rule (Stage 2 DBPR) to
ensure cross-rule consistency and to enable cross-rule comparisons.
(2) The scores presented above represent the worst case scenario; the requirements of this rule are expected to have less impact on the
capacity of most systems affected by each requirement.
Source: Number and percent of systems subject to each rule activity from Exhibit 6.5b. Impact on capacity is determined relative to previous regulations based on
the cost and number of systems/plants that require additional capacity to comply with each requirement, as described in section 7.5.4.
Economic Analysis for the
Final Ground Water Rule
7-7
October 2006
-------
Exhibit 7.1 b Estimated Impact of the GWR on Large System's Technical, Managerial, and Financial Capacity
(0 = no impact, 1 = minimal impact, and 5 = very significant impact)
Requirement
Familiarization with rule
requirements
Sanitary surveys
Triggered monitoring
Corrective actions for significant
deficiencies
Compliance monitoring
Number and
Percent of Large
Systems
1,750(100%)
1,750(100%)
964 (55%)
298(17%)
123(7%)
Technical Capacity
.2
"ro
§ o
01 §
i— O~
=3 O>
O T3
1
0
0
3
2
0
ro
cr
T3
^f
CD
-§->
CD
M—
1
0
0
3
2
o3
o>
D)
T3
-^
5 c
o o
Ł= i=
^ (0
1 1
o "o_
1
0
0
3
2
Managerial Capacity
Ł
o
o
o
Q.
0
1
0
0
3
2
o
ro
N
c
ro
D)
O
D)
it
OT
1
0
0
3
2
ro
LU
= D)
o ro
•2 "c
LU _l
1
0
0
3
2
Financial Capacity
o
o>
o
o>
CD
CD
CD
cc
1
0
0
3
2
u)
O1
_c
T3
0
O
1
0
0
3
2
o
c
o
O
I
D)
2
ro
o
LL
1
0
0
3
2
Notes: (1) To analyze the impact of these requirements on system capacity, the requirements believed to have the most and the least impact on
affected systems were analyzed first. These initial analyses were then used as the bases against which the relative impact of the remaining
requirements were assessed. The impact estimates developed for each requirement were also compared to those developed for the
LT2ESWTR and the Stage 2 DBPR to ensure cross-rule consistency and to enable cross-rule comparisons.
(2) The scores presented above represent the worst case scenario; the requirements of this rule are expected to have less impact on the
capacity of most systems affected by each requirement.
Source: Number and percent of systems subject to each rule activity from Exhibit 6.5b. Impact on capacity is determined relative to previous regulations based on
the cost and number of systems/plants that require additional capacity to comply with each requirement, as described in section 7.5.4.
Economic Analysis for the
Final Ground Water Rule
7-&
October 2006
-------
7.5.4 Derivation of GWR Scores
EPA developed a 5-point scoring system to analyze the impact compliance with all new
regulations will have on the TMF capacity of PWSs. For each regulation, it is necessary to complete the
following steps:
1. Determine the type and number of PWSs to which the regulation applies.
2. List all of the requirements of the regulation.
3. Determine the type and number of PWSs to which each requirement applies.
4. Evaluate the impact of each requirement on the capacity of affected PWSs.
The determination of the universe of affected systems and the evaluation of the capacity impact
of individual requirements requires the use of the cost and technical information contained in SDWIS,
EAs developed for other rules, information collection requests, and other supporting documentation for
the rule. These data sources are also used to develop a qualitative description of the expected response of
affected systems to each requirement.
The overall evaluation of the impact of a requirement on the affected systems, presented in
Exhibit 7.2, is based on the impact each requirement has on nine sub-categories of capacity—three sub-
categories under each of the broader divisions of TMF capacity. Within these sub-categories, a
professional engineer with extensive water system experience reviewed the costs, number of systems
affected, and complexity of each requirement. After estimating the technical, managerial, and financial
impacts within each sub-category, the professional engineer assigned the scores using best professional
judgment. Costs were considered cumulatively for each requirement for small and large systems. This
score reflects the additional capacity that systems will need to develop to comply with each requirement.
Due to a lack of available information on operating budgets, this analysis does not include a quantitative
component.
To ensure the ability to make cross-rule comparisons, to standardize the assignment of numerical
scores, and to minimize the subjectivity of the scoring system, the requirements made on systems by the
regulation in question are compared to the requirements of those regulations for which capacity impact
analyses have already been conducted (e.g., Long Term 1 Enhanced Surface Water Treatment Rule
(LT1ESWTR), LT2ESWTR, Stage 2 DBPR). Similar requirements are assigned similar impact scores.
These group assignments are reviewed by the EPA Rule Manager and other EPA staff cognizant
of small system issues to ensure that they accurately reflect the cumulative impact of the rule
requirements on system capacity. Any disagreements over the assignments are discussed. The EPA Rule
Manager and other EPA staff discuss the rationale for the disagreement and evaluate whether the
assignments need to be adjusted. EPA adjusts the assignments only after review of the rule support
documents and an analysis of the expected system response to the rule requirements.
7.5.4.1 Small Water Systems (Those Serving 10,000 or Fewer People)
Small systems will likely face only a minimal challenge to their technical and managerial
capacity as a result of efforts to familiarize themselves with the GWR and aid the State in conducting
sanitary surveys. Total coliform sampling is already required under the TCR and, therefore, it is not
expected to pose any new technical or managerial capacity issues for systems. On average, PWSs serving
fewer than 3,000 people will only need to take less than 1 triggered samples a year, and small systems
Economic Analysis for the 7-9 October 2006
Final Ground Water Rule
-------
serving fewer than 1,000 people will be able to use one of their required TCR repeat samples to satisfy
triggered monitoring requirements, eliminating any extra burden or cost.
Small system technical and managerial capacity may be affected by requirements to monitor the
effectiveness and reliability of their disinfection or removal, especially systems not currently using
disinfection. Ground water systems serving 3,300 people or fewer and using disinfection can conduct
daily grab samples to measure disinfection levels instead of installing more costly continuous monitoring
equipment. However, this may also require the system to increase staffing levels in addition to providing
training to ensure that system staff understand the compliance monitoring requirements. Reporting,
record-keeping, and data administration requirements will also affect the managerial capacity of small
systems.
Small systems that are required to take corrective action are expected to experience the most
significant financial challenge since many corrective actions consist of a large, one-time capital
expenditure to resolve the problem. Changes in treatment may also significantly impact the managerial
and technical capacity of the system.
7.5.4.2 Large Water Systems (Those Serving at Least 10,000 People)
Large systems will likely not face more than a minimal challenge to their technical and
managerial capacity as a result of efforts to familiarize themselves with the GWR and assist the State with
sanitary surveys. Although larger systems may need to take many source water samples a year to meet
the triggered monitoring requirements, most larger systems are familiar with total coliform monitoring
and already have the TMF capacity to address this increased burden.
Many larger systems already have the TMF capacity to address additional monitoring of the
effectiveness and reliability of disinfection or removal. These systems will be affected less significantly
than smaller systems, especially since some may already be disinfecting and conducting monitoring.
Large systems are expected to face the most significant challenge meeting the technical, managerial, and
financial requirements associated with corrective action. However, this requirement is only necessary
when a sanitary survey identifies a significant deficiency or when a source water monitoring sample tests
positive for fecal indicators.
7.6 Paperwork Reduction Act
The information collection requirements for the GWR have been approved by the Office of
Management and Budget (OMB) under the Paperwork Reduction Act, 44 U.S.C. 3501 et seq. The
information collected as a result of this rule will allow the States/Primacy Agencies and EPA to determine
appropriate requirements for specific systems and evaluate compliance with the rule.
The Paperwork Reduction Act requires EPA to estimate the burden on public water systems
(PWSs) and States/Primacy Agencies of complying with the rule. Burden means the total time, effort,
and financial resources required to generate, maintain, retain, disclose, or provide information to or for a
Federal agency. This burden includes the time needed to conduct these activities:
• Review instructions.
• Develop, acquire, install, and employ technology and systems for the purposes of
collecting, validating, verifying, processing, maintaining, and disclosing information.
Economic Analysis for the 7-10 October 2006
Final Ground Water Rule
-------
Adjust the existing ways to comply with any previously applicable instructions and
requirements.
• Train personnel to respond to information collected.
• Search data sources.
• Complete and review the collection of information.
• Transmit or otherwise disclose the information.
For the first 3 years after publication of the final GWR in the Federal Register, the major
information requirements are for States/Primacy Agencies and PWSs to prepare for implementation of the
rule. The information collection requirements are mandatory under Part 141 for systems and Part 142 for
States/Primacy Agencies. The calculation of GWR information collection burden and costs can be found
in the Information Collection Request for National Primary Drinking Water Regulations: Final Ground
Water Rule (USEPA 2006c).
The total burden associated with GWR requirements over the 3 years covered by the Information
Collection Request is 1,155,791 hours, an average of 385,264 hours per year. This is based on an
estimate that 57 States and territories will each need to provide 1 response each year with an average of
2,193 hours per response, and that 49,110 systems will each provide 2 responses each year with an
average burden of 2.6 hours per response.
The total reporting and recordkeeping cost over the 3-year clearance period of the Information
Collection Request is $30.3 million, an average of $10.1 million per year (simple average over 3 years).
The average annual cost per response is $103. The recordkeeping and reporting burden does not include
any capital costs for the first 3-year Information Collection Request period. Exhibit 7.2 provides a
summary of the results of the Information Collection Request calculations.
Exhibit 7.2 Average Annual Burden Hours and Costs for the GWR Information
Collection Request 3-Year Approval Period
PWSs
States &
Territories
Total
Responses
98,220
57
98,277
Burden
Hours
260,244
125,020
385,264
Labor Costs
$5,890,508
$4,200,914
$10,091,422
O&M
Costs
$0
$0
$0
Capital
Costs
$0
$0
$0
Total
Annual
Costs
$5,890,508
$4,200,914
$10,091,422
Note: Data represent burden and cost for only the 3-year Information Collection Request approval period. Data are
based on nominal (or undiscounted) values. Detail may not add due to independent rounding.
Source: Information Collection Request for the National Primary Drinking Water Regulations: Final Ground Water
Rule (USEPA 2006c).
Economic Analysis for the
Final Ground Water Rule
7-11
October 2006
-------
7.7 Unfunded Mandates Reform Act
The UMRA of 1995, Public Law 104-4, consists of four Titles and numerous sections. Sections
201 through 205 of Title II, entitled "Regulatory Accountability and Reform," are relevant to the GWR
and are discussed in this section. Title II, Section 201 of the UMRA, requires Federal agencies to assess
the effects of their regulatory actions on State, Local, and Tribal governments, and the private sector.
Under UMRA Section 202, EPA generally must prepare a written statement, including a cost-benefit
analysis, for proposed and final rules with "Federal mandates" that may result in expenditures by State,
Local, and Tribal governments, in the aggregate, or by the private sector, of $100 million or more in any 1
year. Section 203 requires the Agency to establish a small government agency plan before establishing
any regulatory requirements that may significantly or uniquely affect small governments.
Section 204 of the UMRA requires the Agency to develop an effective process to permit elected
officers of State, Local, and Tribal governments to provide meaningful and timely input in the
development of regulatory proposals that contain significant Federal intergovernmental mandates.
Finally, Section 205 generally requires EPA to identify and consider a reasonable number of regulatory
alternatives and adopt the least costly, most cost-effective, or least burdensome alternative that achieves
the objectives of the rule before promulgating a rule for which a written statement is needed under
Section 202. The provisions of Section 205 do not apply when they are inconsistent with applicable law.
Moreover, Section 205 allows EPA to adopt an alternative other than the least costly, most cost-effective,
or least burdensome alternative if the Administrator publishes with the final rule an explanation why that
alternative was not adopted.
EPA has determined that this rule will not result in expenditures of $ 100 million or more for
State, Local, and Tribal governments, in the aggregate or the private sector in any 1 year, as shown in
Exhibit 7.3.
Economic Analysis for the 7-12 October 2006
Final Ground Water Rule
-------
Exhibit 7.3 Ground Water System, State, and Tribal Estimated Costs
Ground Water Systems Costs
State Costs
Tribal Costs
Total Public
Ground Water Systems Costs
Total Private
GRAND TOTAL
3% Discount
Rate
$ 29.5
$ 11.8
$ 0.2
$ 41.5
$ 20.3
$ 20.3
$ 61.8
7% Discount
Rate
$ 30.0
$ 11.7
$ 0.2
$ 41.9
$ 20.4
$ 20.4
$ 62.3
Percent of 3%
Grand Total
Costs
48%
19%
0%
67%
33%
33%
100%
Percent of 7%
Grand Total
Costs
48%
19%
0%
67%
33%
33%
100%
Note: Detail may not add due to independent rounding. All values annualized in millions of 2003$.
Source: State Costs from Appendix D; Total Public and Private Costs from GWR Cost Model Output.
Although the GWR is not subject to the requirements of Sections 202 and 205 of UMRA, EPA
has prepared a written statement addressing the following items:
The authorizing legislation (Chapter 2)
Benefit-cost analysis including an analysis of the extent to which the costs of State, Local
and Tribal governments will be paid for by the Federal government (Chapter 8, section
7.7.1)
Estimates of future compliance costs and disproportionate budgetary effects (Chapter 6,
section 7.7.2)
• Macroeconomic effects (section 7.7.3)
• A summary of EPA's consultation with State, Local, and Tribal governments and their
concerns, including a summary of the Agency's evaluation of those comments and
concerns (sections 7.2, 7.7.7, 7.8, 7.11)
• Identification and consideration of regulatory alternatives and the selection of the least
costly, most cost-effective, or least burdensome alternative that achieves the objectives of
the rule (Chapters 3 and 8)
7.7.1 Social Benefits and Costs
The social benefits are those that primarily accrue to the public through an increased level of
protection from viral and bacterial illness due to exposure to microbial pathogens in drinking water. To
assign a monetary value to the illness, EPA used cost-of-illness (COI) estimates by age categories to
estimate the benefits from the reduction in acute viral illnesses and deaths avoided. This is considered to
be a lower-bound estimate of actual benefits because it does not include the pain and discomfort
associated with the illness. Mortalities were valued using a value of statistical life estimate consistent
Economic Analysis for the
Final Ground Water Rule
7-13
October 2006
-------
with EPA policy. Chapter 5 presents the benefits analysis, which includes both qualitative and monetized
benefits of improvements to health and safety. The estimated annualized benefit of the GWR using a 3
percent discount rate and the Enhanced COI approach is $19.7 million under the Final GWR.
In addition to reducing the number of viral illnesses and deaths, the GWR will also decrease
bacterial illness associated with fecal contamination of ground water. EPA did not directly calculate the
actual numbers of illnesses associated with bacterially contaminated ground water because the Agency
lacked the necessary pathogen occurrence data to include it in the risk model. However, in order to get an
estimate of the number of bacterial illnesses from fecally contaminated ground water, the Agency
developed an alternative calculation. This analysis of nonqualified benefits from avoided bacterial
illnesses and deaths is presented in section 5.4. This rule also considered but did not monetize the health
benefit from the reduction in chronic illness associated with some viral and bacterial infections.
Measuring the social costs of the rule requires identifying affected entities by ownership (public
or private), considering regulatory alternatives, calculating regulatory compliance costs, and estimating
any disproportionate impacts. Chapter 6 of this document details the cost analysis performed for the
GWR. Under the Preferred Alternative, the likely compliance scenario is expected to result in total
annualized costs of approximately $61.8 million using a 3 percent discount rate (or $62.3 million using a
7 percent discount rate). Exhibit 7.4 summarizes the range of annualized costs and benefits for each
regulatory alternative.
Exhibit 7.4 Mean Total Annualized Benefits and Costs of Regulatory Alternatives
(SMillions, 2003$)
Regulatory Alternative
Final GWR
(Risk Targeted Approach)
Alternative 1
(Sanitary Survey and Corrective
Action)
Alternative 3
(Multi-barrier Approach)
Alternative 4
(Across-the-Board Disinfection)
Enhanced
Annualized
Benefits (3%)
($Millions)
$19.7
$3.6
$21.3
$70.2
Traditional
Annualized
Benefits (3%)
($Millions)
$10.0
$1.9
$10.8
$35.5
Enhanced
Annualized
Benefits (7%)
($Millions)
$16.8
$2.9
$18.2
$61.9
Traditional
Annualized
Benefits (7%)
($Millions)
$8.6
$1.5
$9.3
$31.5
Annualized
Costs (3%)
($Millions)
$61.8
$15.3
$67.9
$686.4
Annualized
Costs (7%)
($Millions)
$62.3
$15.3
$69.4
$665.3
Source: Benefits from Exhibit 5.31. Costs from Exhibit 6.35.
Various Federal programs exist to provide financial assistance to State, Local, and Tribal
governments in complying with this rule. The Federal government provides funding to States/Primacy
Agencies that have primary enforcement responsibility for their drinking water programs through the
Public Water Systems Supervision (PWSS) Grants Program. States/Primacy Agencies may use these
funds to develop primacy programs or to contract with other State agencies to assist in the development or
implementation of their primacy programs. However, they may not use these funds to contract with
regulated entities (i.e., water systems). States/Primacy Agencies may use PWSS Grants to set up and
administer a State program that includes such activities as public education, testing, training, technical
assistance, development and administration of a remediation grant and loan or incentive program
(excluding the actual grant or loan funds), or other regulatory or nonregulatory measures.
Economic Analysis for the
Final Ground Water Rule
7-14
October 2006
-------
Additional funding is available from other programs administered by EPA or other Federal
agencies. These include EPA's Drinking Water State Revolving Fund (DWSRF), the U.S. Department of
Agriculture's Rural Utilities' Loan and Grant Program, and the Department of Housing and Urban
Development's Community Development Block Grant (CDBG) Program.
SDWA authorizes the EPA Administrator to award capitalization grants to States/Primacy
Agencies, which in turn can provide low-cost loans and other types of assistance to eligible PWSs. The
DWSRF assists PWSs with financing the costs of infrastructure needed to achieve or maintain compliance
with SDWA requirements. Each State has considerable flexibility to determine the design of its DWSRF
Program and to direct funding toward its most pressing compliance and public health protection needs.
States/Primacy Agencies may also, on a one-to-one matching basis, use up to 10 percent of their DWSRF
allotments for each fiscal year to assist in running the State drinking water program. In addition,
States/Primacy Agencies have the flexibility to transfer a portion of funds from their Clean Water State
Revolving Fund accounts to their DWSRF accounts.
A State/Primacy Agency can use the financial resources of the DWSRF to assist small systems.
In fact, a minimum of 15 percent of a State/Primacy Agency's DWSRF grant must be used to provide
infrastructure loans to systems serving 10,000 or fewer people. Two percent of the State/Primacy
Agency's grant is set-aside funding that can only be used to provide technical assistance to small systems.
In addition, up to 14 percent of the State/Primacy Agency's grant may be used to provide TMF assistance
to all system sizes. For small systems that are disadvantaged, up to 30 percent of a State/Primacy
Agency's DWSRF may be used for increased loan subsidies. Tribes have separate set-aside funding that
they can use under the DWSRF.
In addition to the DWSRF, money is available from the Department of Agriculture's Rural Utility
Service (RUS) and Housing and Urban Development's CDBG Program. RUS provides loans, guaranteed
loans, and grants to improve, repair, or construct water supply and distribution systems in rural areas and
towns with a population of up to 10,000 people. In fiscal year 2003, RUS had over $1.5 billion of
available funds for water and environmental programs. Also, three sources of funding exist under the
CDBG program to finance building and improvements of public faculties such as water systems. These
include: 1) direct grants to communities with populations over 200,000; 2) direct grants to States/Primacy
Agencies, which in turn are awarded to smaller communities, rural areas, and colonas in Arizona,
California, New Mexico, and Texas; and 3) direct grants to U.S. territories and trusts. The CDBG budget
for the formula program for fiscal year 2003 totaled over $4.4 billion.
7.7.2 Disproportionate Budgetary Effects
UMRA is intended to reduce the burden on State, Local, and Tribal governments of Federal
mandates that are not accompanied by adequate Federal funding. Section 202 of UMRA requires an
analysis of possible disproportionate budgetary effects of certain classes of rules, in which the GWR
falls.1 EPA believes that the cost estimates presented in Exhibit 7.5 accurately characterize future
compliance costs of the GWR. EPA explored possible disproportionate impacts of the GWR on particular
geographic areas and groups of customers. In general, the costs that a PWS, whether publicly- or
privately-owned, will incur to comply with this rule will depend on many factors that are not generally
1 "...[T]he agency shall prepare a written statement containing. . . (3) estimates by the agency, if and to the
extent that the agency determines that accurate estimates are reasonably feasible, of.. . (B) any disproportionate
budgetary effects of the Federal mandate upon any particular regions of the nation or particular State, Local, or
Tribal government, urban or rural or other types of communities, or particular segments of the private sector..."
Economic Analysis for the 7-15 October 2006
Final Ground Water Rule
-------
based on location. However, the data needed to confirm this assessment and to analyze other impacts of
this problem are not available; therefore, EPA looked at two other factors:
• The impacts of the rule on small versus large systems and the impacts within the five
small system size categories
The costs to publicly-owned versus privately-owned water systems
It is also possible that some States or EPA regions may face greater challenges from the GWR
than other States or regions because they have comparatively more ground water systems. However,
costs are not expected to be highly focused on a particularly geographic region or sector. In addition,
States that have a larger percentage of systems also receive a greater share of money under the PWSS
Grants Program and the DWSRF.
One measure performed of disproportionate impact is the cost incurred by small and large
systems. As a group, small systems will experience a greater impact than large systems under the GWR.
The higher total cost to the small ground water systems is due to the large number of these types of
systems (i.e., 99 percent of ground water systems serve fewer than 10,000 people). Other reasons for the
disparity include the following: 1) large systems are more likely to already be disinfecting their ground
water (disinfection exempts a system from triggered monitoring); 2) they typically have greater technical
and operational expertise; and 3) they are more likely to engage in source protection programs. The total
impacts on small systems (those serving fewer than 10,000 people) as well as large and medium systems
(those serving at least 10,000) are presented in Exhibit 7.5 for 3 percent and 7 percent discount rates.
Exhibit 7.5 Annualized Compliance Costs by Type of Ground Water System
Source Water Category
Annualized Cost to
Systems Serving
< 10,000 People
($ Millions)
3 Percent
7 Percent
Annualized Cost to
Systems Serving
> 10,000 People
($ Millions)
3 Percent
7 Percent
CWSs
Non-tribal Systems
Tribal Systems
Total
12.0
0.2
12.2
12.0
0.2
12.1
6.5
0
6.5
7.0
0
7.1
NTNCWSs
Non-tribal Systems
Tribal Systems
Total
4.8
0.0
4.9
4.8
0.0
4.8
0.1
0
0.1
0.1
0
0.1
TNCWSs
Non-tribal Systems
Tribal Systems
Total
Grand Total
26.3
0.0
26.3
$ 43.4
26.4
0.0
26.4
$ 43.4
0.1
0
0.1
$ 6.7
0.1
0
0.1
$ 7.2
Source: Total Costs from Appendix D minus Tribal Costs; Tribal Costs calculated from
Appendix D.
The mean cost per system for compliance is shown for large and small systems in Exhibit 7.6. The
cost per system is greater for larger systems, due to increased costs of disinfection and other costs that
increase with the size of the system. Among the small systems, the potential system-level economic
impact will be the greatest for systems serving 3,301 to 10,000 people.
