i
-------
-------
Contents
Executive Summary
ES I Protection of Public Health
ES 2 Exposure . .
ES 3 Health Hazards
ES 4 Risk Assessment and Uncertainty
ES 5 Benefits of the IESWTR
ES 6 Compliance Costs and Treatment Effectiveness
ES 7 National Benefits Comparisons
ES 8 Household Cost Comparisons
ES 9 Conclusion
Chapter 1:
i.i
12
1.3
14
15
16
17
Chapter 2:
2.1
2.2
2.3
Chapter 3:
31
32
3.3
Chapter 4:
41
42
43
Chapter 5:
51
52
53
5.4
5.5
56
5.7
5.8
5.9
5.10
1-1
i—I
1-2
1-4
1-5
1-8
1-8
1-9
2-1
2-1
2-2
2-5
3-1
3-I
3-3
3-5
ES-i
ES-i
ES-2
ES-2
. . ES-3
ES-4
. . ES-S
. . ES-6
. ES-6
. ES-7
Introduction and Summary
Introduction
Public Health Concerns Addressed by the IESWTR
Regulatory History
Summary of the Rule
Environmental Justice
Unfunded Mandates Reform Act Analysis
Regulatory Flexibility Analysis
Consideration of Regulatory Alternatives
Chronological Review of Regulatory Options Considered
Summary of Regulatory Alternatives Considered
Cost and Benefit Analyses Conducted
BaselineAnalysis
Industry Profile
Cost Analysis
Benefit Analysis
Benefits A .naIy! 4—1
Introduction 4-I
Health Benefits from Reducing Exposure to Cryp:osporidiwn 4 1
Other Benefits 4-23
CostAnatysis 5-1
Introduction 5-1
Total National Costs of Compliance 5-i
Turbidity Treatment 5-2
Monitoring Individual Filter Turbidity 5 _i 1
State and Utility Turbidity Exceptions Reporting Costs 5-14
Disinfection Profiling and Benchmarking 516
Sanitary Surveys 520
Covered Finished Water Reservoirs 5-23
Household Costs 524
Combined Effect of the Stage 1 DBPR and the IESWTR 5-26
-------
Chaper t et Beneftis . 6-1
Chaper 7. The Economic Rationak for Regulation . 7-1
71 Introduction 7-I
7 2 Statutory Authority for Promulgating the Rule 7-I
7 3 The Economic Rationale for Regulation 7-1
References
-------
Exhibits
Executive Summary ES-1
ESI Summary of Potential Annual Benefits .... . . ES-8
ES 2 Summary of Costs under the Interim Enhanced Surface Water Treatment Rule ES-9
ES 3 Cumulative Distribution of Annual Average Cost per System of the IESWTR
Cumulative Distribution of Annual Cost per Household of the IESWTR . ES-10
ES 4 Characteristics of Surface Water Systems that Use Filtration ES-11
Chapter 3: Baseline Analysis 3-1
3 1 Systems and Plants Using Rapid Granular Filtration per Size of
Population Served . 3-2
Chapter 4: Benefits Analysis 4-1
4.1 Symptoms of 205 Patients with Confirmed Cases of Cryptosporidiosis
during the Milwaukee Outbreak 4-2
4 2 Waterborne Cryptospondium Outbreaks in the U.S. Associated with Drinking
Water by Type of Water Source 4-3
4 3 Summary of Surface Water Monitoring Data for Cryptospondium Oocysts 4-4
4.4 Steps in the Risk Assessment Process for Cryptospondium 4-6
4.5 Baseline Expected National Source Water and Finished Water
Cryptospondium Distributions, Based on Current Treatment 4-9
4.6 Frequency Distribution of Annual Illnesses (Morbidity) Current Treatment
Assumption of 2.5 Log Cryptospondium Removal, without IESWTR ... 4-12
4.7 Frequency Distribution of Annual Illnesses (Morbidity) Current Treatment
Assumption of 3.0 Log Cryptospondium Removal, without IESWTR ... 4-13
4.8 Expected Number of Systems Requiring Additional Treatment if Monthly
Turbidity Standard is Reduced to 0.2 NTU 4-14
4.9 Cumulative Probability Distribution of Aggregate Pilot Plant Data for C. parvum '
Removal 4-15
4 10 Improved Cryptospohdivm Removal Assumptions Additional
Cryptosporidium Log Removal with IESWTR 4-15
4.11 Expected National Source Water and Finished Water Cryptosporidium
Distributions with Improved Removal 4-16
4.12 Number of Infections and Illnesses 4-17
4 13 Losses per Case of Giardiasis by Category 4-19
4.14 Frequency Distribution of Adjusted Cost of Illness Estimate 4-20
4 15 Value of Illnesses Avoided Annually 4-21
4 16 Number of Mortalities among Exposed Population 4-22
4.17 Value of Mortalities Avoided Annually 4-24
Chapter 5: Cost Analysis 5-1
5.1 Summary of Costs under the Interim Enhanced Surface Water Treatment Rule ... 5-3
5 2 Number of Systems Modifying Treatment Practices to Meet Limit 5-5
5.3 Cost Estimates for Alternative Combined Filter Effluent Turbidity Limits 5-7
-------
5 4 > ationaI Cost Estimates for Alternau e \laumum Combined Fitter Effluent
Turbidir Limits 5-9
55 Cost Estimates For Cr.piosporidiuin Inactivation by Ozone 5-to
5 6 Final Annual Cost Estimates for Turbidirv Treatment Reauirernents 5-I I
5 7 Utility Turbidity Start-Up and Monitoring Annual Cost 5-13
58 State Turbidity Start-Up and Monitoring Annual Costs . . . 5-14
5 9 Utility and State Turbidity Exception Costs . 5-16
510 TTHM and HAA5 Data from the Water Industry Database (WIDB) 5-18
S 11 Annual Utility Disinfection Benchmarking Cost Estimates . 5-19
5 12 Annual State Disinfection Benchrnarking Cost Estimates .. 5-20
5(3 Revised Baseline of Systems Based on 1994 Stage I DBPR RIA 5-22
514 Total Start-Up and Annual Cost of Sanitary Surveys Based on Revised Baseline
of Systems 5-22
5 15 Annual Cost of Covered Finished Water Reservoirs . 5-24
5 16 Cumulative Distribution of Annual Cost per Household of the IESWTR 5-26
517 Cost Impact of Current and Expected Rulemakings 5-27
Chapter 6: Net Benefib 6-1
6.1 Summary of Costs under the Interim Enhanced Surface Water Treatment Rule... 6-2
62 Summary of Potential Annual Benefits 6-3
-------
ACKNOWLEDGMENTS
This document v as prepared for the U S Environmental Protection Agency, Office of
Ground Water and Dnnking Water (OGV DW) by Science Application International Corporation
(SAIC) (Contract No. 68-C6-0059) and its subcontractor. The Cadmus Group. Inc. Overall
planning and management for the preparation of this manual was provided by Stig Regli and
Vaterie Blank of OGWDW and Tom Carpei tter of SAIC
EPA acknowledges the valuable contributions of those who wrote and reviewed this
document They include. John Cromwell, James Aibright, Rosemarie Odom and Elena Ryan of
The Cadmus Group, Inc , Tom Carpenter, Mike Lustic and Frank Letkiewicz of SAIC; Enc
Bissonette, Jon Bender, Philip Berger, PhD , Meloñie Williams, PhD, Elizabeth McCIelland.,
PhD . Chris Dockins, PhD , and Rebecca Calderon, PhD., of U.S EPA. EPA also thanks the
following external peer reviewers for their excellent review and valuable comments on the draft
manuscript Charles Abdulla, PhD., (Pennsylvania State University), Gunther Craun (0. Craun
and Associates), and Joseph Eiseriberg, PhD., (University of California at Berkeley).
-------
Executive Summary
ES.l Protection of Public Health
The primary mission of the Environmental Protection Agency (EPA) is to safeguard human health and
the environment This document addresses the expected impacts—both improvements to public health
and the costs to industry and consumers—of one EPA regulation that will make water safer to drink
One of the most difficult challenges facing water systems is reducing the health nsk caused by disease-
causing microbial contaminants (i.e., bacteria, protozoa, and viruses). Many water systems treat their
water with a chemical disinfectant to prevent diseases from microbial contaminants. Disinfection,
however, may pose risks of its own. Disinfectants and their byproducts have been associated with
potential health risks that include cancer and reproductive and developmental effects. EPA has identified
'ways to significantly lessen the potential risks associated with microbial contaminants without increasing
the use and potential risks posed by disinfectants at reasonable costs. To implement these changes, EPA
is publishing a final Interim Enhanced Surface Water Treatment Rule (IESWTR) that contains the new
requirements for water systems and this Regulatory Impact Analysis (RIA), which documents the costs
and benefits of the rule.
The primary goal of the IESWTR is to improve public health by increasing the level of protection from
exposure to Cryptosporidium and other pathogens in dcinking water supplies. The Safe Drinking Water
Act (SDWA) requires the setting of drinking water standards at contaminant levels designed to avoid
adverse effects on health while allowing for a margin of safety. The rule is expected to reduce the level
of Cryptosporidium and other pathogen contamination in finished drinking water supplies through
improvements in filtration at water systems. The rule is also expected to provide a larger margin of
safety, particularly by reducing the likelihood of the occurrence of Cryptosporidium outbreaks.
In the classic paradigm of public health decision-making, it is necessary to decide upon a prudent course
of action despite confounding factors. The decision process consists of weighing available evidence to
gain as much insight as possible into expected or possible health outcomes while also weighing the costs
and technological realities of available responses. At one end of the spectrum, a "No Action" option
might be justified when the balance of health evidence suggests low exposure, low probability, and low
severity while the response technologies imply high costs and limited effectiveness. At the opposite
extreme, urgent and forceful action might be warranted when the health evidence suggests high
exposure, high probability, and high severity while the response technologies have modest costs and
good effectiveness. Based on the risk assessment presented in this RIA, EPA believes that there is
sufficient exposure, probability, and severity on the health side to warrant a public health decision to
accept the cost and technology impacts of the IESWTR in order to obtain the projected exposure
reduction. Highlights of this balancing analysis are summarized in the following discussion.
IESWTR Fined RIA ES-f November 12, 1998
-------
ES.2 Exposure
Exposure to Crypiosporidium is potentially quite large The presence of Crypwspor:dium in surface
water sources is common, as oocySts have been found in wastewater, pristine surface water, surface
water receiving agricultural runoff, water for recreational use, and drinking water The over-139 million
people in the U S served by utilities covered by the major provisions of the IESWTR are potentially at
risk from exposure to Crypiosporidium and other microbial contaminants.
ES.3 Health Hazards
In 1990, EPA’s Science Advisory Board (SAB), an independent panel of experts established by
Congress, cited drinking water contamination as one of the most important environmental risks and
indicated that disease-causing microbial contaminants (i.e., bacteria, protozoa and viruses) are probably
the greatest remaining health risk management challenge for drinking water suppliers (EPA/SAB, 1990)
Information on the number of waterborne disease outbreaks from the U.S. Centers for Disease Control
(CDC) underscores this concern. CDC indicates that, between 1980 and 1996, 401 waterborne disease
outbreaks were reported, with over 750,000 associated cases of disease. During this period, a number of
agents were implicated as the cause, including protozoa, viruses and bacteria, as well as several
chemicals. Most of the cases (but not outbreaks) were associated with surface water, and specifically
with a single outbreak of cryptosporidiosis in Milwaukee (over 400,000 cases) (MacKenzie, et al., 1994)
It is important to note that for a number of reasons, the CDC reports may substantially understate the
actual number of waterborne disease outbreaks and cases in the U.S. First, few States have an active
outbreak surveillance program. Second, disease outbreaks are often not recognized in a community or, if
recognized, are not traced to the drinking water source. Third, a large number of people experiencing
gastrointestinal illness (predominantly diarrhea) do not seek medical attention. Fourth, physicians may
often not have a broad enough community-wide basis of information to attribute gastrointestinal illness
to y specific origin such as a drinking water source. Finally, an unknown but probably significant
portion of waterbome disease is endemic (i.e., not associated with an outbreak), and thus is even more
difficult to recognize.
Waterborne disease is usually acute (i.e., sudden onset and typically lasting a short time in healthy
people). Some pathogens (e.g., Giardia, Crypto .sporidium) may cause extended illness, sometimes lasting
months or longer, in otherwise healthy individuals. Most waterborne pathogens cause gastrointestinal
illness, with diarrhea, abdominal discomfort, nausea, vomiting, and/or other symptoms. Other
waterborne pathogens cause, or at least are associated with, more serious disorders such as hepatitis,
gastric cancer, peptic ulcers, myocarditis, swollen lymph glands, meningitis, encephalitis, and a myriad
of other diseases.