Economic Analysis for the
Final Ground Water Rule
7-16
October 2006
-------
Exhibit 7.6 Mean Annualized Compliance Cost per Ground Water System by
System Size and Type
System Size
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
>1 Million
CWSs
$ 226
$ 251
$ 312
$ 331
$ 773
$ 1,973
$ 10,622
$ 18,369
$ 14
NTNCWSs
$ 187
$ 275
$ 354
$ 646
$ 1,361
$ 3,769
$ 5,813
$ 8,383
$
TNCWSs
$ 275
$ 372
$ 506
$ 721
$ 1,741
$ 4,127
$ 5,320
$ 11,921
$
Source: Derived from Appendix D.
A second measure of impact performed on small systems is the total cost to privately-owned
water systems compared to that incurred by publicly-owned water systems. Exhibit 7.3 reveals that 55
percent of small system compliance costs are borne by publicly-owned PWSs, while 45 percent is borne
by privately-owned PWSs. This difference results from the fact that more than 55 percent of small PWSs
using ground water are owned by public entities. EPA, therefore, expects publicly-owned systems as a
group to have a slightly larger share of the total costs of the rule, but it does not expect cost per system to
differ systematically with ownership. Most importantly, the rule protects the health of customers of all
covered drinking water systems regardless of the size or type of system.
7.7.3 Macroeconomic Effects
Under UMRA Section 202, EPA is required to estimate the potential macroeconomic effects of
the regulation. These include effects on productivity, economic growth, full employment, and Gross
Domestic Product (GDP) (USEPA 2000e). Macroeconomic effects tend to be measurable in nationwide
econometric models only if the economic impact of the regulation reaches 0.25 percent to 0.5 percent of
GDP. In 2003, real GDP was $10,321 billion (U.S. Department of Commerce BEA 2004); thus, a rule
would have to cost at least $26 billion annually to have a measurable effect. A regulation with a smaller
aggregate effect is unlikely to have any measurable impact, unless it is highly focused on a particular
geographic region or economic sector. The GWR should not have a measurable effect on the national
economy; the total annualized costs for the rule range from $61.8 to $62.3 million using a 3 and 7 percent
discount rate, respectively. Using these annualized figures as a measure, the annual cost of the GWR is
an insignificant fraction of a $26 billion annual cost that would be considered a measurable
macroeconomic impact. Thus, annualized GWR costs measured as a percentage of the national GDP will
only decline over time as GDP grows.
Economic Analysis for the
Final Ground Water Rule
7-17
October 2006
-------
7.7.4 Consultation with Small Governments
Before the Agency establishes any regulatory requirements that may significantly or uniquely
affect small governments, including Tribal governments, it must have developed, under Section 203 of
UMRA, a small government agency plan. The plan must provide for the notification of potentially
affected small governments, enabling officials of affected small governments to have meaningful and
timely input in the development of EPA regulatory proposals with significant Federal intergovernmental
mandates and informing, educating, and advising small governments on compliance with the regulatory
requirements. EPA consulted with small governments to address impacts of regulatory requirements in
the GWR that might significantly or uniquely affect small governments. A variety of stakeholders,
including small governments, were provided with several opportunities to participate early in the
regulatory development process, as described in section 7.4.
7.7.5 Consultation with State, Local, and Tribal Governments
Section 204 of UMRA requires the Agency to develop an effective process to permit elected
officers of State, Local, and Tribal governments (or their designated authorized employees) to provide
meaningful and timely input in the development of regulatory proposals that contain significant Federal
intergovernmental mandates. Consistent with these provisions, EPA held consultations with affected
governmental entities prior to proposal of the rule, as described in sections 7.4 and 7.7. EPA conducted
four public meetings for all stakeholders and two Association of State Drinking Water Administrators
(ASDWA) early involvement meetings. Because of the GWR's impact on small entities, the Agency
convened a SBAR Panel in accordance with the RFA as amended by the SBREFA to address small entity
concerns, including small local governments specifically. EPA consulted with small entity stakeholders
prior to convening the SBAR Panel to get their input on the GWR. Of the 22 small entity participants,
five represented small governments. EPA also made presentations on the GWR to the national and local
chapters of the American Water Works Association, the Ground Water Foundation, the National Ground
Water Association, the National Rural Water Association, and the National League of Cities. Twelve
State drinking water representatives also participated in the Agency's GWR workgroup.
In addition to these consultations, EPA circulated a draft of the proposed rule and requested
comment from the public through an informal process. Specifically, on February 3, 1999, EPA posted a
draft proposal on their webpage and mailed out over 300 copies to people who had attended the 1997 and
1998 public stakeholder meetings, as well as people on the EPA workgroup. EPA received 79 letters or
electronic responses to this draft: 34 from State governments (representing 30 different States), 25 from
local governments, 10 from trade associations, 6 from Federal government agencies, and 4 from other
people/organizations. No comments were received from Tribal governments. EPA reviewed the
comments and carefully considered their merit. The GWR reflects many of the commentors' points and
suggestions.
EPA will educate, inform, and advise small systems, including those operated by small
governments, about the GWR requirements. One of the most important components of this process will
be the Small System GWR Implementation Guidance, which is required by SBREFA of 1996. This
plain-English guide will explain what actions a small entity must take to comply with the rule. The
Agency is also developing fact sheets that concisely describe various aspects and requirements of the
GWR. Additional details on Tribal involvement in the rulemaking process can be found in section 7.7.
Economic Analysis for the 7-18 October 2006
Final Ground Water Rule
-------
7.7.6 Regulatory Alternatives Considered
As required under Section 205 of UMRA, EPA considered several regulatory alternatives and
numerous methods that would reduce microbial contamination in ground water systems. Chapter 3
provides a detailed discussion of these alternatives. EPA believes that Alternative 2, the risk targeted
approach, is the most cost effective alternative that achieves the rule's objective to reduce the risk of
illness and death from microbial contamination in PWSs relying on ground water. This alternative is a
targeted approach where costs are driven by the number of systems having to fix fecal contamination
problems and correct significant deficiencies that could lead to fecal contamination.
7.7.7 Impacts on Small Governments
In developing this rule, EPA consulted with small governments pursuant to Section 203 of
UMRA to address impacts of regulatory requirements in the rule that might significantly or uniquely
affect small governments. In preparation for the GWR, EPA conducted an analysis on small government
impacts and included small government officials or their designated representatives in the rulemaking
process. A variety of stakeholders, including small governments, had the opportunity for timely and
meaningful participation in the regulatory development process through the SBREFA process, public
stakeholder meetings, and Tribal meetings. Representatives of small governments took part in the
SBREFA process for this rulemaking and attended public stakeholder meetings. Through participation
and exchange in the SBREFA process and various meetings, EPA notified some potentially affected small
governments of requirements under consideration and provided officials of affected small governments
with an opportunity to have meaningful and timely input into the development of regulatory proposals.
EPA has determined that this rule contains regulatory requirements that might significantly or
uniquely affect small governments. As shown in Exhibit 7.6, estimated annual expenditures per small
system for the GWR range from $226 to $773 for CWSs, $187 to $1,361 for NTNCWSs, and $275 to
$1,741 for TNCWSs (at a 3 percent discount rate).
7.8 Indian Tribal Governments
Executive Order 13175, entitled "Consultation and Coordination with Indian Tribal
Governments" (65 FR 67249; November 6, 2000), requires EPA to develop "an accountable process to
ensure meaningful and timely input by Tribal officials in the development of regulatory policies that have
Tribal implications." The Executive Order defines "policies that have Tribal implications" to include
regulations that have "substantial direct effects on one or more Indian Tribes, on the relationship between
the Federal government and the Indian Tribes, or on the distribution of power and responsibilities
between the Federal government and Indian Tribes."
Under Executive Order 13175, EPA may not issue a regulation that has Tribal implications, that
imposes substantial direct compliance costs, and that is not required by statute, unless the Federal
government provides the funds necessary to pay the direct compliance costs incurred by Tribal
governments, or EPA consults with Tribal officials early in the process of developing the proposed
regulation and develops a Tribal summary impact statement.
EPA performed an analysis to estimate the impact of the GWR on Tribal systems. More than 87
percent of PWSs in Indian Country (727 systems) are ground water systems and will, therefore, be
affected by the GWR. Based on the analysis, EPA concluded that the GWR may have Tribal implications
because it may impose substantial direct compliance costs on Tribal governments, and the Federal
Economic Analysis for the 7-19 October 2006
Final Ground Water Rule
-------
government will not provide the funds necessary to pay the direct costs incurred by the Tribal
governments in complying with the rule. Accordingly, EPA provides the following Tribal summary
impact statement as required by Section 5(b) of Executive Order 13175.
As described in section 7.5.5, EPA held extensive public meetings that provided tribes with the
opportunity for meaningful and timely input into the development of the GWR. Summaries of the
meetings have been included in the public docket for this rulemaking. In addition, the Agency presented
the rule and asked for comment at three Tribal conferences. Two outreach efforts were conducted at
national conferences; one for the National Indian Health Board and the other for the National Tribal
Environmental Council. The third outreach effort took place in conjunction with the Inter-Tribal Council
of Arizona, Inc.
Tribal Summary Impact Statement
EPA performed an analysis to estimate the impact of the GWR on Tribal systems. EPA has
identified 727 Indian Tribal systems that might be subject to the GWR. As seen in Exhibit 7.7, all but
three Tribal systems are classified as small systems (serving fewer than 10,000 people).
Economic Analysis for the 7-20 October 2006
Final Ground Water Rule
-------
Exhibit 7.7 Annual Cost of Compliance for Tribal Systems by System
Type and Size (Annualized at 3 Percent)
System
Size/Type
Number of
Systems
Affected
byGWR
A
Systems
Conducting
Implementation
Activities
%
B
No.
C = B*A
Systems
Conducting
Santitary
Surveys
%
D
No.
E = D*A
Systems
Conducting
HSAs
%
F
No.
G = F*A
Systems
Conducting
Triggered
Monitoring
%
H
No.
I = H*A
Systems
Conducting
Assessment
Monitoring
%
J
No.
K=J*A
Sytems Taking
Corrective
Action
%
L
No.
M = L*A
Systems
Conducting
Compliance
Monitoring
%
N
No.
0 = N*A
Mean
Annualized
Cost per
System
P
Estimated
Total Tribal
Costs
Q = A*P
Primarily Disinfecting Ground Water CWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
Subtotal
171
223
69
69
21
3
0
0
0
556
100%
100%
100%
100%
100%
NA
NA
NA
NA
171
223
69
69
21
NA
NA
NA
NA
553
100%
100%
100%
100%
100%
NA
NA
NA
NA
171
223
69
69
21
NA
NA
NA
NA
553
0%
0%
0%
0%
0%
NA
NA
NA
NA
0
0
0
0
0
NA
NA
NA
NA
0
76%
62%
61%
63%
62%
NA
NA
NA
NA
130
138
42
43
13
NA
NA
NA
NA
367
0%
0%
0%
0%
0%
NA
NA
NA
NA
0
0
0
0
0
NA
NA
NA
NA
0
24%
24%
24%
22%
24%
NA
NA
NA
NA
42
53
16
15
5
NA
NA
NA
NA
132
31%
45%
46%
42%
45%
NA
NA
NA
NA
53
100
32
29
9
NA
NA
NA
NA
224
$ 226
$ 251
$ 312
$ 331
$ 773
$
$
$
$
$ 38,686
$ 56,084
$ 21 ,562
$ 22,851
$ 16,230
$
$
$
$
$ 155,413
Primarily Disinfecting Ground Water NTNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
Subtotal
34
28
13
21
2
0
0
0
0
98
100%
100%
100%
100%
100%
100%
NA
NA
NA
34
28
13
21
2
0
NA
NA
NA
98
100%
100%
100%
100%
100%
100%
NA
NA
NA
34
28
13
21
2
0
NA
NA
NA
98
0%
0%
0%
0%
0%
0%
NA
NA
NA
0
0
0
0
0
0
NA
NA
NA
0
91%
91%
91%
91%
91%
91%
NA
NA
NA
31
25
12
19
2
0
NA
NA
NA
89
0%
0%
0%
0%
0%
0%
NA
NA
NA
0
0
0
0
0
0
NA
NA
NA
0
24%
25%
25%
29%
31%
33%
NA
NA
NA
8
7
3
6
1
0
NA
NA
NA
25
16%
17%
17%
21%
23%
25%
NA
NA
NA
6
5
2
4
0
0
NA
NA
NA
17
$ 187
$ 275
$ 354
$ 646
$ 1 ,361
$ 3,769
$
$
$
$ 6,371
$ 7,689
$ 4,607
$ 13,568
$
$
$
$
$
$ 32,235
Primarily Disinfecting Ground Water TNCWSs
<100
101-500
501-1,000
1,001-3,300
3,301-10,000
10,001-50,000
50,001-100,000
100,001-1 Million
> 1 Million
Subtotal
TOTALS
57
11
3
1
1
0
0
0
0
73
727
100%
100%
100%
100%
NA
100%
NA
NA
NA
57
11
3
1
NA
0
NA
NA
NA
72
723
100%
100%
100%
100%
NA
100%
NA
NA
NA
57
11
3
1
NA
0
NA
NA
NA
72
723
0%
0%
0%
0%
NA
0%
NA
NA
NA
0
0
0
0
NA
0
NA
NA
NA
0
0
98%
98%
98%
98%
NA
98%
NA
NA
NA
56
11
3
1
NA
0
NA
NA
NA
71
527
0%
0%
0%
0%
NA
0%
NA
NA
NA
0
0
0
0
NA
0
NA
NA
NA
0
0
28%
28%
28%
30%
NA
35%
NA
NA
NA
16
3
1
0
NA
0
NA
NA
NA
20
177
13%
12%
13%
15%
NA
20%
NA
NA
NA
7
1
0
0
NA
0
NA
NA
NA
9
250
$ 275
$ 372
$ 506
$ 721
$
$ 4,127
$
$
$
$ 15,688
$ 4,096
$ 1,517
$ 721
$
$
$
$
$
$ 22,022
$ 209,670
Source: Derived from Appendix D.
Economic Analysis for the
Final Ground Water Rule
7-21
October 2006
-------
EPA has estimated costs for Indian Tribal systems to comply with the GWR, based on the
assumption that the percentages of systems expected to incur costs for each size category will be the same
for Tribal systems as for systems nationwide. The costs for Tribal systems are calculated in two steps.
First, the number of Indian Tribal systems in each size category is multiplied by the percentage of systems
nationally in each size category expected to incur costs for various rule activities. Second, the average
cost of each rule requirement is multiplied by the number of Tribal systems expected to incur costs.
Exhibit 7.7 shows the percentage of systems expected to incur costs for various compliance activities.
These costs result in an estimated total annualized cost to Indian Tribes of $209,670 for the GWR.
7.9 Impacts on Sensitive Subpopulations
EPA's Office of Water has historically considered risks to sensitive subpopulations (including
children) in establishing drinking water assessments, advisories or other guidance, and standards.
Generally, the health effects of many pathogens and viruses on sensitive subpopulations is much more
severe and debilitating than on the general population. These sensitive subpopulations include the young,
elderly (especially those weakened by other conditions), malnourished and disease-impaired (especially
those with diabetes), and a broad category of those with compromised immune systems, such as Acquired
Immune Deficiency Syndrome (AIDS) patients, individuals with Lupus or cystic fibrosis, transplant
recipients, and individuals on chemotherapy (Rose 1997). In total, these subgroups represent almost 20
percent of the current population of the United States.
Pregnant and lactating women may be at an increased risk from enteric viruses as well as act as a
source of infection for neonates. Infection during pregnancy may also result in the transmission of
infection from the mother to the child in utero, during birth, or shortly thereafter. Since very young
children do not have fully developed immune systems, they are at increased risk and are particularly
difficult to treat.
Infectious diseases are also a major problem for the elderly because immune function declines
with age. As a result, outbreaks of waterborne diseases can be devastating on the elderly community
(e.g., nursing homes) and may increase the possibility of significantly higher mortality rates in the elderly
than in the general population.
Immunocompromised individuals are a growing proportion of the population with the relatively
new and severe problem magnified by the AIDS epidemic and the escalation in organ and tissue
transplantations. Enteric pathogens take advantage of the impaired immune systems of these individuals
and set up generalized and persistent infections in the immunocompromised host. These infections are
particularly difficult to treat and can result in a significantly higher mortality than in immunocompetent
persons.
With regard to sensitive sub-populations, EPA explicitly examined the effects of the GWR on
young children, the elderly, and immunocompromised individuals. Exhibit 5.24 in Chapter 5 shows the
estimated number of illnesses and deaths avoided in each of these categories and the values of the
associated benefits. In addition to the information presented in Chapter 5 of this EA, research outlining
the potential health benefits of the GWR to both sensitive subpopulations and the general public is
discussed in greater detail in the Occurrence and Monitoring Document for the Final Ground Water Rule
(USEPA 2006b).
Economic Analysis for the 7-22 October 2006
Final Ground Water Rule
-------
7.9.1 Protection of Children from Environmental Health Risks and Safety Risks
Executive Order 13045 (62 FR 19885; April 23, 1997) applies to any rule initiated after April 21,
1998, that (1) is determined to be "economically significant" as defined under Executive Order 12866;
and (2) concerns an environmental, health, or safety risk that EPA has reason to believe may have a
disproportionate effect on children. If the regulatory action meets both criteria, EPA must evaluate the
environmental, health, or safety effects of the planned rule on children, and explain why the planned
regulation is preferable to other potentially effective and reasonably feasible alternatives considered by
the Agency. This final rule is not subject to the Executive Order because it is not economically
significant as defined in Executive Order 12866. As a matter of policy, EPA has examined the
environmental health or safety effects of viruses on children.
The risk of illness and death due to viruses depends on several factors, including the type of virus,
age, nutrition, exposure, genetic variability, presence of disease, and the immune status of the individual.
Rotavirus infections can occur in people of all ages, but they primarily affect young children. In addition,
infants and young children have higher rates of infection and disease from enteroviruses than other age
groups (USEPA 1999). Several viruses that can be transmitted through water, including poliovirus,
coxsackievirus, and echovirus, can have serious health consequences in children, which are discussed in
detail in Chapter 5.
In developing the risk and benefits analysis for the GWR, the effects on children, both in terms of
unique risk and cost-of-illness estimates, were explicitly taken into consideration, as discussed in Chapter
5 of this EA. This analysis suggests that the rule provides a greater per capita health benefit to children
than to adults, mostly due to the high cost-of-illness associated with viral illnesses avoided in young
children. In other words, the analysis suggests that the viral and bacterial illnesses of concern to the
GWR disproportionately affect children, and therefore, the benefits of the proposed rule accrue
disproportionately to children.
7.10 Environmental Justice
Executive Order 12898 (59 FR 7629) establishes a Federal policy for incorporating
environmental justice into Federal agency missions by directing agencies to identify and address
disproportionately high adverse human health or environmental effects of its programs, policies, and
activities on minority and low-income populations. The Agency has considered environmental justice
related issues concerning the potential impacts of this action and consulted with minority and low-income
stakeholders.
Two aspects of the GWR comply with the order that requires the Agency to consider
environmental justice issues in the rulemaking and to consult with stakeholders representing a variety of
economic and ethnic backgrounds. These are: (1) the overall nature of the rule, and (2) the convening of
a stakeholder meeting specifically to address environmental justice issues.
The GWR applies uniformly to CWSs, NTNCWSs, and TNCWSs that use ground water as their
source. Consequently, this rule provides health protection from pathogen exposure equally to all income
and minority groups served by ground water systems. Existing regulations, such as the Surface Water
Treatment Rule (SWTR) and the Interim Enhanced Surface Water Treatment Rule (IESWTR), provide
similar health benefit protection to communities that use surface water or GWUDI.
The Agency built on the efforts conducted during the IESWTR's development to comply with
Executive Order 12898. On March 12, 1998, the Agency held a stakeholder meeting to address various
Economic Analysis for the 7-23 October 2006
Final Ground Water Rule
-------
components of pending drinking water regulations and how they might impact sensitive subpopulations,
minority populations, and low-income populations. This meeting was a continuation of stakeholder
meetings that started in 1995 to obtain input on the Agency's Drinking Water Programs. Topics
discussed included treatment techniques, costs and benefits, data quality, health effects, and the regulatory
process. Participants were national, State, Tribal, municipal, and individual stakeholders. EPA
conducted the meeting by video conference call among 11 cities. The major objectives for the March 12,
1998 meeting included the following:
Solicit ideas from stakeholders on known issues concerning current drinking water regulatory
efforts.
• Identify key areas of concern to stakeholders.
• Receive suggestions from stakeholders concerning ways to increase representation of
communities in OGWDW regulatory efforts.
In addition, EPA developed a plain-English guide for this meeting to assist stakeholders in
understanding the multiple and sometimes complex issues surrounding drinking water regulations.
The GWR and other drinking water regulations are expected to have a positive effect on human
health regardless of the social or economic status of a specific population. The GWR serves to provide a
similar level of drinking water protection to all groups. To the extent that levels of bacteria and viruses in
drinking water might be disproportionately high now among minority or low-income populations (which
is unknown), the GWR will work to remove those differences. Thus, the GWR meets the intent of
Federal policy requiring incorporation of environmental justice into Federal agency missions.
The GWR applies uniformly to PWSs that use ground water as their source. Consequently, this
rule provides health protection from pathogenic bacteria and viruses exposure equally to all income and
minority groups served by ground water systems.
7.11 Federalism
Executive Order 13132, "Federalism" (64 FR 43255; August 10, 1999), requires EPA to develop
an accountable process to ensure "meaningful and timely input by State and Local officials in the
development of regulatory policies that have Federalism implications." "Policies that have Federalism
implications" are defined in the Executive Order to include regulations that have "substantial direct
effects on the States, on the relationship between the national government and the States, or on the
distribution of power and responsibilities among the various levels of government."
Under Section 6(b) Executive Order 13132, EPA may not issue a regulation that has Federalism
implications, that imposes substantial direct compliance costs, and that is not required by statute, unless
the Federal government provides the funds necessary to pay the direct compliance costs incurred by State
and Local governments, or EPA consults with State and Local officials early in the process of developing
the proposed regulation.
If EPA complies by consulting, Executive Order 13132 requires EPA to provide to OMB, in a
separately identified section of the preamble to the final rule, a Federalism Summary Impact Statement
(FSIS). The FSIS must include a description of the extent of EPA's prior consultation with State and
Local officials, a summary of the nature of their concerns, and the Agency's position supporting the need
to issue the regulation, and a statement of the extent to which the concerns of State and Local officials
Economic Analysis for the 7-24 October 2006
Final Ground Water Rule
-------
have been met. Also, when EPA transmits a draft final rule with Federalism implications to OMB for
review pursuant to Executive Order 12866, EPA must include a certification from the Agency's
Federalism Official stating that EPA has met the requirements Executive Order 13132 in a meaningful
and timely manner.
EPA has concluded that this rule does not have Federalism implications because it does not
impose substantial direct compliance costs on State and Local governments. The cost to State, Local, and
Tribal governments in the aggregate is $41.5 million (see Exhibit 7.3) on average annually at a 3 percent
discount rate.