Gastrointestinal illness may be chronic in vulnerable populations (e.g., immunocompromised
individuals). The severity and duration of illness is often greater in immunocompromised persons than in
healthy individuals and may be fatal among this population. For instance, a follow-up study of the 1993
Milwaukee waterborne disease outbreak reported that at least 50 Ci ptosporidiwn-associated deaths
occurred among the severely immunocompromised (Hoxie, et al., 1997). Immunocompromised persons
include infants, pregnant women, the elderly, and especially those with severely weakened immune
systems (e.g., AIDS patients, those receiving treatment for certain types of cancer, organ-transplant
recipients and people on immunosuppressant drugs) (Gerba et al., 1996).
!ESWTR Final RJA FS-2 November 12. 1998
-------
ith speciñc retererice to cr.ptosporidiosis. the disease is caused b rngestion of envirorimentalk
resistant Cr pi’ospor:dwni ooc’ .sts that are readily carried by the waterborne route Both human and other
animals may excrete these oocysts Transmission of this disease often occurs through ingestion of the
infective oocysts from contaminated water or food, but may also result from direct or indirect contact
with infected persons or animals (Casemore and Jackson, 1983, Cordell and Addiss, 1994) Symptoms of
cryptosporidiosis include typical gastrointestinal symptoms (Current, et al, 1983), and as noted above,
these may persist for several days to several months.
While cryposporidiosis is generally a self-limiting disease with a complete recovery in otherwise
healthy persons, it can be very serious in immunosuppressed persons EPA has a particular concern
regarding drinking water exposure to Cryprosporidiurn, especially in severely immunocompromised
persons, because there is no effective therapeutic drug to cure the disease. There have been a number of
waterborne disease outbreaks caused by Cryptosporrdium in the U. S, United Kingdom and many other
countries (Rose, 1997). There appears to be an immune response to Cryptosporidium, but it is not known
if this results in protection (Fayer and Ungar, 1986).
One of the key regulations EPA has developed and implemented to counter pathogens in drinking water
is the Surface Water Treatment Rule (SWTR). Among its provisions, the rule requires that a surface
water system have sufficient treatment to reduce the source water concentration of G:ardia and viruses
by at least 99.9 percent (3 log) and 99.99 percent (4 log), respectively. A shortcoming of the SWTR is
that the rule does not specifically control for the protozoan Cryptosporidium. The first report of a
recognized outbreak caused by Cryptosporidiwn was published during the development of the SWTR
(D’Antonio, et at., 1985). A particular public health challenge is that simply increasing existing
disinfection levels above those most commonly practiced in the United States today does not appear to
be an effective strategy for controlling Cryptospor:dium, because the oocyst is especially resistant to
disinfection.
In terms of occurrence, Cryptosporidnim is common in the environment. Runoff from unprotected
watersheds allows transport of these microorganisms to water bodies used as intake sites for drinking
water treatment plants. One of the particular challenges of Cryptosporidium is its resistance to
disinfection practices used at water treatment plants. Today’s rule addresses the concern of passage of
Cryptospor:diwn through physical removal processes during water treatment It also strengthens the
effectiveness and reliability of physical removal for particulate matter and microorganisms in general,
thereby reducing the likelihood of the disinfection banier being over-challenged. Waterborne disease
outbreaks have been associated with a high level of particles passing through a water treatment plant
(Fox, et al., 1996). This presents a significant public health concern. Hence, there is a need to optimize
treatment reliability and to enhance physical removal efficiencies to minimize the Crypt ospor;dium
levels in finished water. This rule, with tightened turbidity performance criteria and required individual
filter monitoring, is formulated to address these public health concerns.
ES.4 Risk Assessment and Uncertainty
As with other microbial contaminants, there are two ways to characterize the risk posed by
cryptospondiosis: I)endemic risk of illness resulting from everyday low-level exposure to the small
percentage of oocysts that might pass through treatment processes without being inactivated; and 2)
epidemic risk of illness resulting from large numbers of viable oocysts that pass through treatment
processes during some sort of non-routine failure or upset of the treatment plant The extent of current
IFS tTR Final PJA ES-i November 12. 1998
-------
ntormation. kno led e. and uncertainty falls in an uneven pattern across these two approaches to
anal s is
Endemic analysis requires knowledge of the occurrence of oocysts in raw water, the efficacy of treatment
processes in reducing concentrations of viable oocysts, and the dose-response relationship applicable to
humans Enough is known about each of these variables to perform risk assessment, but each factor
contributes variability to the result The existence of endemic risk has been investigated at an individual
water system level with epidemiological studies and some corroboration of the risk assessment
methodology has been established, but there are still broad uncertainty bounds associated with these
attempts at calibration.
Epidemic disease incidence is often reported to the CDC, but the reporting system is believed to be
affected by under-reporting. There is no reliable means of projecting the total incidence of outbreaks
from these data. In addition, there is no simple way to predict the likelihood of future outbreaks that may
be caused by uncommon combinations of natural and human events.
This RIA presents a quantitative risk assessment only for endemic incidence of cryptosporidiosis. In this
analysis, there is uncertainty associated with several key points, requiring assumptions and sensitivity
analyses to quantify risk. The result is a broad range of answers. Assuming a baseline 2.5 log removal of
Cryptosporidium for current treatment, this RIA estimates an expected value (mean) of 1,503,000
cryptosporidiosis endemic infections per year resulting in 643,000 illnesses from exposure to drinking
water supplies in the water systems that will require changes under the rule. The 90 percent confidence
range of this estimate extends from a low of 8,000 to a high of 1,241,000 illnesses per year. Under the
comparison assumption of a 3.0 log removal, this RIA estimates an average of 208,500 illnesses, with a
90 percent confidence range of 2,500 to 384,500.
ES.5 Benefits of the LESWTR
According to the risk assessment performed for this RIA, the IESWTR is estimated to reduce the mean
annual number ot illnesses caused by Cryptosporidium in water systems improving filtration by 110,000
to 463,000 cases depending on which of the six scenarios describing baseline removal (2.5 and 3.0 log)
and improved Crypiosporidiwn removal (low-, mid-, and high-improved) is assumed. Based on these
values, the mean estimated annual benefits of reducing the illness ranges from SO.263 billion to SI .240
billion per year. This calculation is based on a valuation of $2,000 per incidence of cryptosporidiosis
prevented, which is the mean of a distribution of values ascribed to health damages avoided.
The risk assessment also indicated that the rule could result in a mean reduction of 14 to 64 fatalities
each year, depending on varied baseline and removal assumptions. Using a mean value of $5.6 million
per statistical life saved, reducing these fatalities could produce benefits in the range of 50.085 billion to
S0.363 billion.
In addition, benefits would accrue from the implementation of the rule in the form of reduced risk of
outbreaks and consequent epidemic illness, enhanced aesthetic water quality, avoided co ts of averting
behavior, and reduced risk from other pathogens, such as Giardia lamblia. Benefits for these categories
were not quantified for this analysis.
JESJVTR Final PJA November 12, 1998
-------
ES.6 Compli ince Costs and Treatment Effectiveness
The total annual cost of the IESWTR is estimated at $307 million (Exhibit ES I) using a 7 percent cost
of capital Utilities incur 95 percent of this annual cost ($291 million), and States incur the remaining 5
percent (about $16 million) The rule elements that most significantly influence the total cost of the
IESWTR include the cost (0 build and maintain new or advanced treatment facilities and the cost to
monitor the performance of systems. EPA estimates that the total capital cost nationwide would be $759
million Total capital costs are those costs associated with the purchase of equipment or systems that will
meet the treatment requirements The largest capital expenditures are associated with installing
individual filter turbidirneters and making hydraulic improvements to account for recycle flow in process
control decisions. These costs are typically one-time investments. To make the costs comparable with
implementation costs that occur each year, these capital or treatment costs are multiplied by a factor that
“annualizes” the total, thus allowing all rule costs incurred in a year to be summed. To operate and
maintain this capital investment will require about $106 million annually. The annual treatment costs
(annualized capital costs and operating and maintenance (O&M) costs) are $192 million (at a 7 percent
cost of capital). The cost to monitor the performance of systems in terms of turbidity is the other major
cost of the rule. Turbidity monitoring is projected to cost utilities about $96 million annually.
The remaining costs ($19 million annually, about 6 percent of the total) include some other costs to
utilities and all of the costs to States. Utilities will also provide reports and respond when filter
performance falls below expectations ($0.20 million annually); establish disinfection benchmarks ($2.80
million annually); and incur one-time start-up costs for monitoring turbidity and haloacetic acid (HA.A5)
benchmarkirig monitoring ($0.65 million annualized).
Annual State COStS are projected at $15 million. Almost all of this cost (96 percent) is for activities
relating to three requirements: turbidity monitoring, sanitary surveys, and disinfection benchmarking.
The remaining 4 percent of State costs are to start up various parts of the program and to implement the
exception reporting process. Detailed tables for treatment Costs, utility costs, and state costs at the 3, 7,
and 10 percent cost of capital rates may be found in Appendices A through E.
Average annual cost per system (large surf ce water systems that filter using rapid granular filtration) are
displayed in Exhibit ES.3. Because each system will implement one or more treatment techniques
depending on its current water quality characteristics, all affected systems will incur different annual
costs under the LESWTR. Additionally, while 691systems will have to modify their treatment techniques
to meet the turbidity requirements, 1,381 large surface water systems will have to monitor for turbidity
and report turbidity exceptions. Thus, 691 systems will incur both treatment and monitoring costs, and
690 s-ystems will incur only monitoring costs. It is important to note that the cost estimates used for this
exhibit are the average cost per system. Within any one size category, systems may use different
treatment techniques with widely varying costs in many different combinations to treat their water. The
average cost per system gives a good approximation of the most likely costs these systems are expected
to incur under the rule. Under this IESWTR, approximately 50 percent of systems are expected to face an
1 Estimated costs are annualized using a range of razes for the cost of capital over 20 ye The 1994 proposed rule
used a 10 percent cost of capital to annualize. To assist the M-DBP Committee in compar ing revised costs, thN tO percent rate is
currently used where appropnaxe and for comparison. The Office of Management and Budget (0MB) recommends that? pcrcan
be used to annualize capital costs. To reflect this recommendation, costs based at the 7 pèwe4 rate are discussed and used
throughout this RJA. in addition, a 3 percent cost of capital, which is used as a sensitivity analysis, is presented in Exhibit I I
and in Appendices B and E
IESF$ TR Final RIA N embe, 12, 1998
-------
a erage annual cost of less (han S 130,000 The highest annual average cost is 53 million, estimated for 4
systems in the largest population size category
ES.7 National Benefits Comparisons
Given the benefits and costs summarized in Exhibits ES I and ES 2, the IESWTR results in positive net
benefits, assuming a mean number of illnesses avoided, under all three improved removal scenarios (low,
mid, and high) assuming that current treatment achieves a removal of 2 5 logs, taking into account only
the cost of endemic illnesses avoided. Using a current treatment removal assumption of 3 0 logs, net
benefits are positive under the high and mid improved removal scenario, but are negative under the low
improved removal assumption using only the cost of endemic illnesses avoided. When the value of
endemic mortalities prevented is added into the benefits, however, all scenarios have positive net
benefits at the mean.
Thus, the monetized net benefits are positive across the range of current treatment assumptions,
improved log removal scenarios, and cost of capital rates at the mean. The benefits due to the endemic
illnesses avoided may be slightly overstated because the mortalities were not netted out of the number of
illnesses. This value is minimal and would not be captured at the level of significance of the analysis.
Several categories of benefits, including reducing the risk of outbreaks, reducing exposure to other
pathogens such as Giardia, and avoiding the cost of averting behavior have not been quantified for this
analysis, but could represent substantial additional economic value. In addition, the estimates for avoided
costs of endemic illness do not include the value for pain and suffering or the risk premium.
These results indicate that the rule is consistent with the SDWA’s focus on avoiding adverse health
impacts while allowing for a margin of safety, with reasonable assurance that the benefits of the rule will
outweigh the costs.
ES.8 Household Cost Comparisons
Another intuitive measure of the cost-effectiveness and public health benefit of the IESWTR is provided
by computation of the household cost of compliance (Exhibits ES.3). A large number (92 percent) of
households will face a maximum increase in cost of $12 per year ($1 per month). In other words, 60
million households will incur no more than a $1 increase in their monthly costs. Five million households
(7 percent) will face an increase in cost of between S12 and $60 per year (S1-$5 per month). The highest
cost faced by 23,000 households is approximately $100 per year ($8 per month).