Nonetheless, as discussed in section 7.5.4, EPA met with a variety of State and Local
representatives, including several local elected officials, who provided meaningful and timely input in the
development of the GWR. Summaries of the meetings have been included in the public record for this
rulemaking. EPA consulted extensively with State and Local governments. For example, four public
stakeholder meetings were held in Washington, DC, Portland, OR, Madison, WI, and Dallas, TX. EPA
also held three early involvement meetings with ASDWA. Several key issues were raised by stakeholder
regarding the GWR provisions, many of which were related to reducing burden and increasing flexibility
by creating a targeted risk based approach that builds upon existing State programs. Its should be noted
that this rule is important because it will reduce the incidence of fecally contaminated drinking water
supplies by requiring corrective actions for fecally contaminated systems or systems with a significant
risk of fecal contamination resulting in a reduced waterborne illness.
Initial consultation of the GWR occurred before November 2, 1999, the effective date of
Executive Order 13132. However, EPA initiated discussions with State and Local elected officials
regarding the implications of the rule during the public comment period and took their recommendations
under consideration during the development of the final rule requirements.
7.12 Actions Concerning Regulations That Significantly Affect Energy Supply,
Distribution, or Use
Executive Order 13211, "Actions Concerning Regulations That Significantly Affect Energy
Supply, Distribution, or Use" (66 FR 28355; May 22, 2001), provides that agencies shall prepare and
submit to the Administrator of the Office of Information and Regulatory Affairs, OMB, a Statement of
Energy Effects for certain actions identified as "significant energy actions." Section 4(b) of Executive
Order 13211 defines "significant energy actions" as "any action by an agency (normally published in the
Federal Register) that promulgates or is expected to lead to the promulgation of a final rule or regulation,
including notices of inquiry, advance notices of proposed rulemaking, and notices of proposed
rulemaking: (l)(i) that is a significant regulatory action under Executive Order 12866 or any successor
order, and (ii) is likely to have a significant adverse effect on the supply, distribution, or use of energy; or
(2) that is designated by the Administrator of the Office of Information and Regulatory Affairs as a
significant energy action."
The GWR has not been designated by the Administrator of the Office of Information and
Regulatory Affairs as a significant energy action because it is not likely to have a significant adverse
effect on the supply, distribution, or use of energy. This determination is based on the analysis presented
below.
Economic Analysis for the 7-25 October 2006
Final Ground Water Rule
-------
Energy Supply
The first consideration is whether the GWR would adversely affect the supply of energy. The
GWR does not regulate power generation, either directly or indirectly, and the public and private PWSs
that the GWR regulates do not, as a rule, generate power. Further, the cost increases borne by customers
of PWSs as a result of the GWR are a low percentage of the total cost of water, except for very few small
systems that will need to spread the cost of installing advanced technologies over a narrow customer base.
Therefore, the customers that are power generation utilities are unlikely to face any significant effects as a
result of the GWR. In summary, the GWR does not regulate the supply of energy, does not generally
regulate the utilities that supply energy, and is unlikely to significantly affect the customer base of energy
suppliers. Thus, the GWR would not adversely affect the supply of energy.
In response to the GWR, some water utilities are expected to increase their energy use, and those
impacts are discussed later in this section.
Energy Distribution
The second consideration is whether the GWR would adversely affect the distribution of energy.
The GWR does not regulate any aspect of energy distribution. PWSs that are regulated by the GWR
already have electrical service. The rule is projected to increase peak electricity demand at PWSs by only
0.001 percent (see below). Therefore, EPA assumes that the existing connections are adequate and that
the GWR has no discernable adverse effect on energy distribution.
Energy Use
The third consideration is whether the GWR would adversely affect the use of energy. Because
some PWSs are expected to add treatment technologies and security that use electrical power, this
potential impact of the GWR on the use of energy requires further evaluation. The analyses that underlay
the estimation of costs in Chapter 6 are national in scope and do not identify specific plants or systems
that may install treatment in response to the GWR. As a result, no analysis of the effect on specific
energy suppliers is possible with the available data. The approach used to estimate the impact of energy
use, therefore, focuses on national-level impacts. It estimates the additional energy use due to the GWR
and compares that to the national levels of power generation in terms of average and peak loads.
The first step is to estimate the energy used by the technologies or corrective action expected to be
installed as a result of the GWR. Energy use is not directly estimated in Technology and Cost Document
for the Final Ground Water Rule (USEPA 2006d), but the annual cost of energy for each technology and
corrective action addition or upgrade necessitated by the GWR is provided. An estimate of plant-level
energy use is derived by dividing the total energy cost per plant for a range of flows by an average national
cost of electricity of $0.076 per kilowatt hour per year (kWh/y) (USDOE EIA 2004a2). The energy use per
plant for each flow range and technology or corrective action is then multiplied by the number of plants
EPA is aware that DOE has updated its 2003 "average national cost of electricity per kilowatt hour per
year" from $0.076 to $0.074. However, EPA continues to use the $0.076 value to maintain consistency with the
Technology and Cost Document for the Final Ground Water Rule (USEPA 2006d).
Economic Analysis for the 7-26 October 2006
Final Ground Water Rule
-------
predicted to install each technology in a given flow range. The energy requirements for each flow range
are then added to produce a national total. No electricity use is subtracted to account for the technologies
that may be replaced by new technologies, resulting in a conservative estimate of the increase in energy
use. An incremental national annual energy usage is estimated at 4,521 megawatt hours (MWh); results of
the analysis are shown in Exhibit 7.8.
Exhibit 7.8 Total Increased Annual National Energy Usage
Attributable to the GWR
Technology/Corrective Action
Gas Chlorination
Hypochlori nation
CIO2
Anodic Oxidation
Ozonation
NF
Total
Plants Selecting
Technology
A
177
3,565
39
16
0
10
3,808
Total Annual
Energy Required
(kWh/yr)
B
190,099
3,726,479
11,928
95,342
17,229
479,477
4,520,555
Sources:
[A] Plants selecting technology taken from Exhibits 6.5b and 6.21b.
[B] CIO2, Ozonation, and NF -Total annual energy required calculated from energy costs
given in the Technology and Cost Document for the Final Ground Water Rule
(USEPA2006d) assuming $0.076/kWh. Gas Chlorination, Hypochlorination, and Anodic
Oxidation - Total annual energy required obtained from the Water and W/W Models.
Exhibit 7.9 provides a sample calculation for chlorine dioxide showing the increase in energy
usage as a result of the GWR.
To determine if the additional energy required for systems to comply with the rule would have a
significant adverse effect on the use of energy, the numbers in Exhibit 7.9 are compared to the national
production figures for electricity. According to the U.S. Department of Energy's Information
Administration, electricity producers generated 3,848 million MWh of electricity in 2003 (USDOE EIA
2004b3). Using the assumed energy use for the GWR (4,520,555 kWh/y), the rule would result in only a
0.0001 percent increase in annual average energy use when fully implemented. This calculation is shown
below:
4,521 MWh/y - 3,848,000,000 MWh/y * 100 = 0.0001%
EPA is aware that DOE has updated its estimate of total electricity produced in 2003 from 3,848 million
to 3,883 million. However, EPA continues to use the 3,848 million estimate to maintain consistency with related
electricity estimates used in this EA and the Technology and Cost Document for the Final Ground Water Rule.
Economic Analysis for the
Final Ground Water Rule
7-27
October 2006
-------
Exhibit 7.9 Sample Calculation for Determining Increase in Energy Usage:
Chlorine Dioxide (CIO2 Dose = 1.25 mg/L)
System Size
(population
served)
Average Daily Flow
Flow
(MGD)
A
Total No.
of
Entry Points
B
Number of
Entry Points
Selecting
C
Annual Energy
Cost per Entry Point
($/EP/yr)
D
Annual Energy
Requirement
(kWhr/EP/yr)
E = D/$0.076 per
kWhr
Total Energy Usage for
Entry Points Selecting
(kWhr/year)
F=C*E
CWSs
< 100
101 -500
501 - 1 ,000
1,001 -3,300
3,301 - 10,000
10,001 -50,000
50,001 -100,000
100,001 - 1 Million
> 1 Million
0.004
0.014
0.037
0.081
0.199
0.441
0.718
2.263
18.107
12,857
14,534
5,536
8,966
5,734
4,257
1,308
730
-
3
3
-
-
-
0
1
-
-
-
-
261
261
261
262
263
272
378
-
-
3,437
3,437
3,437
3,444
3,455
3,576
4,973
-
-
-
-
-
1,114
4,919
-
-
NTNCWSs
<100
101 -500
501 - 1 ,000
1,001 -3,300
3,301 - 10,000
10,001 -50,000
50,001 -100,000
100,001 - 1 Million
> 1 Million
0.004
0.022
0.071
0.175
0.637
2.856
9.918
17.188
-
8,606
6,150
1,724
651
66
9
1
1
-
2
2
-
-
-
0
0
-
-
-
261
261
261
262
276
310
370
-
-
3,437
3,437
3,437
3,452
3,626
4,080
4,868
-
-
5,839
-
-
-
14
6
-
-
TNCWSs
<100
101 -500
501 - 1 ,000
1,001 -3,300
3,301 - 10,000
10,001 -50,000
50,001 -100,000
100,001 - 1 Million
> 1 Million
TOTALS
0.003
0.017
0.068
0.159
0.619
2.124
7.649
19.724
-
63,288
18,651
1,905
574
73
19
1
1
-
155,641
21
7
-
-
-
0
0
-
-
39
-
-
261
261
262
273
297
416
-
-
-
3,437
3,437
3,451
3,587
3,908
5,475
-
75,392
-
-
-
-
-
31
6
-
-
11,928
Notes: Detail may not add due to independent rounding.
Sources: [A] Flows taken from Exhibit 4.6
[B] Total number of entry points taken from Exhibit 4.3
[C] Number of entry plants selecting chlorine dioxide taken from Exhibit 6.21 b
[D] Energy cost per entry point interpolated from CIO2 energy costs in the Technology and Cost Document
for the Final Ground Water Rule (USEPA 2006d)
[E] Electricity cost is $0.076/KWh, as presented in the Technology and Cost Document for the Final Ground
Water Rule (USEPA 2006d).
Economic Analysis for the
Final Ground Water Rule
7-28
October 2006
-------
In addition to average energy use, the impact at times of peak power demand is important. To
examine whether increased energy usage might significantly affect the capacity margins of energy
suppliers, their peak-season generating capacity reserve was compared to an estimate of peak incremental
power demand by water utilities. Both energy use and water use peak in the summer months, so the most
significant effects on supply would be seen then. During the summer of 2003, U.S. generation capacity
exceeded consumption by 15 percent, or approximately 160,000 MW (USDOE EIA 2004b4). Assuming
around-the-clock operation of water treatment plants, the total energy requirement for technologies can be
divided by 8,760 hours per year. Twelve hours of operation per day was assumed for the of security
system light bulbs. The sum of these two average power demands was 0.52 MW. Assuming that power
demand is proportional to water flow through the plant and that peak flow can be as high as twice the
average daily flow during the summer months, about 1.03 MW could be needed for treatment technologies
and security installed to comply with the GWR. This is only 0.001 percent of the capacity margin
available at peak use. This calculation is presented below:
1. Treatment Technologies: 4,520,555 kWh/y * (y/8,760 hr) * (MW/1,000 kW) * 2 = 1.03 MW
2. 1.03 MW - 160,000 MW * 100 = 0.001%
Although EPA recognizes that not all regions have a 15 percent capacity margin and that this
margin varies across regions and through time, this analysis reflects the effect of the rule on national
energy supply, distribution, and use. While certain areas have experienced shortfalls in generating
capacity in the recent past, a peak incremental power requirement of 1.03 MW nationwide is not likely to
significantly change the energy supply, distribution, or use in any given area.
Conclusion
The GWR is not a "significant energy action" as defined in Executive Order 13211, "Actions
Concerning Regulations That Significantly Affect Energy Supply, Distribution, or Use" (66 FR 28355;
May 22, 2001) because it is not likely to have a significant adverse effect on the supply, distribution, or
use of energy (as a function of annual average use and conditions of peak power demand).
The total increase in energy usage by water systems as a result of the GWR is predicted to be
approximately 4.5 million kWh/y, which is only one-ten-thousandth of 1 percent of the total energy
produced in 2003. While the rule may have some adverse energy effects, EPA does not believe that this
constitutes a significant adverse effect on the energy supply.
4 EPA is aware that DOE has updated its estimate of capacity exceeding consumption in the summer of
2003 from 160,000 to 159,000 MW. However, EPA continues to use the estimate of 160,000 MW to maintain
consistency with related electricity estimates used in this EA and the Technology and Cost Document for the Final
Ground Water Rule (USEPA 2006d).
Economic Analysis for the 7-29 October 2006
Final Ground Water Rule
-------
8. Comparison of Quantified Benefits and Costs of the GWR
8.1 National Quantified Benefits and Costs of the GWR
This chapter presents a summary of the quantified benefits and costs of the Ground Water Rule
(GWR). The nonquantified benefits are summarized in Chapter 5, Section 5.4. EPA estimates that the
quantified benefits are small compared with the nonquantified benefits. In particular, EPA estimates that
the nonquantified benefits associated with bacterial illness would increase total benefits to about five
times the quantified benefits associated with viral illness from Type A virus (represented by data from
rotavirus) and Type B virus (represented by data from enterovirus and echovirus). Other non-bacterial
health and non-health benefits also accrue but are not quantified.
The following sections present summary results from the quantified economic analysis, followed
by a discussion of the results based on evaluation of the total benefits, both quantified and nonquantified.
The first sections of this chapter focus on analysis of the final GWR, followed by a comparison of these
requirements to the other regulatory alternatives considered.
8.1.1 National Quantified Benefits Summary
The quantified benefits of the GWR derive from the reduction in risk of endemic, acute illness,
specifically the morbidity and mortality from viral illness attributable to consumption of drinking water
from the PWSs affected by the rule. The quantified acute viral illnesses are those associated with Type A
(represented by data from rotavirus) and Type B (represented by data from enterovirus and echovirus) and
are a subset of the total illnesses, both acute and chronic, from all waterborne bacteria and viruses.
The quantified benefits are presented in two forms. Exhibit 8.1 presents a summary of quantified
benefits in terms of the annual endemic, acute illnesses and deaths avoided after full implementation.
Exhibit 8.2 monetizes estimates of endemic, acute illnesses and deaths avoided into annualized present
values using both the Enhanced and the Traditional COI approaches to allow comparison to cost
estimates. The mean annualized value of quantified benefits of reduced risk ranges from $8.6 million to
$19.7 million, depending on the COI approach and the discount rate used.
There are substantial benefits attributable to the GWR that are not quantified within this EA as
part of the main analyses because of data limitations. Beneficial aspects of the rule not quantified are
characterized as either health benefits or non-health benefits. Nonquantified health-related benefits
include reducing other acute viral illness (other than those caused by rotavirus and enterovirus); endemic,
acute bacterial illnesses and deaths; and epidemic bacterial and viral acute illness and death (associated
with outbreaks, disinfection failures, and distribution system contamination). Chronic illness, both
bacterial and viral, are also not quantified. The rule will also result in many nonhealth benefits such as
reduced costs for responding to outbreaks, costs for averting behavior, and reduced uncertainly regarding
drinking water safety (see Exhibit 8.3).
Economic Analysis for the 8-1 October 2006
Final Ground Water Rule
-------
Exhibit 8.1 Summary of Annual Avoided Viral Illnesses and Deaths
by System Type
System Type
cws
NTNCWS
TNCWS
Total
Annual Viral Illnesses Avoided
Mean
32,031
2,094
7,743
41 ,868
90% Confidence Bounds
5th
8,704
533
1,037
10,274
95th
68,994
4,308
14,738
88,039
Annual Viral Deaths Avoided
Mean
0.62
0.03
0.09
0.74
90% Confidence Bounds
5th
0.07
0.00
0.01
0.08
95th
1.81
0.09
0.21
2.11
Note: Detail may not add due to independent forecasting. Values presented are average annual illnesses and deaths
avoided over the 25 year period of analysis following rule promulgation.
Source: Appendix C
As mentioned above, there are substantial benefits attributable to the GWR that are not quantified
within this EA as part of the main analyses. These nonqualified benefits are shown in relation to
quantified benefits as part of the total benefits of the GWR in Exhibit 8.3. The nonquantified benefits
result from multiple factors. First, the quantified benefits are based on limited, well-defined data and key
assumptions that restrict the input parameters in the quantified benefit calculation. Typically, these
assumptions resulted in low mean values and narrow uncertainty ranges in the benefit analysis. This EA,
where applicable, discusses alternative assumptions. For example, the enterovirus morbidity fractions
are, by assumption, not determined using coxsackievirus (an enterovirus) data although the enterovirus
severity data use all enterovirus data. If coxsackievirus data were available, the mean morbidity values
would be greater. Choosing alternative values and ranges and differing key assumptions, which might
also be deemed reasonable, would increase the quantified benefits in this EA.
Second, the quantified benefits are based on data and assumptions that pertain to only partial
representation of Type A and Type B viruses potentially found in PWS wells with fecal contamination.
Due to limited available data, only rotavirus and some enterovirus data were used to calculate the
quantified benefits. As is more completely discussed in Section 5.4. other viruses as well as pathogenic
bacteria may contribute to the disease burden, both acute and chronic, associated with PWS wells with
fecal contamination. Most importantly, bacterial illnesses can result in more frequent and lengthier
hospitalization and more frequently have fatal outcomes. If bacterial diseases were considered in the
quantified benefits, the monetized benefits could be substantially greater because bacterial disease can be
more severe and can result in higher mortality rates.
Third, the quantified benefits are based on data and assumptions that limit the characterization of
acute disease. For rotavirus, only acute gastroenteritis illness and fatal dehydration associated with that
illness are monetized. Norovirus disease is not considered. For the enteroviruses, all acute disease
endpoints are considered, but the prevalence of severe endemic cases may be substantially diluted by the
large number of hand, foot, and mouth disease cases that are not likely to be waterborne. Thus, the
proportion of severe cases in the quantitative benefits is likely to be underestimated. As is discussed more
completely in Section 5.4, in neither instance, either for rotavirus or the enteroviruses, are chronic
diseases identified or monetized in the quantitative benefits calculation.
Economic Analysis for the
Final Ground Water Rule
8-2
October 2006
-------
Fourth, the quantified benefits are based explicitly on what has been directly measured in PWS
wells, yet there is great difficulty in identifying and counting all infectious viral pathogens in dilute
drinking water samples. Indeed, some viral pathogens like infectious norovirus can never be identified in
any sample. Section 4.3.2 discusses these difficulties in more detail. Standard fecal indicator data such
as total coliforms and E. coll, commonly used to identify water treatment deficiencies and potential
human health hazards, are explicitly not used to determine human exposure for the purposes of
quantifying the benefits in this EA.
Fifth, the quantified benefits are assumed to be based only on one contamination scenario, fecal
contamination of source water. Other contamination scenarios are thoroughly documented in the ground
water contamination and outbreak scientific literature. However, these scenarios, such as inadequate
disinfection, are not explicitly considered in calculating the quantified benefits in this EA.
Sixth, the quantified benefits are assumed to be based only on avoidance of endemic disease. The
GWR will likely also decrease the incidence of epidemic disease (outbreaks). If epidemic illnesses and
the avoided non-health-related costs of ground waterborne disease outbreaks were included, the
quantified benefits would increase.
In summary, this EA quantifies a subset of the total health and non-health related benefits. In a
sample calculation, discussed in Section 5.4.3.2, EPA estimated that the total benefits could increase by a
factor of five by only accounting for additional deaths and hospitalizations caused by bacterial illness
being avoided. While EPA recognizes that this estimate includes substantial uncertainty, given all the
other nonquantified factors described above, EPA believes that the total benefits from the GWR are likely
to be more than five times those which have been quantified. See Exhibit 8.5b for estimates of net
benefits that incorporate nonquantified benefits.
Economic Analysis for the 8-3 October 2006
Final Ground Water Rule
-------
Exhibit 8.2 Summary of Annualized Present Value Quantified Benefits
(SMillions, 2003$)
System Type
Annualized Benefits
at 3% Discount Rate
Mean
90% Confidence Bounds
5th
95th
Annualized Benefits
at 7% Discount Rate
Mean
90%Confidence Bounds
5th
95th
Enhanced COI
cws
NTNCWS
TNCWS
Total
$ 16.0
$ 0.9
$ 2.7
$ 19.7
$ 5.4
$ 0.3
$ 0.8
$ 6.5
$ 37.0
$ 2.2
$ 6.2
$ 45.4
$ 13.7
$ 0.8
$ 2.3
$ 16.8
$ 4.6
$ 0.2
$ 0.7
$ 5.5
$ 31.6
$ 1.8
$ 5.1
$ 38.6
Traditional COI
CWS
NTNCWS
TNCWS
Total
$ 8.2
$ 0.5
$ 1.3
$ 10.0
$ 1.9
$ 0.1
$ 0.3
$ 2.2
$ 22.3
$ 1.3
$ 3.4
$ 27.0
$ 7.1
$ 0.4
$ 1.1
$ 8.6
$ 1.6
$ 0.1
$ 0.2
$ 1.9
$ 19.1
$ 1.0
$ 2.8
$ 22.9
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical costs and
lost work time (including some portion of unpaid household production). The Enhanced COI also factors in valuations for
lost personal time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI),
time with family, and recreation, and lost productivity at work on days when workers are ill but go to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified benefits are
expected to comprise a significant portion of the overall benefits of the Rule and are presented in Section 5.4 of Ch. 5 of the
EA.
Source: Exhibit 5.23a, 5.23b
Economic Analysis for the
Final Ground Water Rule
8-4
October 2006
-------
Exhibit 8.3 Summary of Benefits of the GWR
Benefit Category
Total Benefits
GWR EA Quantified Benefits
Health Benefits
Reduction in
endemic illness
incidence
Reduction in epi-
demic (outbreak)
illness incidence
Reduction in
treatment failures
viral exposure risk reduction (morbidity
and mortality)
bacterial exposure risk reduction
(morbidity and mortality)
chronic sequelae reduction
reduction in secondary transmission of
viral or bacterial illness from
symptomatic and asymptomatic
individuals
viral exposure risk reduction (morbidity
and mortality)
bacterial exposure risk reduction
(morbidity and mortality)
chronic sequelae reduction
reduction in secondary transmission of
viral or bacterial illness from
symptomatic and asymptomatic
individuals
Decreased illness through minimizing
treatment failures or fewer episodes with
inadequate treatment
acute rotavirus (Type A) illnesses and
deaths avoided
acute enterovirus (Type B) illnesses
and deaths avoided
reduction in secondary transmission of
viral illness from symptomatic
individuals
Not quantified
Not quantified
Non-Health Benefits
Outbreak
responses avoided
Avoided costs of
averting behavior
Increased
confidence
Avoided costs to affected water systems,
local governments (provision of alternate
water, issuing warnings and alerts), and
community (decreased tourism due to bad
press).
reduced need or perceived need to
use bottled water, point-of-use
devices, etc. (includes time and
material costs)
less time spent on averting behavior:
hauling/boiling water, etc.
Perceived reduction in risk associated with
perceived improvement in drinking water
quality
Not quantified
Not quantified
Not quantified
Economic Analysis for the
Final Ground Water Rule
8-5
October 2006
-------
8.1.2 National Cost Summary
The national annual cost of the GWR results from activities associated with rule implementation,
sanitary surveys, triggered source water monitoring, corrective actions, and compliance monitoring. The
estimated annualized cost of the GWR is $61.8 million at a three percent discount rate and $62.3 million
at a seven percent discount rate. Exhibit 8.4 presents these costs further broken out by system type.