Taking the $1 per month figure as a measure of implied public health benefit at the household level, it is
useful to ask what benefits can be identified that could balance a $1 per month expenditure. First, it is
entirely possible that there is much more than a dollar-a-month’s worth of tangible health benefit based
on reduced risk of cryptosporidiosis alone. Second, the broad exposure to microbial pathogens and the
myriad possible health effects involved offer the possibility that there are significant additional health
benefits of a tangible nature. Finally, however, the preventive weighing and balancing of public health
protection provides also a margin of safety—a hedge against uncertainties. Recent survey research
conducted in the drinking water field provides compelling empirical evidence that the number one
priority of water system customers is the safety of their water. Although definitive economic research
has not been performed to investigate the extent of household willingness-to-pay for such a margin of
safety, there is very strong evidence from conventional customer survey research implying a demand for
this benefit.
IFSfl TR Final RJA November 12, 1998
-------
ES.9 Conclusion
In the final analysis, the various benefit/cost comparisons developed in this RJA are quite useful in
assisting the balancing arid weighing analyses that must be performed to support public health decisio-
making Based on a careful weighing of the projected costs against the potential quantified and non-
quantified benefits, EPA has determined that the benefits of the rule Justif i its costs.
IESWTR Final RJA ES -7 November 12. 1998
-------
Exhibit ES.3
Cumulative Distribution of Annual Average Cost per System of the IESWTR
100%
90%
80%
70%
C ’,
I’w
o uv,a
40%
20%
10%
0%
Cumulative Distribution of Annual Cost per Household of the IESWTR
month
(92nd percendls)
$0 $500 $1000 $1,500 $2,000 $2,500 $3,000 $3500
Annual Average Coat par System ($000)
100%
90%
80%
70%
C.
60%
50%
40%
U
0%
$20 $40 $60 $80 $100 $120
Annual Cost par house hold
1ESP TR Final 1A
ES-JO
November 12, 1998
-------
Exhibit ES.4
Characteristics of Surface Water Systems that Use Filtration
Numb rof
S .stem Size Systems to Number of
(population Number of Number of : Modify Systems to Number of
served) Systems Plants Treatment Monitor Only Households
< 10K 4,880
4 880
0
4,880 j 4,122,000
IOK-25K : 94
594
303
291 4,553,000
25K .S OK 316
316
161
155
5,767,000
SOK -75K
124
124 63
61
3,983,000
75K-lOOK
52
104
27
25
2,467,000
10 0K-500K
259
518
122
137
25,524,000
500K-tM
26
52
I I
15 j 12,414,000
> IM
10
20
4
6
10,515,000
Total
6,261 J 6,608
691 5,570
69,345,000
IESWTR Final RIA November 12. 1998
-------
1: Introduction
1.1 Introduction
This document analyzes the impacts of the final Interim Enhanced Surface Water Treatment Rule
(IESWTR). Executive Order 12866, Regulatory Planning and Review, requires EPA to estimate the costs
and benefits of the IESWTR in a regulatory impact analysis (RIA) and to submit the analysis in
conjunction with publishing the final rule.
The IESWTR applies to public drinking water systems using surface water or ground water under the
direct influence of surface water (GWUDI) as a souree, using rapid granular filtration as a treatment
technology, and serving 10,000 or more persons, with the exception of a provision that States perform a
sanitary survey for all surface and OWL DI systems. It builds on the 1989 Surface Water Treatment Rule
(SWTR) and will improve control of microbial pathogens such as Cryptosporidiwn as well as assure
there will be no significant increase in microbial risk for those systems that may need to change their
disinfection practices in order to meet new disinfection byproduct (DBP) standards under the Stage I
Disinfectants/Disinfection Byproducts Rule (Stage I DBPR).
This RIA provides background on the rule, summarizes the key components, discusses alternatives to the
rule, and estimates costs and benefits to the public and State governments. This chapter summarizes the
technical and regulatory issues associated with the rule. It explains the nature of microbial
contamination, reviews the potential health effects of exposure to microbial pathogens, details how the
final IESWTR will address the health effects, and then summarizes the estimated costs and benefits of
rule implementation. In addition, this section includes a statement addressing the potential
disproportionate impact of the rule on low-income or minority communities.
Subsequent chapters are intended to meet the requirements of the Executive Order by responding to
specific analytical questions. Chapter 2 reviews alternative approaches considered as the rule was being
developed. Chapter 3 presents utility data and discusses the changes systems would have to make as a
result of the rule; this approach will establish a baseline for use in the following three chapters. Chapter 4
examines the rule’s potential benefits through the development of a risk assessment. Chapter 5 presents
an estimate of the costs to implement the rule. Chapter 6 provides a comparison of estimated costs and
benefits and summarizes the results of this RIA. Chapter 7 examines the economic rationale for
regulating microbial contaminants.
The IESWTR will be followed by additional rules to improve microbial protection for public water
systems that use surface water and address risk-risk trade-offs with disinfection byproducts. These
include
I) The Long-Term I Enhanced Surface Water Treatment Rule (LT1ESWTR), which will primarily
address public water systems serving fewer than 10,000 people, to be promulgated in November
2000;
IESWTR Final RJA 1-! November 12. 1998
-------
.2) The Long-Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR), which w lI be
promulgated simultaneously with the Stage 2 DBPR in May 2002, and,
3) The Filter Backwash Recycling Rule, to be promulgated in August 2000.
1.2 Public Health Concerns Addressed by the IESWTR
In 1990, EPA’ s Science Advisory Board (SAB), an independent panel of experts established by
Congress, cited dnnking water contamination as one of the most important environmental risks and
indicated that disease-causing microbial contaminants (i.e., bacteria, protozoa and viruses) are probably
the greatest remaining health risk management challenge for drinking water suppliers (EPA/SAB, 1990).
Information on the number of waterborne disease outbreaks from the U.S. Centers for Disease Control
(CDC) underscores this concern. CDC indicates that, between 1980 and 1996, 401 waterborne disease
outbreaks were reported, with over 750000 associated cases of disease. During this period, a number of
agents were implicated as the cause, including protozoa, viruses and bacteria, as well as several
chemicals. Most of the cases (but not outbreaks) were associated with surface water, and specifically
with a single outbreak of cryptospondiosis in Milwaukee (over 400,000 cases) (MacKenzie, et al., 1994).
It is important to note that for a number of reasons, the CDC reports may substantially understate the
actual number of waterborne disease outbreaks and cases in the U.S. First, few States have an active
outbreak surveillance program. Second, disease outbreaks are often not recognized in a community or, if
recognized, are not traced to the drinking water source. Third, a large number of people experiencing
gastrointestinal illness (predominantly diarrhea) do not seek medical attention. Fourth, physicians may
often not have a broad enough community-wide basis of information to attribute gastrointestinal illness
to any specific origin such as a drinking water source. Finally, an unknown but probably significant
portion of waterborne disease is endemic (i.e., not associated with an outbreak), and thus is even more
difficult to recognize.
Waterborne disease is usually acute (i.e., sudden onset and typically lasting a short time in healthy
people). Some pathogens (e.g., Giardia, Cryptosporidiwn) may cause extended illness, sometimes lasting
months or longer, in otherwise healthy individuals. Most waterborne pathogcns cause gastrointestinal
illness, with diarrhea, abdominal discomfort, nausea, vomiting, and/or other symptoms. Other
waterborne pathogens cause, or at least are associated with, more serious disorders such as hepatitis,
gastric cancer, peptic ulcers, myocarditis, swollen lymph glands, meningitis, encephalitis, and a myriad
of other diseases.
Gastrointestmal illness may be chronic in vulnerable populations (e.g., imnunocompromised
individuals). The seventy and duration of illness is often greater in immunocomprom sed persons than in
healthy individuals and may be fatal among this population. For instance, a follow-up study of the 1993
Milwaukee waterborne disease outbreak reported that at least 50 Cryptosporidium-associated deaths
occurred among the severely immunocompromised (Hoxie, et at, 1997). Jmmunocompromised persons
include infants, pregnant women, the elderly, and especially those with severely weakened immune
systems (e.g., AIDS patients, those receiving treatment for certain types of cancer, organ-transplant
recipients and people on immunosuppressant drugs) (Gerba et aL, 1996).
With specific reference to cryptosporidiosis, the disease is caused by ingestion of environmentally
resistant Cryptosporidium oocysts that are readily carried by the waterborne route. Both human and other
animals may excrete these oocysts. Transmission of this disease often occurs through ingestion of the
IESWTR Final S JA 1-2 November 12, 1998
-------
it ecci e ooc’ sts from cornaminaced water or food. but may also result from direct or indirect contact
ith infected persons or animals (Casemore and Jackson. 1983. Corde!l and Addiss, 1994) Symptoms of
cr ptosportdiosis inc’ude typical gastrointestinal symptoms (Current, et al. 1983), and as noted above,
these may persist for several days o several months
While crypiospondiosis is generally a self-limiting disease with a complete recovery in otherwise
healthy persons, it can be very serious in immunosuppressed persons. EPA has a particular concern
regarding drinking water exposure to Crypzospor:dwm, especially in severely immunocompromised
persons, because there is no effective therapeutic drug to cure the disease. There have been a number of
waterborne disease outbreaks caused by Cryptosporidium in the U. S , United Kingdom and many other
countries (Rose, 1997). There appears to be an immune response to Cryptosporidium, but ii is not known
if this results in protection (Fayer and Ungar, 1986).
One of the key regulations EPA has developed and implemented to counter pathogens in drinking water
is the Surface Water Treatment Rule (SWTR). Among its provisions, the rule requires that a surface
water system have sufficient treatment to reduce the source water concentration of G:ardia and viruses
by at least 99.9 percent (3 log) and 99.99 percent (4 log), respectively. A shortcoming of the SWTR is
that the rule does not specifically control for the protozoan Crypiosporidium. The first report of a
recognized outbreak caused by Crypiosporidium was published during the development of the SWTR
(D’Aritonio, et al., 1985). A particular public health challenge is that simply increasing existing
disinfection levels above those most commonly practiced in the United States today does not appear to
be an effective strategy for controlling Crypiosporidium, because the oocyst is especially resistant to
disinfection.
In addition to these microbial issues, there is another potentially confounding public health concern. The
disinfectants used to control pathogens may produce toxic or carcinogenic disinfection byproducts
(DBPs) when they react with organic chemicals in the source water. An important question facing water
supply professionals is how to minimize the risk from both microbial pathogens and DBPs
simultaneously.
At the time the SWTR was promulgated, EPA had limited data concerning Giardia and Cr,ptospondium
occurrence in source waters and treatment efficiencies. The 3-log rernovalJinactivation of Giardia
lamblia and 4-log removal/inactivation of enteric viruses required by the SWTR were developed to
provide protection from most pathogens in source waters. However, additional data have become
available since promulgation of the S’WTR. concerning source water occurrence and treatment
efficiencies for G,ardia as well as for Crypt osporidiurn (LeChevallier, et al., 1991 a,b). A major concern
is that if systems currently provide four or more logs of removal/inactivation for Giardia, such systems
might reduce existing levels of disinfection to more easily meet the new DBP regulations, and thus only
marginally meet the three-log removal/inactivation requirement for Giardia lamblia specified in the
current SWTR. Depending upon source water Gzardia concentrations, such treatment changes could lead
to significant increases in microbial risk (Regli, et al., 1993; Grubbs, et al., 1992; EPA, 1994). As
discussed below, the disinfection benchmarking required under today’s rule is specifically designed as a
process by which a utility and the State, working together, assure that there will be no significant
reduction in microbial protection as the result of modifying disinfection practices in order to meet
maximum contaminant level goals (MCLs) for Total Trihalomethanes (TTHM) and five Haloacetic
Acids (l-IAAS) under the Stage I DBPR..
lES VTR Final RJA 1-3 November 12. 1998
-------
1.3 Regulatory History
The primary responsibility for regulating the quality of drinking water lies with EPA. The Safe Drinking
Water Act (SDWA) establishes this responsibility and defines the mechanisms at the Agency’s disposal
to protect public health EPA sets water quality standards by identifying which contaminants should be
regulated. arid establishes levels of contaminant reduction to be attained by utilities.
To regulate a contaminant, EPA first establishes a maximum contaminant level goal (MCLG) that
establishes the contaminant level at which no known or anticipated adverse health effects occur. MCLGs
are unenforceable health goals. EPA then sets an enforceable maximum contaminant level (MCL) as
close as technologically possible to the MCLG. If it is not feasible to measure the contaminant, a
treatment technique is specified.
For utilities, compliance with a regulation means not exceeding the MCL. However, when M Ls are not
economically or technologically feasible, an approved treatment technique can be used. A treatment
technique requirement is a regulatory approach that specifies a technology that reduces exposure to
contaminants to the extent feasible.