Exhibit 8.4 Summary of Quantified Costs, Final Rule ($Millions, 2003$)
System Type
cws
NTNCWS
TNCWS
States
Total
Annualized Costs
at 3% Discount Rate
Mean
$ 18.7
$ 4.9
$ 26.4
$ 11.8
$ 61.8
90% Confidence Bounds
5th
$ 12.4
$ 3.3
$ 18.6
$ 10.9
$ 45.2
95th
$ 26.3
$ 6.8
$ 35.7
$ 12.6
$ 81.4
Annualized Costs
at 7% Discount Rate
Mean
$ 19.2
$ 4.9
$ 26.5
$ 11.7
$ 62.3
90%Confidence Bounds
5th
$ 13.0
$ 3.4
$ 18.8
$ 10.9
$ 46.1
95th
$ 26.9
$ 6.7
$ 35.5
$ 12.6
$ 81.6
Notes: Detail may not add due to independent rounding.
Source: Exhibits 6.33 and 6.34
8.1.3 Comparison of National Quantified Benefits and Costs
Exhibit 8.5a compares estimated quantified benefits with estimated costs. Based on the
comparison of these values, the estimated quantified benefits of the rule range from approximately one-
third to about one-seventh of the value of the costs, depending on the discount rate and COI approach.
The estimated quantified benefits for the Enhanced COI approach are greater than the corresponding
estimated benefits for the Traditional COI approach. The quantified estimate of the benefits significantly
understates the true benefit of the rule. As discussed in Section 8.1.1 and Exhibit 8.3, the nonqualified
health and non-health benefits far exceed those that EPA was able to quantify, and are the primary basis
for supporting the preferred regulatory alternative. Section 5.4.3 of the EA discusses the potential value
of nonquantified benefits, in particular avoided bacterial illnesses and deaths, which would significantly
increase the net benefits of the final GWR. Exhibit 8.5b presents these estimates based on primary
benefits using the Enhanced COI approach.
Economic Analysis for the
Final Ground Water Rule
October 2006
-------
Exhibit 8.5a Estimated Annualized National Benefits and Costs for the GWR
(SMillions, 2003$)
Estimate
Category
3% Discount Rate
Mean
90% Confidence Bounds
5th
95th
7% Discount Rate
Mean
90%Confidence Bounds
5th
95th
Enhanced COI
Benefits
Costs
Net Benefits
$ 19.7
$ 61.8
$ (42.1)
$ 6.5
$ 45.2
Note 1
$ 45.4
$ 81.4
Note 1
$ 16.8
$ 62.3
$ (45.5)
$ 5.5
$ 46.1
Note 1
$ 38.6
$ 81.6
Note 1
Traditional COI
Benefits
Costs
Net Benefits
Nonquantified
Benefits
Nonquantified
Costs
$ 10.0
$ 61.8
$ (51.8)
$ 2.2
$ 45.2
Note 1
$ 27.0
$ 81.4
Note 1
$ 8.6
$ 62.3
$ (53.7)
$ 1.9
$ 46.1
Note 1
$ 22.9
$ 81.6
Note 1
Decreased incidence of other acute viral disease endpoints
Decreased incidence of bacterial illness and death
Decreased incidence of chronic bacterial and viral illness sequelae
Decreased incidence of waterborne disease outbreaks and epidemic illness
Decreased illness through minimizing treatment failures or fewer episodes with inadequate treatment
Decreased use of bottle water and point-of-use devices (material costs)
Decreased time spent on averting behavior
Avoided costs associated with outbreak response
Perceived improvement in drinking water quality and reduction in risk associated with ingestion
Benefits from optional Assessment Monitoring
Benefits from correction of sanitary survey deficiencies identified in the distribution systems and treatment
plant
Costs for optional Assessment Monitoring
Costs from correction of sanitary survey deficiencies identified in the distribution systems and
treatment plant
Costs for compliance monitoring for some systems that already disinfect
Some land costs depending on the treatment technology
Cost for five repeat samples but this is small compared to the overestimate of cost for the initial
fecal-indicator sample that systems would take.
Note 1: Because benefits and costs are calculated using different model modules, bounds are not calculated on net benefits.
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical costs and lost
work time (including some portion of unpaid household production). The Enhanced COI also factors in valuations for lost personal
time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI), time with family, and
recreation, and lost productivity at work on days when workers are ill but go to work anyway.
Source: Exhibts 8.2 and 8.4
Economic Analysis for the
Final Ground Water Rule
8-7
October 2006
-------
Exhibit 8.5b Estimated Net Benefits Including Annual Bacterial Illness and Death
Avoidance Estimate ($Millions 2003$)
Estimate Category
Benefits
Quantified Benefits (ECOI)
Bacterial Illness and Death Avoidance Benefits
Total
Costs
Net Benefits
Discount Rate
3%
$
$
$
$
$
19.7
98.6
118.3
61.8
56.5
7%
$
$
$
$
$
16.8
83.8
100.6
62.3
38.3
Notes: Costs and Quantified Benefits are output from the GWR model as annualized estimates that incorporate annual
adjustments for income elasticity, changes in real income, and a CPI increase factor (all affecting the VSL). The GWR model
also incorporates a schedule for implementation that affects the timing of benefits accrual throughout the analysis period for the
GWR, which affects all components of benefits (direct and indirect). The model also discounts both costs and benefits (shown
above at 3 and 7 percent). See Ch. 5 and Appendix B of the EA for detail on these adjustments. The estimated value of
avoiding bacterial illness and death is not generated in the primary analysis model but is calculated as shown in Ch. 5, Sec.
5.4.3. Therefore, the components of the estimated value of avoiding bacterial illness and death (affecting VSL and Direct
Medical Costs) are not adjusted for these factors.
8.2 Effect of Uncertainties and Nonquantified Benefit/Cost Estimates on the Estimation
of Net National Benefits
Detailed discussions of the assumptions and uncertainties associated with national benefits and
costs are contained in Chapters 4, 5, and 6. Several of the most important assumptions and data
uncertainties, and the effect of those uncertainties on the benefits and cost analyses, are discussed below.
The GWR EA attempted to capture the full range of uncertainty in the analysis. In the quantified
analysis, parameters were described using a distribution rather than a point value where sufficient
information is available. Where information is limited, distributions are chosen based on informed
professional judgment. Otherwise, distributions were determined by statistical models that allowed the
data to inform the distribution. Choice of statistical models were based on examining the data combined
with use of scientific principles. Alternative assumptions were tested and described in alternative
analyses. The EA also includes detailed discussion of the additional uncertainty associated with
nonqualified factors and untested assumptions.
The quantified GWR benefits are calculated using Monte Carlo sampling of the parameter
distributions and are expressed in monetary terms. The calculated results are presented as ranges with
means and confidence bounds. The most likely quantified value is within this calculated range but the
total benefit range, considering the non quantified benefits, is greater than the calculated benefit range.
As summarized in Exhibit 5.32 and described throughout Chapter 5, because of limited available data,
almost all of the factors in the quantified analysis were chosen to provide a conservative estimate of
benefit, thus underestimating the quantified benefits. Furthermore, the factors discussed but not quantified
in the benefits analysis primarily act to provide additional GWR benefits, albeit nonqualified. Based on
these two uncertainty sources, EPA concludes that the total mean benefits are expected to be several times
greater than the quantified mean benefits and are likely to be greater than the calculated upper confidence
bound as well. The formal peer review of the GWR EA yielded comments and summary statements that
endorse EPA's conclusion that the GWR EA quantified benefits are small compared to the total benefits.
Economic Analysis for the
Final Ground Water Rule
8-8
October 2006
-------
Most of the significant costs that EPA has identified have been quantified. The only significant
costs that have not been quantified are for certain corrective actions that are a result of significant
deficiencies identified during sanitary surveys. Exclusion of these costs from the EA cost analysis results
in an underestimate of potential rule costs. However, as described in Chapter 6 of the EA, the impact on
the overall cost/benefit ratios from excluding costs for correction of treatment or distribution system
significant deficiencies is minimal since data limitations also exclude quantifying any benefits that may
be realized for correcting these significant deficiencies.
There is uncertainty in the cost analysis that could result in either an over- or underestimate of the
costs as presented in this chapter. Exhibit 6.32 in Chapter 6 of the EA presents a summary of these issues
and estimates the effects that each may have on national costs. The greatest uncertainties affecting the
costs of the GWR are in the percentages used to estimate compliance and costs for each regulatory
alternative. However, by using a Monte-Carlo analysis, a best estimate (mean of range of estimates) of
costs can be calculated as well as the uncertainty bounds around that estimate.
Overall, EPA believes that the nonquantified costs are much smaller than the nonqualified
benefits and the estimates of the national net benefits are conservative. Detailed discussion of
nonquantified benefits and costs are presented in chapters 5 and 6, respectively.
8.3 Breakeven Analysis
In the face of uncertainties, it is helpful to evaluate the range over which the regulation meets the
critical test of benefits exceeding costs. Thus, the number of illnesses or deaths that would have to be
avoided annually to compensate for costs, or for the GWR to "break even," can be calculated. If the
estimated present annual cost of the rule is $61.8 million using a three percent discount rate and $62.3
million using a seven percent discount rate, then 131,234 or 155,609 illnesses, respectively, would need
to be avoided annually using the Enhanced COI approach. Using the Traditional COI approach, 257,887
or 304,696 illnesses would have to be avoided annually. The breakeven number of illnesses that need to
be avoided for both the Enhanced and Traditional COI approaches is above the estimated number of
quantified illnesses (41,868 as shown in Exhibit 8.1) avoided by the final GWR. To "break even" using
deaths avoided as a measure, approximately 8 deaths would have to be avoided for either a three or seven
percent discount rate. Although this number of deaths exceeds the bounds of the quantified number of
deaths avoided in the primary analysis, it is small in absolute terms when considering potential effects of
viruses and bacteria not quantified as part of the analysis. Section 5.4.3.2 describes the methodology for
estimating ten additional deaths due to bacterial illness from contaminated PWS wells. Bacterial illness
and death is not considered in the breakeven analysis.
It must be noted, again, that this analysis does not account for the nonquantified health and non-
health benefits, which far exceed those that EPA was able to quantify, and are the primary basis for
supporting the preferred regulatory alternative. The nonquantified benefits of the final GWR include all
of the nonquantified benefits that accrue to each alternative, as described in Chapter 5.
Economic Analysis for the 8-9 October 2006
Final Ground Water Rule
-------
Exhibit 8.6 Estimated Breakeven Points
Measure
Estimated Number of Cases to Avoid to Break Even1
3% Discount Rate
7% Discount Rate
llnesses
Enhanced COI
Traditional COI
131,234
257,887
155,609
304,696
Deaths
VSL
8.3
8.4
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical costs and
lost work time (including some portion of unpaid household production). The Enhanced COI also factors in valuations for
lost personal time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI),
time with family, and recreation, and lost productivity at work on days when workers are ill but go to work anyway.
1 Breakeven illnesses and deaths are derived by dividing the regulation cost by an estimate of the average benefit per
illness or death avoided as shown in Exhibit 8.1. Illnesses or deaths avoided may be of bacterial or viral etiology; since this
analysis is based on the viral COI described in Ch. 5 of the EA for Type A and Type B viruses, for reasons also explained in
Ch. 5 the COI could be an underestimate for cases of other types of viral or bacterial illness avoided. This could result in
an overestimate of the number of illnesses or deaths that need to be avoided in order to break even with the costs of the
rule.
Source: Derived from Exhibits 8.1 and 8.5a
8.4 Comparison of Regulatory Alternatives
As discussed in Chapter 3, the development and evaluation of several regulatory alternatives was
undertaken as part of a consultation process that included stakeholder meetings and public comments.
The four alternatives in the final EA are listed below.
Alternative 1—Sanitary surveys and corrective action.
Final GWR—Risk Targeted Approach (Sanitary surveys, triggered monitoring, optional
assessment monitoring, corrective action, and compliance monitoring).
Alternative 3—Multi-Barrier Approach (Sanitary surveys, triggered monitoring, optional
hydrogeologic sensitivity assessment, assessment monitoring, corrective action, and compliance
monitoring).
Alternative 4—Across-the-board disinfection (Sanitary surveys, install/upgrade and maintain
treatment).
The following sections evaluate the benefits and costs for the Final GWR requirements in
comparison to the three other alternatives.
Economic Analysis for the 8-10 October 2006
Final Ground Water Rule
-------
8.4.1 Comparison of Benefits and Costs
To make meaningful comparisons between regulatory alternatives, it is first necessary to look at
the final benefit and cost numbers derived for each. Exhibit 8.7 presents the annualized present value
costs for each alternative considered, followed by presentations of benefits in Exhibits 8.8 and 8.9.
Exhibit 8.7 Annualized Costs, by Regulatory Alternative ($Millions, 2003$)
Rule Alternative
Alternative 1
Final Rule
Alternative 3
Alternative 4
Annualized Costs
at 3% Discount Rate
Mean
$ 15.3
$ 61.8
$ 67.9
$ 686.4
90% Confidence Bounds
5th
$ 11.8
$ 45.2
$ 49.4
$ 636.8
95th
$ 19.2
$ 81.4
$ 89.5
$ 735.4
Annualized Costs
at 7% Discount Rate
Mean
$ 15.3
$ 62.3
$ 69.4
$ 665.3
90% Confidence Bounds
5th
$ 11.9
$ 46.1
$ 51.0
$ 612.3
95th
$ 19.0
$ 81.6
$ 90.6
$ 717.0
Source: Exhibit 6.35
Exhibit 8.8 Number of Annual Quantified Viral Illnesses and Deaths Avoided
Regulatory Alternatives
Rule Alternative
Alternative 1
Final Rule
Alternative 3
Alternative 4
Illnesses Avoided
Mean
7,497
41,868
45,419
155,282
90% Confidence Bounds
5th
1,618
10,274
11,639
27,824
95th
17,007
88,039
95,166
399,085
Deaths Avoided
Mean
0.14
0.74
0.80
2.67
90% Confidence Bounds
5th
0.01
0.08
0.09
0.21
95th
0.44
2.11
2.33
9.25
Source: Exhibit 5.30
Economic Analysis for the
Final Ground Water Rule
8-11
October 2006
-------
Exhibit 8.9 Annualized Value of Quantified Viral Illnesses and Deaths Avoided,
by Regulatory Alternative ($Millions, 2003$)
Rule Alternative
At 3%
Mean
90% Confidence Bounds
5th
95th
At 7%
Mean
90% Confidence Bounds
5th
95th
Enhanced COI
Alternative 1
Final Rule
Alternative 3
Alternative 4
$ 3.6
$ 19.7
$ 21.3
$ 70.2
$ 0.9
$ 6.5
$ 7.1
$ 18.3
$ 9.3
$ 45.4
$ 48.7
$ 177.0
$ 2.9
$ 16.8
$ 18.2
$ 61.9
$ 0.7
$ 5.5
$ 6.0
$ 16.1
$ 7.5
$ 38.6
$ 41.6
$ 156.3
Traditional COI
Alternative 1
Final Rule
Alternative 3
Alternative 4
$ 1.9
$ 10.0
$ 10.8
$ 35.5
$ 0.3
$ 2.2
$ 2.5
$ 6.5
$ 5.5
$ 27.0
$ 28.9
$ 102.4
$ 1.5
$ 8.6
$ 9.3
$ 31.5
$ 0.2
$ 1.9
$ 2.1
$ 5.7
$ 4.5
$ 22.9
$ 24.8
$ 90.8
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical costs and lost
work time (including some portion of unpaid household production). The Enhanced COI also factors in valuations for lost
personal time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI), time with
family, and recreation, and lost productivity at work on days when workers are ill but go to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unqualified benefits are expected
to compose a significant portion of the overall benefits of the Rule and are presented in Section 5.4 of Ch. 5 of the EA.
Source: Exhibit 5.31
Net benefit is the difference between the monetized benefit and cost estimates. Exhibit 8. lOa
presents net benefits based on the annualized present value of quantified benefits at three and seven
percent discount rates, using both the Enhanced and the Traditional COI approach. The data are based on
the average benefits less the average values for costs. As discussed in Section 8.1.1 and Exhibit 8.3, the
nonqualified health and nonhealth benefits far exceed those that EPA was able to quantify, and their
incorporation into the quantified analysis would result in substantial increases in net benefits. Exhibit
8.1 Ob demonstrates how by using a multiplier (based on bacterial hospitalizations and deaths avoided),
the net benefits of the final GWR are positive using the enhanced COI and are very close to positive using
the traditional COI. Based on consideration of the large number of nonqualified benefits, EPA believes
that this level of additional nonquantified benefits in relation to quantified benefits will be easily
achieved. Therefore, consideration of these nonquantified benefits provides a basis for supporting the
preferred regulatory alternative. To the extent that additional benefits are realized, those populations at
greatest risk from ground water contamination (e.g., sensitive sub-populations, including children and the
immunocompromised) will be better protected.
Economic Analysis for the
Final Ground Water Rule
8-12
October 2006
-------
Exhibit 8.1 Oa Annualized Net Benefits by Regulatory Alternative
($Millions, 2003$)
Rule Alternative
Annualized Value
3%
7%
Enhanced COI
Alternative 1
Final Rule
Alternative 3
Alternative 4
$
$
$
$
(11.7)
(42.1)
(46.6)
(616.2)
$
$
$
$
(12.4)
(45.5)
(51.2)
(603.4)
Traditional COI
Alternative 1
Final Rule
Alternative 3
Alternative 4
$
$
$
$
(13.5)
(51.8)
(57.1)
(650.9)
$
$
$
$
(13.8)
(53.7)
(60.1)
(633.8)
Notes: Detail may not add due to independent rounding. The Traditional COI
only includes valuation for medical costs and lost work time (including some
portion of unpaid household production). The Enhanced COI also factors in
valuations for lost personal time (non-work time) such as childcare and
homemaking (to the extent not covered by the Traditional COI), time with family,
and recreation, and lost productivity at work on days when workers are ill but go
to work anyway.
The figures presented in this exhibit represent only the quantifiable benefits of
the GWR. The unquantified benefits are expected to compose a significant
portion of the overall benefits of the Rule and are presented in Section 5.4 of Ch.
5 of the EA.
Source: Derived from Exhibits 8.7 and 8.9
Exhibit 8.1 Ob Annualized Mean Net Benefits for Final Rule Including Estimates
for Nonquantified Benefits ($Millions, 2003$)
Multiple of Benefits Representing
Nonquantified Benefits
Annualized Net Benefit Value -
Final Rule
3%
7%
Enhanced COI
Quantified Net Benefits Only
Nonquantified = 5X Quantified
$
$
(42.1)
56.5
$
(45.5)
$ 38.3
Traditional COI
Quantified Net Benefits Only
Nonquantified = 5X Quantified
$
$
(51.8)
(1.6)
$
$
(53.7)
(10.9)
Notes: Detail may not add due to independent rounding. The Traditional COI only includes
valuation for medical costs and lost work time (including some portion of unpaid household
production). The Enhanced COI also factors in valuations for lost personal time (non-work time)
such as childcare and homemaking (to the extent not covered by the Traditional COI), time with
family, and recreation, and lost productivity at work on days when workers are ill but go to work
anyway.
Source: Derived from Exhibits 8.7 and 8.9
Economic Analysis for the
Final Ground Water Rule
8-13
October 2006
-------
8.4.2 Cost-Effectiveness Measures
Cost-Effectiveness-Traditional Approach
Cost-effectiveness analysis is a policy evaluation tool that allows comparisons of regulatory
alternatives. The concept of cost-effectiveness can be defined simply as getting the greatest benefits for a
given expenditure or imposing the least cost for a given level of benefits. In Exhibits 8.1 la and 8.1 Ib, the
test is to see if any alternative falls to the right and completely below any other alternative on the graph.
If so, the alternative to the right and below would be more cost-effective and "dominate" the alternative
that provided fewer benefits at higher costs.
In the strict sense, each of the regulatory alternatives is cost effective—no regulatory alternative
provides more benefits at the same or a lower cost than another, and no alternative can achieve lower
costs for the same or a greater level of benefits than another. Thus, no alternative dominates any other or
is more cost effective. Instead, the alternatives offer increasing levels of benefits at increasing levels of
cost, as seen in Exhibits 8.1 la and 8.1 Ib.
Economic Analysis for the 8-14 October 2006
Final Ground Water Rule
-------
Exhibit 8.11 a Mean Annualized Costs at Mean Benefit Level, Enhanced COI,
by Regulatory Alternative
ur
0
i
i/>
0
o
3 Percent Discount Rate
$800.0 -,
tvnn n
tfinn n
$4000 -
<[Onn ,-.
CpoUU.U
Q*onn 0
'Rinn n
Alt 4
S*
/S
^^
jX^
^r
Final x^
Rule ^^r
$- $25.0 $50.0 $75.0 $100.0
Benefits ($Millions)
Source: Exhibits 8.7 and 8.9
v>
c.
o
g
«»
in
in
o
O
7 Percent Discount Rate
$700.0 -,
$500.0 -
$400.0 -
---. .
-poUU.U
$200.0 -
«-inn n -
$- -
$-
* Alt 4
/
f
/
Final /
K^^ĄfK^
$25.0 $50.0 $75.0 $100.0
Benefits ($Millions)
Source: Exhibits 8.7 and 8.9
Notes: Detail may not add due to independent rounding. The Traditional COI only includes valuation for medical
costs and lost work time (including some portion of unpaid household production). The Enhanced COI also
factors in valuations for lost personal time (non-work time) such as childcare and homemaking (to the extent not
covered by the Traditional COI), time with family, and recreation, and lost productivity at work on days when
workers are ill but go to work anyway.
Economic Analysis for the
Final Ground Water Rule
8-15
October 2006
-------
Exhibit 8.11 b Mean Annualized Costs at Mean Benefit Level, Traditional COI,
by Regulatory Alternative
(A
C
O
i
<&
•4-i
(A
O
O
3 Percent Discount Rate
$800.0 n
^
^r
Alt 1 Illl'-Sr
^--^^ /\lt 3
i i i i i
$- $10.0 $20.0 $30.0 $40.0 $50.0
Benefits ($Millions)
Source: Exhibits 8.7 and 8.9
Note: The Traditional COI only includes valuation for medical costs and lost work time (including some
portion of unpaid household production). The Enhanced COI also factors in valuations for lost personal
time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional
COI), time with family, and recreation, and lost productivity at work on days when workers are ill but go to
work anyway.
Economic Analysis for the
Final Ground Water Rule
8-16
October 2006
-------
Cost per Case of Illness or Death Avoided
Another measure related to cost-effectiveness that EPA used to evaluate the regulatory
alternatives is the cost for each case of illness and death avoided. EPA has performed this analysis for the
quantified benefits of the GWR. For purposes of evaluating the alternatives, the lower the cost per case
or death avoided, the more cost-effective the alternative is believed to be. Exhibit 8.12 presents the
average cost per case of viral illness and death avoided for each regulatory alternative.
Exhibit 8.12 Cost Per Viral Illness or Death Avoided
by Regulatory Alternative (2003$)
Rule Alternative
Alternative 1
Final Rule
Alternative 3
Alternative 4
Cost per Viral Illness Avoided
($)
3%
$ 2,045
$ 1 ,476
$ 1 ,495
$ 4,420
7%
$ 2,044
$ 1 ,488
$ 1 ,527
$ 4,284
Cost per Viral Death Avoided
($Millions)
3%
$ 107.4
$ 83.1
$ 85.0
$ 257.4
7%
$ 107.4
$ 83.8
$ 86.9
$ 249.5
Note: The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unqualified benefits are
expected to compose a significant portion of the overall benefits of the Rule and are presented in Section 5.4 of Ch. 5 of the EA.
Source: Derived from Exhibits 8.7 and 8.8
Exhibit 8.12 shows that the cost per case avoided for the final GWR is lower than the cost per
case avoided for each of the other alternatives considered. Thus, the Final GWR is the most efficient
alternative by this measure. Given the substantial nonqualified benefits of the final GWR as described
in Chapter 5, the actual cost per case is lower than calculated in Exhibit 8.12.
Incremental Net Benefits
EPA also assessed the incremental net benefits of the regulatory alternatives. Incremental costs
and benefits are those that are incurred or realized in reducing viral illness, bacterial illness, and outbreaks
from one rule alternative to the next. Estimates of incremental costs and benefits are useful in considering
the economic efficiency of different regulatory options considered by the Agency. Generally, the goal of
an incremental analysis is to identify the option where incremental benefits most closely equal
incremental costs (net social benefits are maximized). However, the usefulness of this analysis is limited
because the benefits from the rule that are not quantified or monetized far exceed those that EPA was able
to quantify, and these nonquantified benefits are the primary basis for supporting the preferred regulatory
alternative.
Exhibit 8.13 presents the four regulatory alternatives in order of increasing level of reduction in
waterborne pathogens, or increasing levels of protection from illness. As a result, it is possible to
compare incremental net benefits from the baseline and alternative to alternative. As shown in Exhibits
8.13a and b, incremental net benefits for all alternatives are negative. The benefits of non-quantified
bacterial illness and death avoided would add benefits to all alternatives without any increase in costs.
Economic Analysis for the
Final Ground Water Rule
8-17
October 2006
-------
EPA estimated that the total benefits are likely to increase by more than a factor of five by accounting for
additional deaths and hospitalizations caused by reduced bacterial illness and death alone. These
non-quantified benefits have a significant positive impact on the incremental benefits and net incremental
benefits. Both Alternative 3 and Alternative 2 would have positive net incremental benefits if the
bacterial benefits are considered. The next highest alternative, Alternative 4, has such highly negative
incremental net benefits, and the difference is so substantial that non-monetized benefits would be
unlikely to compensate. However, comparisons between Alternative 4 and the other alternatives may be
between two separate sets of benefits, in the sense that they may be distributed to somewhat different
populations. However, based on consideration of all factors, EPA has determined that the final GWR
provides the maximum benefits at a cost that is justified.
The cost-effectiveness of the final GWR and alternatives was also considered in terms of the
quality-adjusted life years (QALYs) saved for avoided viral illnesses or deaths. The QALYs analysis
shows that the final GWR performs the best in terms of this measure. Appendix H provides the detail of
this analysis.
Exhibit 8.13a Incremental Net Quantified Benefits by Rule Alternative -
Enhanced COI (Annualized Present Value Mean, $Millions, 2003$)
Rule Alternative
Annual
Costs
A
Annual
Benefits
B
Incremental
Costs
C
Incremental
Benefits
D
Incremental
Net Benefits
E=D-C
3% Discount Rate
Alternative 1 : Sanitary Survey and Corrective Action
Alternative 2 (Final Rule): Risk Targeted Approach
Alternative 3: Multi-Barrier Approach
Alternative 4: Across-the-board Disinfection
$ 15.3
$ 61.8
$ 67.9
$ 686.4
$ 3.6
$ 19.7
$ 21.3
$ 70.2
$ 15.3
$ 46.5
$ 6.1
$ 618.5
$ 3.6
$ 16.1
$ 1.6
$ 48.9
$ (1 1 .7)
$ (30.4)
$ (4.5)
$ (569.6)
7% Discount Rate
Alternative 1 : Sanitary Survey and Corrective Action
Alternative 2 (Final Rule): Risk Targeted Approach
Alternative 3: Multi-Barrier Approach
Alternative 4: Across-the-board Disinfection
$ 15.3
$ 62.3
$ 69.4
$ 665.3
$ 2.9
$ 16.8
$ 18.2
$ 61.9
$ 15.3
$ 47.0
$ 7.1
$ 595.9
$ 2.9
$ 13.9
$ 1.4
$ 43.8
$ (12.4)
$ (33.1)
$ (5.7)
$ (552.2)
Notes: The Traditional COI only includes valuation for medical costs and lost work time (including some portion of unpaid household production). The Enhanced COI
also factors in valuations for lost personal time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI), time with family,
and recreation, and lost productivity at work on days when workers are ill but go to work anyway. All values are annualized in 2003$.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified benefits are expected to compose a significant portion of the
overall benefits of the Rule and are presented in Section 5.4 of Ch. 5 of the EA.
Sources: (A) Exhibit 8.7, (B) Exhibit 8.9
Economic Analysis for the
Final Ground Water Rule
8-18
October 2006
-------
Exhibit 8.13b Incremental Net Quantified Benefits by Rule Alternative -
Traditional COI (Annualized Present Value Mean, $Millions, 2003$)
Rule Alternative
Annual
Costs
A
Annual
Benefits
B
Incremental
Costs
C
Incremental
Benefits
D
Incremental
Net Benefits
E=D-C
3% Discount Rate
Alternative 1 : Sanitary Survey and Corrective Action
Alternative 2 (Final Rule): Risk Targeted Approach
Alternative 3: Multi-Barrier Approach
Alternative 4: Across-the-board Disinfection
$ 15.3
$ 61.8
$ 67.9
$ 686.4
$ 1.9
$ 10.0
$ 10.8
$ 35.5
$ 15.3
$ 46.5
$ 6.1
$ 618.5
$ 1.9
$ 8.2
$ 0.8
$ 24.7
$ (13.5)
$ (38.3)
$ (5.3)
$ (593.8)
7% Discount Rate
Alternative 1 : Sanitary Survey and Corrective Action
Alternative 2 (Final Rule): Risk Targeted Approach
Alternative 3: Multi-Barrier Approach
Alternative 4: Across-the-board Disinfection
$ 15.3
$ 62.3
$ 69.4
$ 665.3
$ 1.5
$ 8.6
$ 9.3
$ 31.5
$ 15.3
$ 47.0
$ 7.1
$ 595.9
$ 1.5
$ 7.1
$ 0.7
$ 22.2
$ (13.8)
$ (39.9)
$ (6.4)
$ (573.7)
Note: The Traditional COI only includes valuation for medical costs and lost work time (including some portion of unpaid household production). The Enhanced COI
also factors in valuations for lost personal time (non-work time) such as childcare and homemaking (to the extent not covered by the Traditional COI), time with family,
and recreation, and lost productivity at work on days when workers are ill but go to work anyway. All values are annualized in 2003$.
The figures presented in this exhibit represent only the quantifiable benefits of the GWR. The unquantified benefits are expected to compose a significant portion of the
overall benefits of the Rule and are presented in Section 5.4 of Ch. 5 of the EA.
Sources: (A) Exhibit 8.7, (B) Exhibit 8.9
8.5 Summary of Conclusions
The Agency also performed a number of other analyses related to the final rule. This process
included an analysis of net benefits, as well as cost effectiveness and efficiency analyses. In addition, the
Agency performed a number of comparisons among the four regulatory alternatives that are described in
more detail earlier in this chapter. The following is a summary of these analyses.
The GWR likely passes economic threshold criteria:
The GWR has positive net benefits when both quantified and nonqualified benefits are
considered. For the Enhanced COI approach, the quantified benefits alone are approximately
27 to 32 percent of the costs of the GWR (Exhibit 8.5a) depending on discount rate. For the
Traditional COI approach, the quantified benefits are approximately 14 to 16 percent of the
costs of the GWR. Considering that nonqualified benefits are expected to be significantly
larger than the quantified benefits, it appears likely that the final GWR would have positive
net benefits regardless of the discount rate or cost of illness approach used. Section 5.4.3.2
presents a discussion of nonquantified benefits and estimates a portion of their value, based
only on bacterial illnesses avoided, at four times the primary analysis benefits (resulting in
Economic Analysis for the
Final Ground Water Rule
8-19
October 2006
-------
total benefits that are five times the primary benefits). This includes consideration of the
value of deaths and hospitalization costs avoided for ground water borne cases of bacterial
illness prevented by the rule. Including only these estimated bacterial illness and death
benefits, the total net benefits of the GWR would be positive using the Enhanced COI
approach. Total net benefits would still be slightly negative using the Traditional COI
approach, however, other nonquantified benefits such as indirect (non-medical) costs
associated with waterborne bacterial illness or the value of avoiding other chronic (either
bacterial or viral) or other viral illnesses (not accounted for in this analysis) would most
likely make this value positive.
• The number of illnesses that must be avoided to break even with costs (Exhibit 8.6) is well
above the estimated number of viral cases avoided (Exhibit 8.1), but is most likely within the
bounds of cases avoided once nonquantified cases (both bacterial and viral) are considered.
The number of deaths that must be avoided to break even, while outside the bounds of the
quantitative analysis, is small in absolute terms. Consideration of all nonquantified benefits
is predicted to result in favorable break even results.
The GWR is cost-effective (using either the Enhanced or the Traditional COI approach): no
other alternative achieves greater benefits at the same cost or the same benefits at lower cost
(Exhibit 8.11).
Final GWR determinations:
• The economic analysis for this rule, considering quantified and nonquantified benefits,
supports the basis for selecting the final GWR over other alternatives. However, the
distinction between Alternative 2 and 3 on an economic basis, is not great.
• EPA chose the final GWR because EPA believes it is more flexible, targeted, and cost-
effectively protective than Alternative 3. Optional assessment monitoring allows States to
most effectively target those systems at greatest risk and minimize unnecessary monitoring.
EPA took the following considerations into account in making this judgment:
1) Under Alternative 3, some States may not be able to conduct HSAs and thereby require
systems in nonsensitive aquifers to conduct assessment monitoring unnecessarily. For
systems not at risk this additional monitoring would provide no benefit.
2) Systems with frequent TC positives in the distributions system (and subsequent frequent
triggered monitoring) would benefit little from assessment monitoring regardless if they
were sensitive or not because the source water would already be thoroughly evaluated.
Under Alternative 3, such systems in sensitive (or undetermined) aquifers would be
required to do assessment monitoring.
3) Systems identified as having significant risk factors pertaining to potential fecal
contamination at their source (e.g., aquifer condition, well characteristics, proximity to
sewage or septic), but infrequent triggered monitoring source water samples, would
benefit from assessment monitoring. States will be able to identify such systems on an
ongoing basis through a variety of tools and information readily available to them.
Economic Analysis for the 8-20 October 2006
Final Ground Water Rule
-------
• The EPA believes that the final rule is a logical outgrowth of the proposed rule, that it is
supported by comments, and that it provides public health benefits while apportioning costs
in a more flexible targeted manner.
As a result of all of these considerations, EPA has determined that the final GWR will provide
important protection against illnesses and deaths attributable to fecally contaminated ground water. EPA
also believes that the GWR will provide a desired level of protection from ground water pathogen
contamination at a justifiable cost.
Economic Analysis for the 8-21 October 2006
Final Ground Water Rule
-------
References
Abbaszadegan, M., P.W. Stewart, M.W. LeChevallier, J.S. Rosen, and C.P. Gerba. 1998/1999.
Occurrence of viruses in ground water in the United States. AWWA Research Foundation:
Denver, CO, 157 p.
Abbaszadegan, M., P.W. Stewart, and M.W. LeChevallier. 1999. A strategy for detection of viruses in
groundwater by PCR. Applied and Environmental Microbiology. 65 (2): 444-449.
Abbaszadegan, M., M. Denhart, M. Spinner, G. Di Giovanni, M. LeChavallier. 1999c. Identification of
viruses present in ground water cell culture harvest by PCR. Proceedings, Water Quality
Technology Conference, AWWA: Denver, CO.
Abbaszadegan, M. 2002. Viruses in drinking water and ground water. In: Encyclopedia of
environmental microbiology, Ed. Bitton. p.3288-3300. New York, NY: G. John Wiley & Sons.
Abbaszadegan, M., M. LeChevallier, and C. Gerba. 2003. Occurrence of viruses in US groundwaters.
Journal of the American Water Works Association. 95(9): 107-120.
Abdalla, C.W. 1990. Measuring economic losses from ground water contamination: An investigation of
household avoidance costs. Water Resources Bulletin. 26:451-63.
Abdalla, C.W. et al. 1992. Valuing environmental quality changes using averting expenditures: An
application to groundwater contamination. Land Economics. 68(2): 163-9.
Altekurse, S. F., N.J. Stern, P.I. Fields, and D.L. Swerdlow. 1999. Camplybacter jejuni - An emerging
foodborne pathogen. Emerging Infectious Diseases. 5(10): 1-10.
American Water Works Association (AWWA). 1998. Ground water system survey.
Anderson, A.D., A.G. Heryford, J.P. Sarisky, C. Higgins, S.S. Monroe, RS. Beard, C.M. Newport, J.L.
Cashdollar, G.S. Fout, D.E. Robbins, S.A. Seys, K.J. Musgrave, C. Medus, Jan Vinje, J.S. Bresee,
H.M. Mainzer, and R.I. Glass. 2003. A waterborne outbreak of Norwalk-like virus among
snowmobilers-Wyoming, 2001. Journal of Infectious Diseases. 187:303-306.
Anderson, E.J., and S.G. Weber. 2004. Rotavirus infection in adults. Lancet Infectious Diseases. 4(2):91-
9.
Anderson, R.M., and R. May. 1991. Infectious diseases of humans: dynamics and control. New York:
Oxford University Press.
Anderson, R.M. and R.M. May. 1982. Directly transmitted infectious-diseases - control by vaccincation.
Science. 215(4536): 1053-1060.
Ando, T., J.S. Noel, and R.L. Fankhauser. 2000. Genetic calssification of 'Norwalk-like viruses'.
Journal of Infectious Diseases. 181:8336-8348.
Economic Analysis for the R-l October 2006
Final Ground Water Rule
-------
Angulo, F.J., S. Tippen, D. Sharp, B.J. Payne, C. Collier, J. Hill, T.J. Barrett, R.M. Clark, E. Geldreich,
H.D. Donnell, and D.L. Swerdlow. 1997. A community waterborne outbreak of salmonellosis
and the effectiveness of a boil water order. American Journal of Public Health. 87(4):580-584.
Association of State Drinking Water Administrators (ASDWA). 1997. Survey of best management
practices for community ground water systems. ASDWA: Washington DC, December.
ASDWA. 2001. Drinking water program resource needs assessment. Version 9. November 27, 2001.
Atherholt, T., E. Feerst, B. Hovendon, J. Kwak, J. and D. Rosen. 2003. Evaluation of indicators of fecal
contamination in groundwater. Journal of the American Water Works Association. 95 (10): 119-
131.
Banks, W.S.L., C.A. Klohe, D.A. Battigelli. 2001. Occurrence and distribution of enteric viruses in
shallow ground water and factors affecting well vulnerability to microbiological contamination in
Worcester and Wicomico Counties, Maryland. U.S. Geological Survey Water-Resources
Investigations Report. 01-4147.
Banks, W.S.L., and D.A. Battigelli. 2002. Occurrence and distribution of microbiological contamination
and enteric viruses in shallow ground water in Baltimore and Harford Counties, Maryland. U.S.
Geological Survey Water-Resources Investigations Report. 01-4216. 32 p.
Banyai, K., M. Angyal, E. Kormendi, F. Lakatos, M. Uj, and G. Szucs. 2002. Outbreak of human
rotavirus infection in an adult community. Orv Hetil. 143(22): 1347-52.
Barron-Romero, B.L., J. Barreda-Gonzalez, R. Doval-Ugalde, J. Zermeno-Eguia Liz, and M. Huerta-
Pena. 1985. Asymptomatic rotavirus infections in day care centers. Journal of Clinical
Microbiology. 22(1): 116-8.
Bartlett, A.V. et al. 1988. Rotavirus in infant-toddler day care centers: epidemiology relevant to disease
control strategies. Journal of Pediatrics. 113:435-441.
Battigelli, D.A. 1999. Monitoring ground waters in Wisconsin, Minnesota, and Maryland for enteric
viruses and candidate viral indicators. Unpublished report. February 23, 1999.
Barwick, R.S., D.A. Levy, G.F. Craun, M.J. Beach, and R.L. Calderon. 2000. Surveillance for
waterborne-disease outbreaks - United States, 1997-1998. Morbidity and Mortality Weekly
Report. 49(SS04):l-35.
Barza, M. 2002. Potential mechanisms of increased disease in humans from antimicrobial resistance in
food animals. In: The need to improve antimicrobial use in agriculture: Ecological and human
health consequences. M. Barza and S.L. Gorbach (eds). Clinical Infectious Diseases.
34(S3):S123-S130.
Beller, M., A. Ellis, S.H. Lee, M.A. Drebot, S.A. Jenkerson, E. Funk, M.D. Sobsey, O.D. Simmons III,
S.S. Monroe, T. Ando, J. Noel, M. Petric, J.P. Middaugh, and J.S. Spika. 1997. Outbreak of
viral gastroenteritis due to a contaminated well. Journal of the American Medical Association.
278:563-568.
Economic Analysis for the R-2 October 2006
Final Ground Water Rule
-------
Bennett, J.V. et al. 1987. Infectious and parasitic diseases. In: Amler, Closing the Gap: The Burden of
Unnecessary Illness, R.W. and H.B. Dull, (eds.). p. 102-14. New York, NY: Oxford Press.
Benson, V. and M.A. Marano. 1998. Current estimates of the national health interview survey, 1995.
National Center for Health Statistics. Vital Health Statistics. 10(199).
Bernstein, D.I., D.S. Sander, V.E. Smith, G.M. Schiff, and R.L. Ward. 1991. Protection from rotavirus
reinfection: 2-year prospective study. Journal of Infectious Diseases. 164(N2):277-283.
Bishop, R. 1996. Natural history of human rotavirus infection. Archive of Virology. 12[Suppl]: 119-128
Bishop, R.F., G.L. Barnes, E. Cipriani, and J.S. Lund. 1983. Clinical immunity after neonatal rotavirus
infection: a prospective longitudinal study in young children. The New England Journal of
Medicine. 309(2):72-76.
Bitton, G., and C.P. Gerba. 1984. Ground water pollution microbiology. New York, NY: John Wiley &
Sons.
Blanton, L.H., S.M. Adams, M.S. Beard, G. Wei, S.N. Bulens, M. Widdowson, R.I. Glass, and S.S.
Monroe. 2006. Molecular and epidemiologic trends of caliciviruses associated with outbreaks of
acute gastroenteritis in the United States, 2000-2004. Journal of Infectious Diseases.
193:413-421.
Borchardt, M.A., N.L. Haas, and R.J. Hunt. 2004. Vulnerability of drinking-water wells in La Crosse,
Wisconsin, to enteric-virus contamination from surface water contributions. Applied and
Environmental Microbiology. 70(10): 5 93 7-5 946.
Boring, J.R., W.T. Martin, and L.M. Elliott. 1971. Isolation of salmonella typhimurium from municipal
water, Riverside, California, 1965. American Journal of Epidemiology. 93:49-54.
Bureau of Labor Statistics (BLS). 2003. Consumer price index: all urban consumers, medical care
services (not seasonally adjusted); labor force statistics from the current population survey.
http://www.bls.gov. Data extracted July 8, 2004.
BLS. 2004. Employment cost index for 2003, not seasonally adjusted, wages and salaries, state and local
government, http://www.bls.com. Data extracted July 8, 2004.
Bowen, G.S. and M.A. McCarthy. 1983. Hepatitis A associated with a hardware store water fountain
and a contaminated well in Lancaster County, Pennsylvania, 1980. American Journal of
Epidemiology. 117:695-705.
Brookhart, M.A., A.E. Hubbard, M.J. van der Laan, J.M. Colford, and J.N.S. Eisenberg. 2002. Statistical
estimation of parameters in a disease transmission model: analysis of a Cryptosporidium
outbreak. Statistics in Medicine. 21(23):3627-3638.
Economic Analysis for the R-3 October 2006
Final Ground Water Rule
-------
Cannon, R.O., J.R. Poliner, R.B. Hirschorn, B.C. Rodeheaver, P.R. Silverman, E.A. Brown, G.H. Talbot,
S.E. Stine, S. Monroe, D.T. Dennis, and R.I. Glass. 1991. A multistate outbreak of 'Norwalk
virus' gastroenteritis associated with consumption of commercial ice. Journal of Infectious
Diseases. 164:860-863.
Canter L. and R.C. Knox. 1984. Evaluation of septic tank system effects on ground water. USEPA:
Washington B.C.
Canter, L.W., R.C. Knox, and D.M. Fairchild. 1987. Ground water quality protection. Chelsea, MI:
Lewis Publishers, Inc.
Carabin, H., T.W. Gyorkos, J.C. Soto, J. Penrod, L. Joseph, and J.P. Collet. 1999. Estimation of direct
and indirect costs because of common infections in toddlers attending day care centers.
Pediatrics. 103(3):556-564.
Carducci A., L. Caniania, R. Moscatelli, B. Casini, E. Rovini, F. Mazzoni, A. Giuninti, and M. Verani.
2002. Interference between enterovirus and reovirus as a limiting factor in environmental virus
detection. Letters in Applied Microbiology. 34:110-113.
Carter, M.J. 2005. A review: Enterically infecting viruses: pathogenicity, transmission and significance
for food and waterborne infection. Journal of Applied Microbiology. 98:1354-1380.
Centers for Disease Control and Prevention (CDC). 1993. Surveillance for waterborne disease outbreaks
- United States, 1991-1992. Morbidity and Mortality Weekly Report. 42(SS-05):l-22.
CDC. 200la. Norwalk-like viruses, public health consequences and outbreak management. Morbidity
and Mortality Weekly Report. 50(RR-9):18.
CDC. 200Ib. Two fatal cases of adenovirus-related illness in previously healthy young adults - Illinois,
2000. Morbidity and Mortality Weekly Report. 50(26):553-555.
CDC. 2002. Human Immunodeficiency Virus (HIV)/AIDS surveillance report: U.S. HIV and AIDS
cases reported through December 2001. Year-end edition 13(2):table 1.
CDC. 1996. Shigella sonnei outbreak associated with contaminated drinking water - Island Park, Idaho,
August 1995. Morbidity and Mortality Weekly Report. 45(11):229-231.
CDC. 1988. Viral gastroenteritis - South Dakota and New Mexico. Morbidity and Mortality Weekly
Report. 37(5):68-71.
Champsaur, H., E. Questiaux, J. Prevot, M. Henry-Amar, D. Goldszmidt, M. Bourjouane, and C. Bach.
1984. Rotavirus carriage, asymptomatic infection and disease in the first two years of life. I.
virus shedding. Journal of Infectious Diseases. 149(5):667-674.
Chanock, R. M., R.G. Wyatt, and A.Z. Kapikian. 1978. Immunization of infants and young children
against rotaviral gastroenteritis - prospects and Problems. Journal of the American Vetrinary
Medical Association. 173(5):570-572.
Economic Analysis for the R-4 October 2006
Final Ground Water Rule
-------
Chang, L., K. Tsao, S. Hsia, S. Shih, C. Huang, W. Chan, K. Hsu, T. Fang, Y. Huang, and L. Lin. 2004.
Transmission and clinical features of Enterovirus 71 infections in household contacts in Taiwan.