As described earlier, one of the key regulations EPA has developed and implemented to counter
pathogens in drinking water is the 19 9 SWTR. Among its provisions, the rule requires that a utility have
sufficient treatment to reduce the source water concentration of Giardia lamblia and viruses by at least
99 9 percent and 99.99 percent, respectively. The SWTR has several shortcomings, including not
specifically controlling for the protozoan Cryprosporidium. Also, the disinfectants used to control
pathogens may either be toxic or carcinogenic directly, or produce toxic or carcinogenic DBPs when they
react with organic chemicals in the source water. An important question facing water supply
professionals is how to minimize the risk from both microbial pathogens and DBPs simultaneously.
To address the complex issues associated with regulating microbial pathogens, EPA launched a rule-
making process in 1992 and convened a Regulatory Negotiation (RegNeg) Advisory Committee under
the Federal Advisory Committee Act (FACA), representing a range of stakehoiders affected by possible
regulation. The RegNeg Committee met repeatedly over a period of 10 months and arrived at a
consensus proposal for taking progressive steps toward addressing both DBPs and microbial pathogens.
The 1992 consensus-building process resulted in the three following regulatory proposals—
1) A staged approach to regulation of DBPs (referred to as the Stage I and Stage 2
DBPRs) incorporating MCLs, MRDLs, and treatment technique requirements;
2) A companion Interim Enhanced Surface Water Treatment Rule (IESWTR) designed to improve
control of microbial pathogens and prevent inadvertent reductions in microbial safety as a result
of DBP control efforts, and;
3) An Information Collection Rule (ICR) to collect information necessary to reduce many key
uncertainties prior to subsequent negotiations regarding the Stage 2 rule requirements.
Congress amended the SDWA in 1996 and affirmed the strategy developed by the RegNeg Committee.
Congress also established a series of new statutory deadlines for the rules. Under the new amendments,
the IESWTR and the Stage I DBPR must both be promulgated by November 1998. The Filter Backwash
Recycle Rule and the Long Term I Enhanced Surface Water Treatment Rule (LTI) are required to be
IESW7’R Final RIA 1-4 November 12. 1998
-------
promulgated b August 2000 a d November 2000, respectively EP8 must promulgate the Stage 2
DBPR b May 2002 In addition, the Agency will promulgate a Final Ground Water Rule by November
2000 and a Long Term 2 ESWTR (LT2) to accompany the Stage 2 DBPR by May 2002
In 1997, a similar FACA process was implemented with the Microbial-DisinfectantsIDisinfection
Byproducts (M-DBP) Advisory Committee. The M-DBP Committee convened to collect, share, and
analyze new information available since 1994, review previous assumptions made during the RegNeg
process, as well as build consensus on the regulatory implications of this new information. The
Committee made recommendations to EPA including the following, performing benchmarking to
provide a methodology and process by which a utility and the State, working together, assure that there
will be no significant reduction in microbial protection as the result of modifying disinfection practices
inorder to meet MCLs for Total Trihalomethanes (1THMs) and 5 haloacetic acids (HAAS); turbidity; -
Cryptosporidwrn MCLG; Crypzosporidiurn removal requirements; and sanitary surveys.
1.4 Summary of the Rule
The IESWTR is intended to improve control of pathogens such as Crypsosporidiwn as well as assure no
significant increase in microbial risk as systems act to meet the new DBP MCLs under the Stage I
DBPR. With the exception of a requirement that States conduct a sanitary survey for all surface water
and GWUDI systems, the IESWTR applies only to public drinking water systems, using surface water or
GWUDI as a source and serving 10,000 or more people.
Major features of the rule include an MCLG of zero for Cryptcsporidium, limitations on turbidity, a
disinfection benchmark and, sanitary survey provisions. In addition, the rule adds Crypt osporidium to the
definition of GWUDI and to watershed control requirements for unfiltered systems, as well as requiring
that newly constructed finished water reservoirs be covered.
Cryptosporidium
The rule sets the MCLG for Crypiosporidiwn at zero. All surface water systems that serve 10,000 or
more people and are required to filter under the SWTR must remove at least 99 percent of influent
Crypsosporadium (referred to as achieving 2 Log removal). Systems that use rapid granular filtration
(direct filtration or conventional filtration treatment) and meet the turbidity requirements contained in the
IESWTR (described below) are assumed to achieve at least a 2 log removal of Cryprosporzdium. Systems
that use slow sand filtration and diatomaceous earth filtration and meet the turbidity performance
requirements contained in the 1989 SWTR also are assumed to achieve at least a 2 log removal of
Cryptosporidium. Systems may demonstrate that they achieve higher levels of physical removal, and
States have the option of determining whether certain systems do not meet the 2 log removal requirement
even though the systems are in compliance with the revised, more stringent, combined effluent turbidity
provisions in the final IESWTR.
Turbidity Requirements
For all surface water and GWUDI systems that use conventional treatment or direct filtration, serve
10,000 or more people, and are required to filter under the SWTR—
IESWTR Final RJA 1-5 November 12, 1998
-------
• The turbid ry level cia svsiems combined filtered effluent (CFE) at each plant must be less than
or equal toO 3 nephelometric turbidity units (NTUs) in at least 95 percent of the measurements
taken each month, and.
The turbidity level of a system’s CFE at each plant must at no time exceed I NTU
For both the maximum and the 95 th percentile requirements, compliance is determined based on
measurements of the CFE at 4-hour intervals.
Individual Filter Requirements
All surface water systems that use rapid granular filtration and that serve 10,000 or more people conduct
continuous monitoring of turbidity for each individual filter and must provide monthly exception reports
to the State as part of the existing CFE reporting process. Exceptions to be reported include the
following: I) any individual filter with a turbidity leveL greater than 1.0 NTU based on 2 consecutive
measurements 15 minutes apart; and 2) any individual filter with a turbidity level greater than 0 5 NTU
at the end of the lirsi 4 hours of filter operation (i.e ., after backwashing or cleaning) based on 2
consecutive measurements 15 minutes apart. Systems must develop a filter profile if there is no apparent
reason for abnormal filter performance.
If an individual filter has turbidity levels greater than 1.0 NTU based on 2 consecutive measurements 15
minutes apart at any time in each of 3 consecutive months, the system shall conduct a self-assessment of
the filter. If an individual filter has turbidity levels greater than 2.0 NTU based on 2 consecutive
measurements 15 minutes apart at any time in each of 2 consecutive months, the system shall arrange for
the conduct of a Comprehensive Performance Evaluation (CPE) by the State or a third party approved by
the State.