Journal of the American Medical Association. 291(2): 222-227.
Chatterjee, N.K., D.W. Moore, S.S. Monroe, R.I. Glass, M.J. Cambridge, S.F. Kondracki, and D.L.
Morse. 2004. Molecular epidemiology of viral gastroenteritis in New York State, 1998-1999.
Clinical Infectious Diseases. 38(Suppl 3):S303-310.
Cherry, J. D. 1995. Enteroviruses. In: Infectious diseases of the fetus and newborn infant, 4th ed.
Remington and Klein, (eds.). p.404-446. Philadelphia, PA: WB Saunders Company.
Chiba, S., S. Nakata, S. Ukae, andN. Adachi. 1993. Virological and serological aspects of immune
resistance to rotavirus gastroenteritis. Clinical Infectious Diseases. 16(Suppl.l2): S117-S121.
demons, J., M. Rao, F. Ahmed, R. Ward, S. Huda, J. Chakraborty, M. Yunus, M.R. Khan, B. Kay, F. van
Loon, and D. Sack. 1993. Breast-feeding and the risk of life-threatening rotavirus diarrhea:
prevention or postponement?. Pediatrics. 92(5):680-685.
Collins, J.G. 1997. Prevalence of selected chronic conditions: United States, 1990-1992. Hyattsville,
MD: National Center for Health Statistics. DHHS Publication. Vital Health Statistics.
[PHS]10(194):97-1522.
Consultants in Environmental and Occupational Health (CEOH). 1998. Draft report on GWDR Risk
Assumptions, unpublished report to SAIC, July 27, 1998.
Couch, R.B., R.G. Douglas, K.M. Lindgren, P.J. Gerone, and V. Knight. 1970. Airborne transmission of
respiratory infection with coxsackievirus A type 21. American Journal of Epidemiology. 91:78-
86.
Craun, G.F. and R. Calderon. 2001. Waterborne disease outbreaks caused by distribution system
deficiencies. Journal of the American Water Works Association. 93(7):64-75.
Cropper, M.L. and A.J. Krupnick. 1990. The social costs of chronic heart and lung disease. Resources
for the Future Discussion Paper QE89-16-REV. June 1990.
D'Antonio, R.G., R.E. Winn, J.P. Taylor, T.L. Gustafson, W.L. Current, M.M. Rhodes, W. Gary, and
R.A. Zajac. 1985. A waterborne outbreak of Cryptosporidiosis in normal hosts. Annals of
Internal Medicine. 103:886-888.
Dahling, D. R. 2002. An improved filter elution and cell culture assay procedure for evaluating public
ground water systems for culturable enteroviruses. Water Environmental Research. 74(6):564-
568.
Davis, J.V. and E.G. Witt, III. 2000. Microbiological and chemical quality of ground water used as a
source of public supply in Southern Missouri- phase I, May 1997-March 1998. Water -Resources
Investigations Report 00-4038. USGS, US DOI, Rolla, MI. 77.
Economic Analysis for the R-5 October 2006
Final Ground Water Rule
-------
De Borde, B.C., and R. Ward. 1995. Results of one year of virus testing at two high-yield water table
wells in areas served by septic systems. Unpublished report to Mountain Water Co., Missoula,
MT.
Denis-Mize, K., G.S. Fout, D.R. Dahling, and D.S. Francy. 2004. Detection of human enteric viruses in
stream water with PCR and cell culture. Journal of Water Health. 2:37-47.
De Serres, G., T.L. Cromeans, B. Levesque, N. Brassard, C. Barthe, M. Dionne, H. Prud'homme, D.
Paradis, C.N. Shapiro, M.V. Nainan, and H.S. Margolis. 1999. Molecular confirmation of
Hepatitis A virus from well water: Epidemiology and public health implications. Journal of
Infectious Diseases. 179:37-43.
DFiHS. 1994. International classification of diseases, 9th revision, clinical modification. 5th edition,
volume 1. Washington, DC: U.S. Government Printing Office. DHHS Publ. No. PHS 94-1260.
December.
Dirckx, J.H., ed. 1997. Stedman's concise medical dictionary for the health professions, 3rd ed.
Baltimore, MD: Williams & Wilkins.
Doherty, K. 1998. Status of the New England ground water viral study. Proceedings, AWWA Annual
Meeting: Dallas, TX. June 23, 1998. AWWA: Denver.
Dworkin M.S., D.P. Goldman, T.G. Wells, J.M. Kobayashi, and B. Herwaldt. 1996. Cryptospsoridiosis in
Washington State: An outbreak associated with well water. Journal of Infectious Diseases.
174:1372-1376.
Eisenberg, J. N., E.Y.W. Seto, A.W. Olivieri, and R.C. Spear. 1996. Quantifying water pathogen risk in
an epidemiological framework. Risk Analysis. 16(4):549-563.
Eisenberg, J. N. S., X.D. Lei, A.H. Hubbard, M.A. Brookhart, and J.M. Colford. 2005. The role of
disease transmission and conferred immunity in outbreaks: Analysis of the 1993 Cryptosporidium
outbreak in Milwaukee, Wisconsin. American Journal of Epidemiology. 161(l):62-72.
Eisenberg, J. N. S., E.Y.W. Seto, J.M. Colford, Jr., A. Olivieri, and RC. Spear. 1998. An analysis of the
Milwaukee cryptosporidiosis outbreak based on a dynamic model of the infection process.
Epidemiology. 9(3):255-263.
Eisenberg, J. N. S., J.A. Seller, J. Scott, D.M. Eisenberg, and J.M. Colford. 2004. A dynamic model to
assess microbial health risks associated with beneficial uses of biosolids. Risk Analysis.
24(l):221-236.
Elixhauser, A. et al. 1993. Clinical classifications for health policy research: Discharge statistics by
principal diagnosis and procedure. Division of Provider Studies Research Note 17, Agency for
Health Care Policy and Research, Rockville, MD: Public Health Services. AHCPR Publ. No. 93-
0043.
Economic Analysis for the R-6 October 2006
Final Ground Water Rule
-------
Fankhauser, R.L., S.S. Monroe, J.S. Noel, C.D. Humphrey, J.S. Bresee, U.D. Parashar, T. Ando, and R.I.
Glass. 2002. Epidemiologic and molecular trends of Norwalk-like viruses' associated with
outbreaks of gastroenteritis in the United States. Journal of Infectious Diseases. 186:1-7.
Felkner, M., K. Hendricks, L. Surarez, and D.K. Waller. 2003. Diarrhea: a new risk factor for neural tube
defects?. Birth Defects Res A din Mol. Teratol. 67(7):504-508.
Femmer, S. 1999. Microbiological quality of older wells in public water supplies in the Ozark Plateaus
aquifer system, Missouri. Unpublished report to Missouri Department of Natural Resources.
Femmer, S. 2000. Microbiological and chemical quality of ground water used as a source of public
supply in southern Missouri - phase II, April-July, 1998. Water-Resources Investigations Report
00-4260, US DOI, USGS, Rolla, MI. 62.
Person, M. J. 1996. Hospitalizations for rotavirus gastroenteritis among children under five years of age
in New South Wales. Medical Journal of Australia. 164(5):273-276.
Person, M., S. Streingfellow, K. McPhie, C. Mclver, and A. Simos. 1997. Longitudinal study of rotavirus
infection in child-care centres. Paediatrics Child Health.
Fletcher, M., M. E. Levy, D.D. Griffin. 2000. Foodborne outbreak of Group A rotavirus gastroenteritis
among college students, District of Columbia, March-April 2000. Morbidity and Mortality
Weekly Report. 49:1131-1133.
Flewett, T. H., A.S. Bryden, and H. Davies. 1975. Epidemic viral enteritis in a long-stay children's ward.
Lancet. l(7897):4-5.
Flewett, T. H., and G.N. Woode. 1978. The rotaviruses: brief review. Archive of Virology. 57:1-23.
Fout, G. S., B.C. Martinson, M.W.N. Moyer, and D.R. Dahling. 2003. A multiplex reverse transcription-
PCR method for detection of human enteric viruses in ground water. Applied and Environmental
Microbiology. 69(6):3158-3164.
Fox, N.R. et al. 1993. Direct and indirect costs of diabetes in the United States in 1992. American
Diabetes Association: Alexandria, VA.
Fox, N.R., et al. 1998. Economic consequences of diabetes mellitus in the U.S. in 1997. Diabetes Care.
21(2):296-309.
Framm, S.R. and R. Soave. 1997. Agents of diarrhea. The Medical clinics of North America. 81(2):427-
447.
Francy, D., D. Helsel, and R. Nally. 2000. Occurrence and distribution of microbiological indicators in
groundwater and stream water. Water Environmental Research. 72(2): 152-161.
Economic Analysis for the R-7 October 2006
Final Ground Water Rule
-------
Francy, D.S., R.N. Bushon, J. Stopar, E.J. Luzano, and G.S. Fout. 2004. Environmental factors and
chemical and microbiological water-quality constituents related to the presence of enteric viruses
in ground water from small public water supplies in Southeastern Michigan. USGS Scientific
Investigations Report. 54:3004-5219.
Frost, F.J., G.F. Craun, and R.L. Calderon. 1996. Waterborne disease surveillance. Journal of the
American Water Works Association. 88(9):66-75.
Fruhwirth, M., K. Berger, B. Ehlken, I. Moll-Schuler, S. Brosl, I. Mutz. 2001. Economic impact of
community and nosocomially acquired rotavirus gastroenteritis in Austria. Pediatric Infection
Disease Journal. 20(2): 184-188.
Fujioka, R.S. and B.S. Yoneyama. 2001. Assessing the vulnerability of groundwater sources to fecal
contamination. Journal of American Water Works Association. 93(8):62-71.
Gallay, A., H. De Valk, M. Cournot, B. Ladeuil, C. Hemery, C. Castor, F. Bon, F. Megraud, P. Le Cann,
and J.C. Desenclos. 2006. A large multi-pathogen waterborne community outbreak linked to
faecal contamination of a groundwater system, France, 2000. Clinical Microbiology and
Infection. 12(6): 561-570.
Garg, A.X., R.S. Suri, N. Barrowman, F. Rehman, D. Matsell, M. Patricia Rosas-Arellano, M. Salvadori,
R.B. Haynes and W.F. Clark. 2003. Long-term renal prognosis of diarrhea-associated hemolytic
uremic syndrome: A systematic review, meta-anlysis and meta-regression. Journal ofAmerincan
Medical Association. 290(10): 1360-1370.
Garg, A. X., L. Moist, D. Matsell, H.R. Thiessen-Philbrook, R.B. Haynes, RS.Suri, M. Salvadori, J. Ray,
and W.F. Clark. 2005. Risk of hypertension and reduced kidney function after acute
gastroenteritis from bacteria-contaminated drinking water. Canadian Medical Association
Journal. 173(3): 1-8.
Garthright, W. E. et al. Estimates of incidence and costs of intestinal infectious diseases in the US.
Public Health Reports. 103(2): 107-115.
Geldreich, E.E. 1996. Microbial quality of water supply in distribution systems. Boca Raton, FL: Lewis
Publishers.
Gelman, A., J.B. Carlin, H.S. Stern, and D.B. Rubin. 1995. Bayesian Data Analysis, New York: Chapman
& Hall.
Gerba, C.P., J.B. Rose, and C.N. Haas. 1996a. Sensitive subpopulations: who is at the greatest risk?
International Journal of Food and Microbiology. 30:113-123.
Gerba, C.P., J.B. Rose, C.N. Haas and K.D. Crabtree. 1996b. Waterborne rotavirus: a risk assessment.
Water Resources. 30(12):2929-2940.
Gewurz, A., R. Potempa, C. Goetz. 1985. Coxsackie A-l 1 encephalitis (CAE) in a patient with common
variable immunodeficiency (CVID): Response to intravenous and intraventricular treatment with
intravenous immune globulin (I VIG). Annals of Allergy. 55:272.
Economic Analysis for the R-8 October 2006
Final Ground Water Rule
-------
Gilks W.R., A. Thomas, and D.J. Spiegelhalter. 1994. A language and program for complex Bayesian
modelling. The Statistician (Volume 43) pp. 169-178.
Glass, R.I., P.E. Kilgore, R.C. Holman, S. Jin, J.C. Smith, P.A. Woods, M.J. Clarke, M.S. Ho, and J.R.
Gentsch. 1996. The epidemiology of rotavirus diarrhea in the United States: surveillance and
estimates of disease burden. Journal of Infectious Diseases. 174(Suppl 1):S5-S11.
Gofti-Laroche, L., B. Gratacap-Cavallier, D. Demanse, O. Genoulaz, J.M. Seigncurin, and D. Zmirou.
2003. Are waterborne astrovirus implicated in acute digestive morbidity (EMIRA Study)?
Journal of Clinical Virology. 27:74-82.
Gomez-Barreto, J., E.L. Plamer, A.J. Nahmias, and M.H. Hatch. 1976. Acute enteritis associated with
reovirus-like agents. Journal of the American Medical Association, 235:1857-1860.
Griffin, D.D, M. Fletcher, M.E. Levy, M. Ching-Lee, R. Nogami, L. Edwards, H. Peters, L. Montague,
J.R. Gentsch, and R.I. Glass. 2002. Outbreaks of adult gastroenteritis traced to a single genotype
of rotavirus. Journal of Infectious Diseases. 185:1502-1505.
Grimwood, K., Q. S. Huang, L.G. Sadleir, W.A. Nix, D.R Kilpatrick, M.S. Oberste, and M.A. Pallansch.
2003. Acute Flaccid Paralysis from Echovirus Type 33 Infection. Journal of Clinical
Microbiology. Vol. 41, No. 5.
Gurwith, M., W. Wenman, D. Hinde, S. Heltham, and H. Greenberg. 1981. A prospective study of
rotavirus infection in infants and young children. Journal of Infectious Diseases. 144(3).
Haas, C. N. 1983. Estimation of risk due to low doses of microorganisms: A comparison of alternative
methodologies. American Journal of Epidemiology. 55:573-582.
Haas, C.N., J.B. Rose, C.P. Gerba. 1999. Quantitative microbial risk assessment, p.267. New York, NY:
John Wiley & Sons, Inc.
Haffejee, I. E. 1995. The epidemiology of rotavirus infections: a global perspective. Journal of
Pediatric Gastroenterology Nutrition. 20:275-86.
Hall, C. E., M. K. Cooney, and J. P. Fox. 1970. The Seattle virus watch program. I. Infection and illness
experience of virus watch families during a community-wide epidemic of echovirus type 30
aseptic meningitis. American Journal of Public Health. 60:1456-1465.
Hamilton, M. S., M. A. Jackson, D. Abel. 1999. Clinical utility of polymerase chain reaction testing for
enteroviral meningitis. Pediatric Infectious Diseases Journal. 18(6):533-537.
Hammond, P. B., and R. Coppick. 1990. Valuing health risks, costs and benefits for environmental
decision making. Washington, DC: National Academy Press.
Hancock, C.M., J.B. Rose, and M. Callahan. 1998. Cryptosporidium and giardia in U.S. ground water.
Journal of the American Water Works Association. 90:58-61.
Economic Analysis for the R-9 October 2006
Final Ground Water Rule
-------
Harrington, W. et al. 1991. Economics and episodic disease: The benefits of preventing a giardiasis
outbreak. Resources for the Future: Washington, B.C.
Health Canada. 2000. Waterborne outbreak of gastroenteritis associated with a contaminated municipal
water supply, Walkerton, Ontario, May-June 2000. Communicable Disease Report. 26-20.
Hegarty, J.P., M.T. Dowd and K.H. Baker. 1999. Occurrence of helicobacter pylori in surface water in the
United States. Journal of Applied Microbiology. 87:697-701.
Heiselman, D. E. 1997. Pericarditis. In: Griffith's 5 minute checklist consult. Dambro, M. R. (ed.).
p.786-787. Baltimore, MD: Williams & Wilkins.
Hejkal, T.W., B. Keswick, R.L. LaBelle, C.P. Gerba, Y. Sanchez, G. Dressman, B. Hafkin, and J.L.
Melnick. 1982. Viruses in a community water supply associated with an outbreak of
gastroenteritis and infectious hepatitis. Journal of the American Water Works Association.
74:318-321.
Hertal, N.T., F.K. Pedersen, and C. Heilmann. 1989. Coxsackie B3 virus encephalitis in a patient with
agammaglobulinaemia. European Journal of Pediatrics. 148:642-3.
Hethcote, H. 1976. Qualitative analyses of communicable disease models. Mathematical Biosciences.
28:335-356.
Hethcote, H. W. 2000. The mathematics of infectious diseases. Siam Review. 42(4):599-653.
Hodgson, T.A. and A.N. Kopstein. 1984. Health care expenditures for major diseases in 1980. Health
Care Financing Review. 5 (4): 1 -12.
Holler, C., S. Koschinsky and D. Witthuhn. 1999. Isolation of enterohaemorrhagic Escherichia coli from
municipal sewage. Lancet. 353:2039.
Hrdy, D. B. 1987. Epidemiology of rotaviral infection in adults. Reviews of Infectious Diseases.
9(3):461-9.
Huang, Q.S., J.M. Carr, W.A. Nix, M.S. Oberste, D.R. Kilpatrick, M.A. Pallansch, M.C. Croxson, J.A.
Lindeman, M.G. Baker and K. Grimwood. 2003. An echovirus type 33 winter outbreak in New
Zealand. Clinical Infectious Diseases. 37:650-657.
Hufnagel, G. 1998. Symptoms, diagnosis and treatment of myocarditis and dilated cardiomyopathy
(DCM). http://www.unmc.edu/Pathology/Myocarditis/whatisdcm.hmtl.
Huhn, G.D., C. Gross, D. Schnurr, C. Preas, Y. Shigeo, S. Reagan, C. Paddock, D. Passaro, and M.S.
Dworkin. 2005. Myocarditis outbreak among adults, Illinois, 2003. Emerging Infectious
Diseases. 11(10): 1621-1624.
Hunter, P. 2003. Drinking water and diarrheal disease due to Escherichia coli. Journal of Water and
Health. 1(2): 65-72.
Economic Analysis for the R-10 October 2006
Final Ground Water Rule
-------
International Life Sciences Institute (ILSI). 1996. A conceptual framework to assess the risks of human
disease following exposure to pathogens. Risk Analysis. 16(6):841:848.
ILSI. 2000. Revised framework for microbial risk assessment. Washington, DC: ILSI Press.
Irving, T.E. J. de Louvois and G.L. Nichols. 1996. Fact sheets on emerging waterborne pathogens: Final
report to the Department of the Environment." AWWA Research Foundation. Report No: DWI
4248/1.
INTELLIMED. 2002. MyHealthScore. Healthcare quality information for consumers. Medical cost
database by CPT-4 code, formerly MECQA. http://www.myhealthscore.com.
Isakbaeva, E.T., M. Widdowson, R.S. Beard, S.N. Bullens, J. Mullins, S.S. Monre, J. Bresee, P. Sassno,
E. H. Cramer, and R.I. Glass. 2005. Noravirus Transmission on Cruise Ship. Emerging
infectious diseases. 1 1(1): 154-157.
Jenista , J.A., K.R. Powell, and M.A. Menegus. 1984. Epidemiology of neonatal enterovirus infection.
Journal of Pediatrics. 104:685-690.
Jiang, B., J.R. Gentsch, and R.I. Glass. 2002. The role of serum antibodies in the protection against
rotavirus disease: an overview. Clinical Infectious Diseases. 34:1351-1361.
Jin, S., P. E. Kilgore, R. C. Holman, M. J. Clarke, E. J. Gangarosa, R. I. Glass. 1996. Trends in
hospitalizations for diarrhea in United States children from 1979 through 1992; estimates of the
morbidity associated with rotavirus. Pediatric Infectious Diseases Journal. 15(5):397-404.
Johnson, P. C., J.J. Mathewson, H.L. DuPont, and H.B. Greenberg. 1990. Multiple -challenge study of
host susceptibility to Norwalk gastroenteritis in US adults. Journal of Infectious Diseases.
Kafetzis, D. A., H. C. Maltezou, A. Zafeiropoulou, A. Attilakos, C. Stavrinadis, M. Foustoukou. 2001.
Epidemiology, clinical course and impact on hospitalization costs of acute diarrhea among
hospitalized children in Athens, Greece. Scandinavian Journal of Infectious Diseases.
33(9):681-685.
Kapikian, A. Z. and R. M. Chanock. 1996. Rotaviruses. In: Virology, 2nd ed., vol. 1 & 2., B.N. Fields et
al. (eds.). New York, NY: Raven Press.
Kapikian, A., R.G. Wyatt, H.B. Greenberg, A.R Kalica, H.W. Kim, C.D. Brandt, W.J. Rodrigues, R.H.
Parrott, and R.M. Chanock. 1980. Approaches to immunization of infants and young children
against gastroenteritis due to rotaviruses. Reviews of Infectious Diseases. 2(3):459-469.
Kapikian, A.Z., Y. Hoshino, and R.M. Chanock. 2001. Rotaviruses. In Fields virology. Howley P.M. Ed.
Philadelphia, Lippincott Williams & Wilkins. pp. 1787-1833.
Kaplan, M.H. and S.W. Klein. 1983. Coxsackievirus infections in infants younger than three months of
age: A serious childhood illness. Reviews of Infectious Diseases. 5: 1019-1033.
Economic Analysis for the R-ll October 2006
Final Ground Water Rule
-------
Karim, M., M. LeChevallier, M. Abbaszadegan, and J.Rosen. 2002. Field testing of USEPA method
1601 for coliphage. AWWA Research Foundation.
Karim, M.R., M. Abbaszadegan, A. Alum, and M. LeChevallier. 2003. Virological quality of
groundwater. In: Proceedings, Water Quality Technology Conference: Philadelphia, PA.
October, 1999.
Karim, M.R., M. LeChevallier, M. Abbaszadegan, A. Alum, J. Sobrinho, and J. Rosen. 2004. Microbial
indicators for assessing the vulnerability of groundwater to fecal contamination. American Water
Co. Report. 106pg.
Karzon, D.T., G.L. Eckert, A.L. Barren et al. 1961. Aseptic meningitis due to Echov 4 virus. American
Journal of Diseases of Children. 101:102-114.
Katz, M., and S.A. Plotkin. 1967. Minimal infective dose of attenuated polio virus for man. American
Journal of Public Health. 37:1837.
Kawai, S. et al. 1987. A morphological analysis of chronic myocarditis. Japanese Circulation Journal
51:1385-91.
Kearney, M.T. J.M. Cotton, P.J. Richardson and A.M. Shah. 2001. Viral myocarditis and dilated
cardiomyopathy: mechanisms, manifestations and management. Postgraduate Medical Journal.
77:4-10.
Khetsuriani, N., R.C. Holman and L.J. Anderson. 2002. Burden of encephalitis-associated
hospitalizations in the United States, 1988-1997. Clinical Infectious Diseases. 35:175-182.
Khetsuriani, N., E.S. Quiroz, R.C. Holman and L.J. Anderson. 2003. Viral meningitis-associated
hospitalizations in the United States, 1988-1999. Neuroepidemiology. 22:345-352.
Kilgore, P.E., R.C. Holman, M.J. Clarke, and R.I. Glass. 1995. Trends of diarrheal disease—associated
mortality in U.S. children, 1968 through 1991. Journal of the American Medical Association.
274(14):1143-1148.
Kim, H.W., C.D. Brandt, A.Z. Kapikian, RG. Witt, J.O. Arrobio, W.J. Rodriguez, RM. Chanock, and
R.H. Parrott. 1997. Human reovirus-like agent infection: occurrence in adult contacts of
pediatric patients with gastroenteritis. Journal of the American Medical Association.
238(5):404-407.
Kim K., G. Hufnagel, N.M. Chapman, and S. Tracy. 2001. The group B coxsackievirues and
myocarditis. Reviews in Medical Virology. 11:355-368.
Kleckner, N. and J. Neumann. 2000. Memorandum: Update to recommended approach to adjusting
WTP estimates to reflect changes in real income. September 30, 2000.
Kleinbaum, D. G., L.L. Kupper, and H. Morgensterm. 1982. Epidemiologic research: principles and
quantitative methods. New York, NY: Van Nostrand Reinhold.
Economic Analysis for the R-12 October 2006
Final Ground Water Rule
-------
Koenraad, P.M.F.J., P.M. Rombouts and S.H.W. Notermans. 1997. Epidemiological aspects of
theromophilic Camplyobacter in water-related environments: A review. Water Environmental
Research. 69(l):52-63.
Kogon, A., I. Spigland, T.E. Frothingham, L. Elveback, C. Williams, C.E. Hall, and J.P. Fox. 1969. The
virus watch program: a continuing surveillance of viral infections in metropolitan New York
families. American Journal of Epidemiology. 89(1):51-61.
Koopman, J. S., S.E. Chick, C.P. Simon, C.S. Riolo, and G. Jacquez. 2002. Stochastic effects on endemic
infection levels of disseminating versus local contacts. Mathematical Biosciences.
180(SI.):49-71.
Koopman, J. S., G. Jacquez, and S.E. Chick. 2001. New data and tools for integrating
discrete and continuous population modeling strategies. Annual New York Academy of Sciences.
954:268-294.
Koopman, J. S., I.M. Longini, J.A. Jacquez, and C.P. Simon. 1991a. Assessing risk factors for
transmission of infection. American Journal of Epidemiology. 133(12): 1199-1209.
Koopman, J. S., and A.S. Monto. 1989. The Tecumseh Study XV: Rotavirus infection and
pathogenicity. American Journal of Epidemiology. 130(4):760-768.
Koopman, J. S., A.S. Monto, and I.M. Longini. 1989. The Tecumseh study XVI: family and community
sources of rotavirus infection. American Journal of Epidemiology. 130(4):760-768.
Koopman, J. S., D.R. Prevots, M.A. Vaca Marin, H. Gomez Dantes, M.L. Zarate Aquino, I.M. Longini,
Jr., and J. Sepulveda Amor. 1991b. Determinants and predictors of dengue infection in Mexico.
American Journal of Epidemiology. 133(11): 1168-78.
Kramer, M.H., B.L. Herwaldt, R.L. Calderon, and D.D. Juranek. 1996. Surveillance for
waterborne-disease outbreaks—United States, 1993-1994. Morbidity and Mortality Weekly
Report. 45(SS-l):l-33.
Lanfear, K.J. 1992. A Database of nitrate in ground-water samples from the conterminous United States.
U.S. Geological Survey Open File Report. 92-652.
Lawson, H.W., M.M. Braun, R.I.M. Glass, S. Stine, S. Monroe, H.K. Atrash, L.E. Lee, and S.J.
Engelender. 1991. Waterborne outbreak of Norwalk virus gastroenteritis at a southwest U.S.
resort: role of geological formations in contamination of well water. Lancet. 337:1200-1204.
LeBaron, C.W., M.D. Furutan, J.F. Lew, J.R. Allen, D.V. Gouvea, D.C. Moe, D.S. Monroe. 1990. Viral
agents of gastroenteritis; public health importance and outbreak management. Morbidity and
Mortality Weekly Report. 39(RR-5): 1-24.
Lee, S.H., D.A. Levy, G.F., Craun, M.J. Beach, and R.L. Calderon. 2002. Surveillance for waterborne-
disease outbreaks—United States, 1999-2000. Morbidity and Mortality Weekly Report.
51(SS08):l-28.
Economic Analysis for the R-13 October 2006
Final Ground Water Rule
-------
Lee, Y, P.W. Johnson, J.L. Call, M.J. Arrowood, B.W. Furness, S. C. Pichette, K.K. Grady, P. Reeh, L.
Mitchell, D. Bermire-Sweat, W. MacKenzie and V.C.W. Tsang. 2001. Development and
application of a quantitative, specific assay for Cryptosporidium parvum oocst detection in
high-turbidity environmental samples. American Journal of Tropical Medicine and Hygiene.
Lehan, P.H., E.W. Chick, I.L. Doto et al. 1957. An epidemic illness associated with a recently
recognized enteric virus (Echo virus type 4), I. Epidemiologic and clinical features. American
Journal of Hygiene . 66 : 63 -75 .
Lepow, M.L., R.J. Warren, and V.G. Ingram. 1962. Sabin type 1 oral poliomyelitis vaccine: effect of
dose upon response of newborne infants. American Journal of Diseases of Children. 104:67.
Levy, D.A., M.S. Bens, G.F. Craun, R.L. Calderon, and B.L. Herwaldt. 1998. Surveillance for
waterborne disease outbreaks - United States, 1995-1996. Morbidity and Mortality Weekly
Report. 47(55-5): 1-34.
Lew, J.F., R.I. Glass, R.E. Gangarosa, I.P. Cohen, C. Bern, and C. Moe. 1991. Diarrheal deaths in the
United States, 1979 through 1987, a special problem for the elderly. Journal of the American
Medical Association. 265(24):3280-3284.
Liddle, J. L. M. et al. Rotavirus gastroenteritis: impact on young children, their families and the health
care system. Medical Journal of Australia. 167:304-307.
Lieberman, R.J., L.C. Shadix, B.S. Newport, S.R. Crout, S.E. Buescher, R.S. Safferman, R.E. Stetler, D.
Lye, G.S. Fout, and D. Dahling. 1994. Source water microbial quality of some vulnerable public
ground water supplies. In: Proceedings, Water Quality Technology Conference: San Francisco,
CA. October, 1994.
Lieberman, R.J., L.C. Shadix, B.S. Newport, M.W.N. Frebis, S.E. Moyer, R.S. Safferman, R.E. Stetler,
D. Lye, G.S. Fout, and D. Dahling. 2002. Microbial Monitoring of Vulnerable Public Ground
Water Supplies. EPA/AWWARF.
Lin, T., S. Twu, M. Ho, L. Chang, and C. Lee. 2003. Enterovirus 71 outbreaks, Taiwan: Occurrence and
recognition. Emerging Infectious Diseases. 9(3):291-293.
Lindesmith, L., C. Moe, S. Marionneou, R. Ruvoen, X. Jiang, L. Lindblad, P. Stewart, J. LePendu, and R.
Baric. 2003. Human susceptibility and resistance to Norwalk virus infection. Nature Medicine.
9(5):548-552.
Lindsey, B.D., J.S. Raspberry, and T.M. Zimmerman. 2002. Microbiological quality of water from
noncommunity supply wells in carbonate and crystalline aquifers of Pennsylvania. U.S.
Geological Survey Water-Resources Investigations Report 01-4268, 30 p.
Linhares, A. C., P.P. Pinheiro, R.B. Freitas, Y.B. Gabbay, J.A. Shirley, and G.M. Beards. 1981. An
outbreak of rotavirus diarrhea among a nonimmune, isolated South American Indian community.
American Journal of Epidemiology . 1 1 3 (6) : 703 - 1 0 .
Economic Analysis for the R-14 October 2006
Final Ground Water Rule
-------
Livernois, J. 2002. The Economic Costs of the Walkerton Crisis. The Walkerton Inquiry. Toronto:
Ontario Ministry of the Attorney General.
Longini I.M. and J.S. Koopman. 1982. Household and community transmission parameters from final
distributions of infections in households. Biometrics. 38:115-126.
Lycke, E., J. Blmber, G. Berg, A. Eriksson, and L. Madsen. 1978. Epidemic acute diarrhea in adults
associated with infantile gastroenteritis virus. Lancet. 2:1056-1057.
Lynch, M., F. O'Halloran, D. Whyte, S. Fanning, B. Cryan, R. I. Glass. 2001. Rotavirus in Ireland:
national estimates of disease burden, 1997 to 1998. Pediatric Infectious Diseases Journal.
(7):693-698.
Maasdam, C.F. and S. Anuras. 1981. Are you overlooking GI infections in your elderly patients?
Geriatrics. 36:127-134.
Magnani, J.W. and G.W. 2006. Myocarditis, current trends in diagnosis and treatment. Circulation
113:876-890.
Marano N., D. Vugia, T. Fiorentino et al. 2000. Fluoroquinolone-resistant Campylobacter causes longer
duration of diarrhea than fluoroquinolone-suceptible Campylobacter strains in FoodNet sites
[abstract]. In: Program and abstracts of the International Conference on Emerging Infectious
Diseases 2000 (Atlanta). Centers for Disease Control and Prevention: Atlanta, 2000.
Maria, H., A. Elshebani, O. Anders, T. Torsten, and F. Gun. 2005. Sumultaneous type 1 diabetes onset in
mother and son coincident with an enteroviral infection. Journal of Clinical Virology.
33:158-167.
Marciano-Cabral, F., R. MacLean, A. Mensah, L. LaPat-Polasko. 2003. Identification of Naegleria
fowleri in domestic water wources by nested PCR. Applied and Environmental Microbiology.
69(10):5864-5869.
Marshall, J., J. Botes, G. Gorrie, C. Boardman, J. Gregory, J. Griffith, G. Hogg, A Dimitriadis, M. Carton
and R. Bishop. 2003. Rotavirus detection and characterization in outbreaks of gastroenteritis in
aged-care facilities. Journal of Clinical Virology. 28:331-340.
Matson, D. O. and M. K. Estes. 1990. Impact of rotavirus infection at a large pediatric hospital. Journal
of Infectious Diseases. 162:598-604.
Matsui, S. M., and H.B. Greenberg. 2000. Immunity to calicivirus infection. Journal of Infectious
Diseases. 181(Suppl 2):S331-5.
Maynard C., J.M. Fairbrother, S. Bekal et al. 2003. Antimicrobial resistance genes in enterotoxigenic
Escherichia coli O149:K91 isolates obtained over a 23-year period from pigs. Antimicrobial
Agents and Chemotherapy. 47(10): 3214-3221.
McKinney, R.E., S.K. Kotz and C.M. Wilfert. 1987. Chronic enteroviral meningoencephalitis in
agammaglobulinemic patients. Reviews of Infectious Diseases. 9:334-56.
Economic Analysis for the R-15 October 2006
Final Ground Water Rule
-------
McLaughlin, J.B., B.D. Gessner, T.V. Lynn, E.A. Funk and J.P. Middaugh. 2004. Association of
regulatory issues with and Echovirus 18 meningitis outbreak at a children's summer camp in
Alaska. Pediatric Infectious Disease Journal. 23(9):875-877.
McMillan, S. 1996. Camp Four Echoes outbreak investigation. Unpublished report by the Panhandle
Health District, Environmental Health Division, Coeur d'Alene, Idaho.
McMinn, P., I. Stratov, L. Nagarajan and S. Davis. 2001. Neurological manifestations of enterovirus 71
infection in children during an outbreak of hand, foot and mouth disease in Western Australia.
Clinical Infectious Diseases. 32:236-242.
Mead, P.S., L. Slutsker, V. Dietz, L.F. McCaig, J.S. Bresee, C. Shapiro, P.M. Griffin, and R.V. Tauxe.
1999. Food-related illness and death in the United States. Emerging Infectious Diseases
5(5):607-625.
Melnick, J.L. 1996. Enteroviruses: Polioviruses, coxsackieviruses, echoviruses, and newer enteroviruses.
In:. Virology, 3rd ed, vol. 2., Fields, B.N. et al., (eds.). pp. 655-712. Philadelphia,?A: Lippincott-
Raven Publishers.
Miller, G. M. 1997. Viral meningitis. In: Griffith's 5 minute clinical consult., Dambro, M. R. (ed.).
p.662-663. Baltimore, MD: Wiliams & Wilkins.
Minnesota Department of Health. 2000. Minnesota Department of Health viral occurrence study.
Minnesota Department of Health: St. Paul, 7 p.
Minor, T.E., C.I. Allen, and A.A. Tsiatis. 1981. Human infective dose determinations for oral poliovirus
type I vaccine in infants. Journal of Clinical Microbiology. 13:388.
Missouri Department of Health. 1992. Summary of investigation, hepatitis A outbreak, Apostolic church
and school, Racine, MO. Unpublished report, 5 p.
Modlin, J. F. 1986. Perinatal echovirus infection: insights from a literature review of 61 cases of serious
infection and 16 outbreaks in nursuries. Reviews of Infectious Diseases. 8(6): 918-926.
Modlin, J. F. 1995. Coxsackieviruses, echoviruses, and newer enteroviruses. In: Principles and practice
of infectious diseases., Mandell, G. L. et al., (eds.). pp. 1620-1636. New York, NY: Churchill
Livingston, Inc.
Moe, C.L., J.A. Frelinger, W. Heizer, P. Stewart. 2001. Final Report: Studies of the Infectivity of
Norwalk and Norwalk-like Viruses. USEPA Report. USEPA Grant Number: R826139.
Molyneaux, P. 1995. Human immunity of rotavirus. Journal of Medical Microbiology. 43 (6): 3 97-404.
Moore, M. 1982. Enteroviral Disease in the United States, 1970-1979. Journal of Infectious Diseases.
146(1):103-108.
Economic Analysis for the R-16 October 2006
Final Ground Water Rule
-------
Moore, A.C., B.L. Herwaldt, G.F. Craun, R. L. Calderon, A.K. Highsmith, and D.D. Juranek. 1993.
Surveillance for waterborne disease outbreaks - United States, 1991-1992. Morbidity and
Mortality Weekly Report. Surveillance Summary SS-5, US Centers for Disease Control and
Prevention.
Moore, M.J. and W. Kip Viscusi. 1990. Compensation mechanisms for job risks: wages, workers'
compensation, and product liability. Princeton, NJ: Princeton University Press.
Moore A.C., B.L. Herwaldt, G.F. Craun, R. L. Calderon, A.K. Highsmith, and D.D. Juranek. 1993.
Surveillance for waterborne disease outbreaks-United States, 1991-1992. Morbidity and
Mortality Weekly Report. 42(SS-5):l-22.
Morens, D.M. 1978. Enteroviral diseases in early infancy. Journal of Pediatrics. 92(3):374-7.
Morens, D. M., M. A. Pallansch, and M. Moore. 1991. Polioviruses and other enteroviruses. In:
Textbook of Human Virology, 2nd ed., Robert B. Belshe, (ed.). St. Louis: Mosby Year Book.
Morpeth, S.C., and N.M. Thielman. 2006. Diarrhea in patients with AIDS. Current Treatment Options in
Gastroenterology. 9(1): 23-37.
Morris, G.J. and M. Potter. 1997. Emergence of New Pathogens as a Function of Changes in Host
Suseptibility. BID. Vol. 3, No. 4.
National Cancer Institute (NCI). 2000. SEER cancer Data, as of January 1, 2000.
National Center for Health Statistics (NCHS). 1994. Current estimates from the National Health
Interview Survey. Vital and Health Statistics. 10:190.
NCHS. 1998. International Classification of Diseases, 9th Revision.
NCHS. 1998. Vital statistics of the United States, 1995, preprint of volume II, mortality, part A sec 6 life
tables. Hyattsville, MD.
National Drinking Water Advisory Council. 2001. Report of the Arsenic Cost Working Group to the
National Drinking Water Council. August 14, 2001.
National Institute of Child Health and Human Development (NICHD), 1999. When the Body's Defenses
are Missing. Primary Immunodeficiency. DHHS, NICHD (NIH Publication No. 99-4149),
Washington, DC, U.S. Government Printing Office.
National Health, Lung and Blood Institute (NHLBI), National Institutes of Health. 1996. Morbidity and
mortality: 1996 chartbook on cardiovascular, lung, and blood diseases. May.
National Marrow Donor Program 2004. Facts & Figures.
http: //www .marrow. org/MEDI A/facts_figures .pdf.
National Research Council (NRC). 1983. Risk assessment in the federal government: Managing the
process. Washington, DC: National Academy Press.
Economic Analysis for the R-17 October 2006
Final Ground Water Rule
-------
NRC. 1997a. Safe water from every tap, improving water service to small communities. Washington,
DC: National Academy Press.
NRC. 1997b. Valuing ground water: economic concepts and approaches. Washington, DC: National
Academy Press.
NRC and Commission on Behavioral and Social Sciences and Education. 2000. Time use measurement
and research: report of a workshop., Committee on National Statistics: Michele Ver Ploeg,
Joseph Altonji, Norman Bradburn, Julie DaVanzo, William Nordhaus, and Francisco Samaniego
(eds.). Washington, D.C: National Academy Press.
National Water Research Institute (NWRI). 1997. Groundwater disinfection regulation workshop.
January 6-8.
Nigrovic, L. E. 2001. What's new with enteroviral infections?. Current Opinion in Pediatrics. 13(1): 89-
94.
Nigrovic, L. E. and V. W. Chiang. 2000. Cost analysis of enteroviral polymerase chain reaction in
infants with fever and cerebrospinal fluid pleocytosis. Archive ofPediatric and Adolescent
Medicine. 154(8):817-821.
Ohio EPA. 2005. South Bass Island, Ottawa County gastrointestinal illness, summer 2004, Ohio
Environmental Protection Agency investigation and actions. Unpublished Report, 42 p.
O'Ryan, M. L., D.O. Matson, M.K. Estes, and L.K. Pickering. 1993. Anti-rotavirus Gtype - specific and
isotype-specific antibodies in children with natural rotavirus infections. Journal of Infectious
Diseases. 169:504-511.
O'Ryan, M., I. Perez-Schael, N. Mamani, A. Pena, B. Salinas, G. Gonzalez, F. Gonzalez, D. O. Matson, J.
Gomez. 2001. Rotavirus-associated medical visits and hospitalizations in South America: a
prospective study at three large sentinel hospitals. Pediatric Infectious Diseases Journal.
20(7):685-693. July.
Parashar, U.D., R.C. Holman, M.J. Clarke, J.S. Bresee, and R.I. Glass. 1998. Hospitalizations associated
with rotavirus diarrhea in the United States, 1993 through 1995: surveillance based on the new
ICD-9-CM rotavirus-specific diagnostic code. Journal of Infectious Diseases. Ill: 13-17.
Parashar, U.D., M.A. Chung, R.C. Holman, RW. Ryder, J.L. Hadler, R.I. Glass. 1999. Use of state
hospital discharge data to assess the morbidity from rotavirus diarrhea and to monitor the impact
of a rotavirus immunization program: a pilot study in Connecticut. Pediatrics. 104(3 Pt. 1):489-
494.
Parasuraman, T. V., K. Frenia, and J. Romero. 2001. Cost of illness and consideration for the economic
evaluation forthe economic evaluation of potential therapies. Pharmacoeconomics. 19(1):3-12.
Parry, S.M. and R. L. Salmon. 1998. Sporadic STEC O157 infection: secondary household transmission
in Wales. Emerging Infectious Diseases. 4(4): 1-6.
Economic Analysis for the R-18 October 2006
Final Ground Water Rule
-------
Parshionikar S.U., S. Willian-True, G.S. Fout, D.E. Robbins, S.A. Seys, J.D. Cassady, and R. Harris.
2003. Waterborne outbreak of gastroenteritis associated with a norovirus. Applied and
Environmental Microbiology . 69(9):5263-5268.
Parsonnet, J, S.C. Track, C.A. Bopp, C.J. Wood, D.G. Addiss, F. Alai, L Gorelkin, N. Hargrett-Bean, R.
A. Gunn, and R.V. Tauxe. 1989. Chronic diarrhea associated with drinking untreated water.
Annals of Internal Medicine. 110(12):985-991.
Perez-Schael et al. 1984. Rotavirus shedding by newborn children. Journal Medical Virology
14(2):127-136.
Pichichero, M. E., S. McLinn, H. A. Rotbart, M. A. Menegus, M. Cascino, B. E. Reidenberg. 1998.
Clinical and economic impact of enterovirus illness in private pediatric practice. Pediatrics.
102(5):1126-1134.
Pickering, L., A. Bartlett, R. Reves, and A. Morrow. 1988. Asymptomatic excretion of rotavirus before
and after rotavirus diarrhea in children in day care centers. Journal of Pediatrics. 1 12(3) 361-5.
Pillai, S. 1997. Virus sampling and microbial analysis at the U.S. -Mexico border for the U.S.
Environmental Protection Agency. Unpublished report for The Cadmus Group, Inc.
Prager, R., S. Annemuller and H. Tschape. 2005. Diversity of virulence patterns among shiga
toxin-producing Escherichia coli from human clinical cases - need for more detailed diagnostics.
International Journal of Medical Microbiology. 295:29-38.
Rangel, J.M., P.H. Sparling, C.Crowe, P.M. Griffin and D. L. Swerdelow. 2005. Epidemiology of the
Escherichia coli O157:H7 outbreaks, United States, 1982-2002. Emerging Infectious Diseases .
Rees, J.R., M.A. Pannier, A. McNees, S. Shallow, F. J. Angulo, and D.J. Vugia. 2004.
Persistent diarrhea, arthritis and other complications of enteric infections: A pilot survey based on
California FoodNet surveillance, 1998-1999. Clinical Infectious Diseases. 38 (Suppl
3):S311-S317.
Regli, S., J.B. Rose, C.N. Haas, and C.P. Gerba. 1991. Modeling the risk from giardia and viruses in
drinking water. Journal of the American Water Works Association. 83(1 1):76-84.
Renken, R.A., K.J. Cunningham, M.R. Zygnerski, M.A. Wacker, A. M. Shapiro, R.W. Harvey, D.W.
Metge, C. L. Osborn, and J.N. Ryan. 2005. Assessing the vulnerability of a municipal well field
to contamination in a karst aquifer. Environmental and Engineering Geoscience.
XI(4):3 19-331.
Research Triangle Institute (RTI). 1997. Valuing drinking water quality: theory, methods, and research
needs. Draft report prepared for the USEPA. April.
Riffard, S., S. Springthorpe, L. Filion and S. Sattar. 2004. Occurrence of Legionella in groundwater.
AWWA Research Foundation Report 90985F: Denver CO, 164 p.
Economic Analysis for the R-19 October 2006
Final Ground Water Rule
-------
Riley, S., C. Fraser, C.A. Donnelly, A.C. Ghani, L.J. Abu-Raddad, A.J. Hedley, G.M. Leung, L.M. Ho,
T.H. Lam, T.Q. Thach, P. Chau, K.P. Chan, S.V. Lo, P.Y. Leung, T. Tsang, W. Ho, K.H. Lee,
E.M. Lau, N.M. Ferguson, and R.M. Anderson. 2003. Transmission dynamics of the etiological
agent of SARS in Hong Kong: impact of public health interventions. Science.
300(5627):1961-6.
Rockx, B., M. de Witt, H. Vennema, J. Vinje, E. de Bruin, Y. van Duynhoven, and M. Koopmans. 2002.
Natural History of Human Calicivirus Infection: A Prospective Cohort Study. Clinical Infectious
Diseases. 35(3):246-253.
Rodriguez, W. J., H.W. Kim, C.D. Brandt, R.H. Yolken, M. Richard, J.O. Arrobio, R.H. Schwartz, A.Z.
Kapikiam, R.M. Chanock, and R.H. Parrott. 1979. Common exposure outbreak of gastroenteritis
due to type 2 rotavirus with high secondary attack rate within families. Journal of Infectious
Diseases. 6:170.
Rodriguez, W.J. H.W. Kim, C.D Brandt, R.H. Scwartz, M.K. Gardner, B. Jeffries, R.H. Parrott, RA.
Kaslow, J.I. Smith, and A.Z. Kapikian. 1987. Longitudinal study of rotavirus infection and
gastroenteritis in families served by a pediatric medical practice: clinical and epidemiologic
observations. The Pediatric Infectious Disease Journal. 6(2): 170-176.
Roivainen, M., M. Knip, H. Hyoty, P. Kulmala, M. Hiltunen, P. Vahasalo, T. Hovi, H.K. Akerblom.
1998. Several different enterovirus serotypes can be associated with prediabetic autoimmune
episodes and onset of overt IDDM. Journal of Medical Virologu. 56:74-48.
Rolle-Kampczyk, U.E., G.J. Fritz, U. Diez, I. Lehman, M. Richter, and O. Herbarth. 2004. Well water -
one source of Helicobacter pylori colonization. International Journal of Hygiene and
Environmental Health. 207:363-368.
Rollins, D.M. and R.Colwell. 1986. Viable but nonculturable stage of Campylobacter jejuni and its role
in survival in the natural aquatic environment. Applied and Environmental Microbiology.
52(3):531-538.
Rose, J. B., R.L. Mullinax, S.N. Singh, M.V. Yates, and C.P. Gerba. 1987. Occurrence of rotaviruses
and enteroviruses in recreational waters of Oak Creek, Arizona. Water Resources. 21(11):1375-
1381.
Rose, J.B. 1997. Environmental ecology of cryptosporidium and public health implications. Annual
Review of Public Health. 18:135-161.
Rotbart, H.A. 1995. Enteroviral infections of the central nervous system. Clinical Infectious Diseases
20:971-981.
R.S. Means. 1998 Mechanical Cost Data. R.S. Means Company, Inc. Kingston, MA.
Rusin, P.A., J.B. Rose, C.N. Haas, and C.P. Gerba. 1997a. Risk assessment of opportunistic bacterial
pathogens in drinking water. Reviews of Environmental Contamination and Toxicology.
152:57-83.
Economic Analysis for the R-20 October 2006
Final Ground Water Rule
-------
Rusin, P.A., J.B. Rose and C.P. Gerba. 1997b. Health significance of pigmented bacteria in drinking
water. Water Science and Technology. 35(ll-12):21-27.
Salvato, J.A. 1992. Environmental Engineering and Sanitation, 4th ed. Hoboken, NJ. John Wiley &
Sons, Inc.
Sawyer, M. H. 2001. Enterovirus infections: diagnosis and treatment. Current Opinions in Pediatrics.
13(l):65-69.
Schaub, S. 2004. A risk assessment framework for waterborne pathogens and requirements for
producing a complete protocol. Human and Ecological Risk Assessment. 10(1):151-158.
Schiff, G.M., G.M. Stefanovic, B.C. Young, D.S. Sander, J.K. Pennekamp, and R.L. Ward. 1984.
Studies of echovirus-12 in volunteers: determination of minimal infectious dose and the effect of
previous infectious dose. Journal oflnfectios Diseases. 150(6):858-866.
Schroeder C.M., J. Meng, Z. Shaohua, et al. 2002. Antimicrobial resistance of Escherichia coli O26,
O103, Oil 1, O128, and O145 from animals and humans. Emerging Infectious Diseases. 8(12).
Schumacher, J. D., C. Chuard, F. Renevey, L. Matter, C. Regamey. 1999. Outbreak of echovirus 30
meningitis in Switzerland. Scandinavian Journal of Infectious Diseases. 31(6):539-542.
Schwab, K. and J. Bae. 2005. Persistence of norovirus and viral surrogates seeded into surface and ground
water. From the Proceedings of the AWWA Water Quality Technology Conference: Baltimore,
MD.
Seunghyun, K., and M.Y. Corapcioglu. 1997. The role of biofilm growth in bacterial-facilitated
contaminant transport in porous media. Transport in Porous Media. 26:161-181.
Shepherd, R. W., S. Truslow, J.A. Walkersmith, R. Bird, W. Cutting, R. Darnell, and C.M. Barker. 1975.
Infantile gastroenteritis - clinical-study of reovirus-like agent infection. Lancet
2(7944): 1082-1084.
Shere JA, K.J. Bartlett, C.W. Kaspar. 1998. Longditudianl study of Escherichia coli O157:H7
dissemination on four dairy farms in Wisconsin. Applied and Environmental Microbiology.
64:1390-1399.
Smith K.E., J.M. Bessner, C.W. Hedberg, et al. 1999. Quinolone-resistant campylobacter jejuni
infections in Minnesota 1992-1998. New England Journal of Medicine. 340:1525-32.
Sobsey, M.D. and J.S. Glass. 1984. Influence of water quality on enteric virus concentration by
microporous filter methods. Applied and Environmental Microbiology. 47:956-960.
Solimena M. and P. De Camilli. 1995. Coxsackievirus and Diabetes. Nature Medicine. l(l):24-25.
Seller, J. A., J. Eisenberg, J. DeGeorge, R. Cooper, G. Tchobanoglous, and A. Olivieri. 2006. A public
health evaluation of recreational water impairment. Journal of Water and Health. 4( 1): 1 -19.
Economic Analysis for the R-21 October 2006
Final Ground Water Rule
-------
Seller, J. A., J.N. Eisenberg, and A.W. Olivieri. 1999. Evaluation of pathogen risk assessment
framework. Prepared by EOAInc. andU.C. Berkeley for ILSI Risk Science Institute.
Seller, J. A., A. Olivieri, J. Crook, R. Parkin, R. Spear, G. Tchobanoglous, and J.N.S. Eisenberg. 2003.
Risk-based approach to evaluate the public health benefit of additional wastewater treatment.
Environmental Science and Technology. 37(9): 1882-1891.
Seller, J. A., A.W. Olivieri, J.N.S. Eisenberg, R. Sakaji, and R. Danielson. 2004. Evaluation of microbial
risk assessment techniques and applications. Water Environment Research Foundation Report,
Project OO-PUM-3.
Statistics Canada. 1999. Average time spent on activities, total population and participants, by sex.
General Social Survey, 1998. http://www.statcan.ca/english/Pgdb/People/Families/famil36a.htm.
as viewed August 2002.
Statistics New Zealand. 1999. Table 19: Average minutes per day spent on each individual activity
counting only primary activities, with subtotals giving the four major types of time, by sex, cross-
classified by family type, for all people aged 12 and over living in private households. New
Zealand Time Use Survey, Ministry of Women's Affairs.
Strikas, R.A., L. J. Anderson, and R.A. Parker. 1986. Temporal and geographic patterns of isolates of
nonpolio enteroviruses in the United States, 1970-1983. Journal of Infectious Diseases.
153(2):346-351.
Sun, J. etal. 1992. Estimating the benefits of groundwater contamination control. Southern Journal of
Agricultural Economics. 24(2): 63-71.
Swerdlow, D.L., B.A. Woodruff, RC. Brady, P.M. Griffin, S. Tippen, H. Donnel, Jr., E. Geldreich, B.J.
Payne, A. Meyer, Jr., J.G. Wells, K.D. Greene, M. Bright, N.H. Bean, and P.A. Blake. 1992. A
waterborne outbreak in Missouri of Escherichia coli O157:H7 associated with bloody diarrhea
and death. Annals of Internal Medicine. 117(10):812-819.
Szucs, G., M. Uj, I. Mihaly, J. Deak. 1999. Burden of human rotavirus-associated hospitalizations in
three geographic regions of Hungary. Acta Paediatrica Supplements. 88(426):61-65.
Taylor, J.W., G.W. Gary, and H.B. Greenberg. 1981. Norwalk-related viral gastroenteritis due to
contaminated drinking water. Am. J. Epi. 114:584-592.
Teunis, P.P.M., O.G. van der Heijden, J.W.B. van der Giessen, and A.H. Havelaar. 1996. The
dose-response relation in human volunteers for gastro-intestinal pathogens. Report no.
284550002.
Teunis, P., A. Havelaar, J. Eisenberg, D. Meuman, and S. Ferenc. 1999. Evaluation of the ILSI risk
assessment framework in assessing risks of waterborne pathogens. Unpublished report to ILSI.
Teunis, P.P.M., W.J. Lodder, S. H. Heisterkamp, A.M. de Roda Husman. 2005. Mixed plaques: Statistical
evidence how plaque assays may underestimate virus concentrations. Water Research
39:4240-4250.
Economic Analysis for the R-22 October 2006
Final Ground Water Rule
-------
Travers K. and M. Barza. 2002. Morbidity of infections caused by antimicrobial-resistant bacteria, In:
The need to improve antimicrobial use in agriculture: Ecological and human health consequences,
M. Barza, S.L. Gorbach (eds.). Clinical Infectious Diseases. 34(S3):S131-S134.
Tucker, A. W., A. C. Haddix, J. S. Bresee et al. 1998. Cost effectiveness analysis of a rotavirus
immunization program for the United States. Journal of the American Medical Association.
279:1371-1376.
Tsang, T.H.F., E.K. Denison, H.V. Williams, L.V. Venczel, M.M. Ginsberg and D. J. Vugia. 2000.
Acute Hepatitis E infection acquired in California. Clinical Infectious Diseases. 30:618-619.
Tufvesson, B., T. Johnsson, and B. Persson. 1977. Family infections by reo-like virus: comparison of
antibody titres by complement fixation and immunoelectroosmophoresis. Scandinavian Journal
of Infectious Diseases. 9(4): 25 7-261.
Turcios, R.M., M. Widdowson, A.C. Sulka, P.S. Mead, and R.I. Glass. 2006. Reevaluation of
epidemiologic criteria for identifying outbreaks of acute gastroenteritis due to norovirus: United
States, 1998-2000. Clinical Infectious Diseases. 42:964-969.
Tyler, K.L., E.S. Barton, M.L. Ibach, C. Robinson, J.A. Campbell, S.M. O'Donnell, T. Valy-Nagy, P.
Clarke, J. D. Wetzel, and T. Dermody. 2004. Isolation and molecular characterization of a novel
type 3 reovirus from a child with meningitis. Journal of Infectious Diseases. 189:1664-1675.
U.S. Bureau of the Census. 1992. Census of Governments, GC92(4)-4:Finaces of Municipal and
Township Governments.
U.S. Bureau of the Census. 1996. National Health History Interview Study; Statistical Abstract of the
U.S.
U.S. Bureau of the Census. 2000. Census, residential population estimates by month and single year of
age. http://www.census.gov/population/estimates/nation/e90s/e9596rmp.txt
U.S. Census Bureau. (2000). "Population Estimates, http://www.census.gov/popest/estimates.php."
U.S. Bureau of the Census. 2001a. Households and Families: 2000. Census 2000 Brief. C2KBR/01-8.
U.S. Bureau of the Census. 2001b. Statistical Abstract of the United States. Tables 194, 567, 582, 621,
and 626.
U.S. Bureau of the Census. 2003. Statistical Abstract of the United States. Tables 641, 602, 646.
U.S. Department of Commerce, Bureau of Economic Analysis. 2004. Table 1.1.6. Real Gross Domestic
Product, Chained Dollars (Billions of chained (2000) dollars).
http://www.bea.doc .gov/bea/dn/nipaweb/SelectTable.asp?Selected=N#Sl
U.S. Department of Energy, Energy Information Administration. 2004a. Table ES1.A. Total Electric
Power Industry Summary Statistics, 2004 and 2003.
http://www.eia.doe.gov/cneaf/electricity/epm/tableesla.html
Economic Analysis for the R-23 October 2006
Final Ground Water Rule
-------
U.S. Department of Energy, Energy Information Administration. 2004b. Table 7.1. Electricity Overview
(Billion Kilowatthours). http://www.eia.doe.gov/emeu/mer/txt/mer7-l
USEPA. 1989. National Primary Drinking Water Regulations; Total Coliforms (Including Fecal
Coliform and E. coli); Final Rule. Federal Register 54(124):27544. June 29, 1989.
USEPA. 1990. Ground Water Volume 1: Ground Water and Contamination. EPA Office of Research
and Development. EPA/625/6-90/016a.
USEPA. 1993. Wellhead Protection: A Guide for Small Communities. Seminar Publication. EPA
Office of Research and Development. EPA/625/R-93/002.
USEPA. 1995a. Guidance for risk characterization. Science Policy Council. February 1995.
USEPA. 1995b. Policy for risk characterization at the U.S. Environmental Protection Agency.
Memorandum. March 21, 1995.
USEPA. 1995c. A Framework for Measuring the Economic Benefits of Ground Water. U.S. EPA Office
of Water and Office of Policy, Planning, and Evaluation. October.
USEPA. 1996a. Amendments to the Safe Drinking Water Act. 63 FR 40585.
USEPA. 1996b. Ground Water Disinfection and Protective Practices in the United States. Prepared by
EPA and Science Applications International Corporation. Office of Ground Water and Drinking
Water, Washington, D.C.
USEPA. 1996c. Economic analysis of Federal regulations under Executive Order 12866. Office of
Management and Budget. January 11, 1996.
USEPA. 1996d. Workshop on Predicting Microbial Contamination of Groundwater Systems, July
10-11, 1996, Proceedings Report. Office of Ground Water and Drinking Water. Washington
DC, September.
USEPA. 1997a. Community Water System Survey (CWSS), Volumes I and II. Office of Water,
Washington, D.C. EPA/815-R-97-001a and -OOlb.
USEPA. 1997b. Policy for use of probabilistic analysis in risk assessment. Office of Research and
Development. May 15, 1997.
USEPA. 1997c. The Benefits and Costs of the Clean Air Act, 1970-1990. Prepared for U.S. Congress.
USEPA. 1998a. Guidance on Implementing the Capacity Development Provisions of the Safe Drinking
Water Act Amendments of 1996.
USEPA. 1998b. "Wisconsin migrant worker camp drinking water quality study." Unpublished report
prepared for US EPA Region V, Safe Drinking Water Branch, July, 1998, p. 10.
Economic Analysis for the R-24 October 2006
Final Ground Water Rule
-------
USEPA. 1998c. "GWR vulnerability assessment study, Aprils, 1998." Unpublished report prepared by
International Consultants, Inc. for the Office of Ground Water and Drinking Water, p. 29.
USEPA. 1998d. National Primary Drinking Water Regulations: Disinfectants and Disinfection
Byproducts; Final Rule. 63 FR 69389, December 16, 1998.
USEPA. 1999a. Drinking Water Criteria Document for Viruses: An Addendum. EPA 822-R-98-042.
January 15, 1999.
USEPA. 1999b. Underground Injection Control Regulations for Class V Injection Wells, Revision; Final
Rule. 64 FR 68545. December 7, 1999.
USEPA. 2000a. Geometries and Characteristics of Water Systems Report (Model Systems Report).
December, 2000.
USEPA. 2000b. Data Reliability Analysis of the EPA SDWIS/FED.
USEPA. 2000c. Estimated Per Capita Water Ingestion in the United States. (Based on Data collected by
the United States Department of Agriculture's 1994-96 Continuing Survey of Food Intakes by
Individuals). EPA Office of Water, Office of Science and Technology. April, 2000
USEPA. 2000d. Handbook for Non-Cancer Health Effects Valuation.
USEPA. 2000e. Guidelines for Preparing Economic Analyses. U.S. EPA Office of the Administrator.
EPA/240-R-00-003. September 2000.
USEPA. 2000f. Development of Cost of Capital Estimates for Public Water Systems, Final Report.
USEPA. 2000g. Health Risks of Enteric Viral Infections in Children. Office of Science and Technology,
Washington, D.C. EPA/822/R/00/010.
USEPA. 2000h. Regulatory Impact Analysis of the Proposed Ground Water Rule.
USEPA. 200 la. Third Edition of the Water Industry Baseline Handbook (Baseline Handbook). May,
2001.
USEPA. 200Ib. Science Advisory Board (SAB) Arsenic review panel report.
USEPA. 2002. Children's Health Valuation Handbook.
USEPA. 2003a. The Safe Drinking Water Information System - Federal Version (SDWIS/FED) data (4th
quarter freeze year 2003 data).
USEPA. 2003b. Labor Costs for National Drinking Water Rules.
USEPA. 2003c. EPA Protocol for Participation in a PWSS Program Data Verification.
Economic Analysis for the R-25 October 2006
Final Ground Water Rule
-------
USEPA. 2004. Developing Dynamic Infection Transmission Models for Microbal Risk Assessment
Applications, EPA-NCEA-C-1463.
USEPA. 2006a. Public Comment and Response Document for the Final Ground Water Rule,
EPA815-R-06-013.
USEPA. 2006b. Occurrence and Monitoring Document for the Final Ground Water Rule,
EPA815-R-06-012.
USEPA. 2006c. Information Collection Request for the National Primary Drinking Water Regulations:
Final Ground Water Rule, EPA 815-R-06-011.
USEPA. 2006d. Technology and Cost Document for the Final Ground Water Rule, EPA 815-R-06-015.
USEPA. 2006e. National Primary Drinking Water Regulations: Stage 2 Disinfectants and Disinfection
Byproducts Rule; Final Rule. 71 FR 388, January 4, 2006. EPA 815-F-05-003.
USEPA and ASDWA. 1995. EPA/State Joint Guidance on Sanitary Surveys.
Van Houtven, G.L., J.C. Whitehead, T.H. Bingham and B. Depro. 1997. Valuing Drinking Water
Benefits: Theory, Methods, and Research Needs. Draft Report.
Varma J.K. K.D. Greene, J. Ovitt, et al. 2005. Hospitalization and antimicrobial resistance in
salmonella outbreaks, 1984-2002. Emerging Infectious Diseases. 11(6).
Vaughn, J.M. 1996. "Sample Analyses." Attachment, unpublished letter on the analysis of alluvial wells
in Missouri by J. Lane and K. Duzan, Missouri Department of Natural Resources, Rolla, MO,
November 7, 1996.
Velazquez, F. R., J.J. Calva, M.L. Guerrero, D. Mass, R.I. Glass, L.K. Pickering, and G. Ruiz-Palacios.
1993. Cohort study of rotavirus serotype patterns in symptomatic and asymptomatic infections in
Mexican children. Pediatric Infectious Disease. 12(1):54-61.
Ventura, S.J, W.D. Mosher, S.C. Curtin, J.C. Abma, and S. Henshaw. 2000. Trends in pregnancies and
pregnancy rates by outcome: Estimates for the United States, 1976-96. Division of Vital and
Health Statistics. 21:56
Vreugdenhil G.R., N.C. Schloot, A. Hoorens. C. Rongen, D. G. Pipeleers, W. J.G. Melchers, B.O.Roep
and J.M.D. Galama. 2000. Acute Onset of Type I Diabetes Mellitus after Severe Echovirus 9
Infection: Putative Pathogenic Pathways. Clinical Infectious Diseases. 31:1025-1031.
Ward, R.L., D.R. Knowlton, and B.J. Pierce. 1984. Efficiency of Human Rotavirus Propagation in Cell
Culture. Journal of Clinical Microbiology. 19(6): 718-753.
Ward, R.L., D.I. Berstein, D.R. Knowlton, J.R. Sherwook, E.G. Young, T.M. Cusack, J.R. Rubino, and
G.M. Schiff 1991. Prevention of surface-to-human transmission of rotaviruses by treatment with
disinfectant spray. Journal of Clinical Microbiology. 29(9): 1991-1996.
Economic Analysis for the R-26 October 2006
Final Ground Water Rule
-------
Ward, R.L. 1986. Human rotavirus studies in volunteers: determination of infectious dose and
serological response to infection. Journal of Infectious Diseases. 154(5):871.
Ward, R. L., and D.I. Bernstein. 1994. Protection against rotavirus disease after natural rotavirus
infection. Journal of Infectious Diseases. 169:900-904.
Washington State Department of Health. 1995. Wellhead Protection Program Guidance Document.
DOH Publication No. 331-018. April.
Wenman, WM, D. Hinde, S. Feltham, and M. Gurwith. 1979. Rotavirus in adults: results of a prospective
family study. 1979. New England Journal of Medicine. 301(6):303- 306.
Widdowson, M., E.H. Cramer, L. Hadley, J.S. Bresee, RS. Beard, S.N. Bulens, M. Charles, W. Chege, E.
Isakbaeva, J.G. Wright, E. Mintz, D. Forney, J. Massey, R.I. Glass, and S.S. Monroe. 2004.
Outbreaks of acute gasteroenteritis on cruise ships and on land: Identification of a predominant
circulating strain of norovirus-United States 2002. Journal of Infectious Diseases. 190:27-36.
Widdowson, M. A., G.J. van Doornum, W.H. van der Poel, A.S. de Boer, R. van de Heide, U. Mahdi, P.
Haanen, J.L. Kool, and M. Koopmans. 2002. An outbreak of diarrhea in a neonatal medium care
unit caused by a novel strain of rotavirus: investigation using both epidemiologic and
microbiological methods . Infection Control and Hospital Epidemiology. 23(11):665-70.
Widdowson, M., S.S. Monroe, and R.I. Glass. 2005. Are Noro viruses emerging?. Emerging Infectious
Diseases. ll(5):735-737.
Widdowson, M., A. Sulka, S.N. Bulens, RS. Beard, S.S. Chaves, R Hammond, E.D.P. Salehi, E.
Swanson, J. Totaro, R. Woron, P.S. Mead, J.S. Bresee, S.S. Monroe, and R.I. Glass. 2005.
Norovirus and foodborne disease, United States, 1991-2000. Emerging Infectious Diseases.
Wilson, R., L.J. Anderson, R.C. Holman, G.W. Gary, and H.B. Greenberg. 1982. Waterborne
gastroenteritis due to the Norwalk Agent: Clinical and epidemiological investigation. American
Journal of Public Health. 72(l):72-74.
Winkelstein, W., D.T. Karzon, A.L. Barren et al. 1957. Epidemiological observations on an outbreak of
aseptic meningitis due to ECHO virus type 6. American Journal of Public Health. 47:741-749.
World Health Organization (WHO). 2003. Hazard characterization for pathogens in food and water:
Guidelines, 61p.
Yanko, W.A., J.L. Jackson, P.P. Williams, A.S. Walker, and M.S. Castillo. 1999. An unexpected
temporal pattern of coliphage isolation in ground waters sampled from wells at varied distance
from reclaimed water recharge sites. Water Resource. 33:53-64
Yates, M., J. Malley, P. Rochelle, R. Hoffman. 2006. Impact of adenovirus resistance on UV
disinfection requirements: report from an expert workshop on the state of the science on
adenoviruses. Journal of the American Water Works Association. Pre-publication draft.
Economic Analysis for the R-27 October 2006
Final Ground Water Rule
-------
Yolken, R.H., C.A. Bishop, T.R. Townsend, E.A. Bolyard, J. Bartlett, G.W. Santos, and R. Sabal. 1982.
Infectious gastroenteritis in bone-marrow transplant recipients. New England Journal of
Medicine. 306(17): 1010-2.
Young B.C. and D.G. Sharp. 1977. Poliovirus aggregates and their survival in water. Applied and
Environmental Microbiology. 33:168-177.
Economic Analysis for the R-28 October 2006
Final Ground Water Rule
------- |