Disinfection Benchmarking
Disinfection benchmarking allows a plant to chart or plot its daily levels of Giardia inactivation on a
graph which, when viewed on a seasonal or annual basis, represents a “profile” of the plant’s inactivation
performance. The system can use the profile to evaluate the effects of possible changes in disinfection
practice on microbial protection. This approach makes it possible for a plant that is considering changing
its disinfection practices to meet DBP MCLs to evaluate whether the particular change under
consideration will result in a lower level of inactivation than the benchmark. Comparison with the
benchmark provides the necessary tool to allow plants, taking source water quality into consideration, to
project or measure the possible impacts of potential changes in disinfection. Only certain systems would
be required to develop a profile and keep it on file for State review during sanitary surveys (i.e., systems
with TTHN/HAAS levels exceeding 80 percent of Stage I DBPB MCLs). Only a subset of those
required to develop a profile (i.e., those intending to make significant changes in their disinfection
practice) would be required to submit their profile and ana1ys s to the State for review.
Sanitary Surveys
The IESWTR requires States to conduct sanitary surveys of alt surface water systems (including GWUDI
systems). Under the IESWTR a sanitary survey is defined as an on-site review of the water source
(identif ’ing sources of contamination using results of source water assessments where available),
IE.SWTR Final RIA 1-6 November 12. 1998
-------
facilities, equipment operalion. maintenance, and monitoring compliance of a system to e a1uate the
adequacy of the system. its sources and operations, and the distribution of safe drinking water Included
in the IESWTR requirements is the concept that components of a sanitary survey may be completed as
part of a staged or phased State review process within the established frequency interval set forth below
Finally, iii order to meet the IESWTR requirements, a sanitary survey must address each of the eight
elements in the December 1995 EPA/State Guidance on Sanitary Surveys
This rule provides that sanitary surveys must be conducted for all surface water systems (including
GW1JDI systems) no less frequently than every 3 years for community systems and no less frequently
than every 5 years for noncommunity systems. Any sanitary survey conducted after December 1995 that
addresses the eight sanitary survey components of the 1995 EPA/State guidance may be counted or
“grandfathered” for purposes of completing the first round of surveys. This approach also provides that
for community systems having outstanding performance based on prior sanitary surveys as determined
by the State, successive surveys may be conducted no less than every 5 years.
In addition, as part of follow-up activity for sanitary surveys, systems must respond to deficiencies
outlined in a State sanitary survey report within 45 days, indicating how and on what schedule the system
will address significant deficiencies noted in the survey. Finally, States must have the appropriate rules
or other authority to assure that facilities take the steps necessary to address significant deficiencies
identified in the survey report that are within the control of the utility and its governing body.
Other Requirements
New provisions under the IESWTR include extending watershed control requirements for unfiltered
systems serving 10,000 or more people to include the control of Crypiosporidium. This builds on the
existing requirements for Giardia Iamb/ia and viruses. Cryprosporidium are included in the watershed
control provisions wherever Giardia Iamb/ia is mentioned. The watershed control program minimizes
the potential for source water contamination and includes a characterization of the watershed hydrology
characteristics, land ownership, and activities that may have an adverse effect on source water quality.
Monitoring for unfiltered systems is not required but will be considered under future microbial rules.
EPA believes that an effective watershed protection program will help to improve source water quality
because existing guidance already references the need to guard against pathogenic protozoa, including
Crypiosporidium specifically. EPA is proceeding on the presumption that existing watershed programs
already consider, and State reviews have evaluated, the adequacy of watershed provisions to assure that
raw drinking water supplies are adequately protected against C,yprosporidium contamination. To the
extent this is not the case, however, EPA expects that unfiltered Systems and States in their annual
review will reassess their program with regard to this concern and take whatever steps are necessary to
ensure that potential vulnerability to Cryp:osporidium contamination is considered and adequately
addressed.
With the IESWTR, EPA includes Cryptosporidium in the definition of GWUDI systems. Systems using
these ground water sources that are considered vulnerable to Crypiosporidiwn contamination would be
subject to the provisions of the SWTR. EPA believes that current GWUDI guidance is adequate and,
based on presently available data, additional changes are not needed to accommodate this provision.
Also included in the IESWTR is a requirement that systems cover finished water reservoirs and storage
IE .SWTR F:nai Pi4 1-7 November /2. 1998
-------
tanks Finished ater reservoirs that are open to the atmosphere may be subject to some of the
environmental factors that surface water is subject to, depending on site-specific characteristics and the
extent of protection provided.
1.5 Environmental Justice
Executive Order 12898 established a presidential policy for incorporating environment.al Justice into
Federal agency missions by directing agencies to identify and address, as appropriate, disproportionately
high and adverse human health or environmental effects of its programs, policies, and activities on
minority and low-income populations.
First, national drinking water regulations apply uniformly to utilities. Although not all utilities have to
modify treatment or operations to reach a particular standard, all must comply with the water quality
standards as promulgated. Thus, the treatment performance level is consistent across all populations
served by surface water systems serving 10,000 or more people. A complementary regulation is under
development that will address similar issues for systems serving fewer than 10,000 people.
In addition, concerns of affected communities, including sensitive subpopulations, were included in the
IESWTR through the RegNeg and M-DBP processes undertaken to craft the regulation. Both committees
were chartered under the FACA and included a broad cross-section of regulators, the regulated
communities, industry, and consumers. Extensive discussion on setting levels that provided the
maximum protection feasible took place, and the final consensus on recommendations to EPA for the
LESWTR considered issues of affordability, equity, and safety.
Finally, the Agency held a stakeholder meeting March 12, 1998 to specifically address environmental
justice issues. The main objectives of the meeting were to solicit ideas from environmental justice
stakeholders on issues surrounding proposed drinking water efforts to increase environmental justice
representation in the regulatory process.
1.6 Unfunded Mandates Reform Act Analysis
Title 11 of the Unfunded Mandates Reform Act (UMRA) of 1995, P.L. 104-4, establishes requirements
for 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 must prepare a written statement
including a benefit/cost analysis, for proposed and final rules with Federal mandates that may result in
expenditures to State, local, and tribal governments, in the aggregate, or to the private sector, of $100
million or more in any one year.
Because EPA believes that this rule may result in expenditures of $100 million or more for State, local,
and tribal governments, in the aggregate, or the private sector in one year, it has prepared Unfimded
Mandaies Reform Act Analysis for the Interim Enhanced Surface Water Treatment Rule to accompany
this RJA. Th4s document reviews the benefit/cost analysis, estimates potential disproportionate budgetary
effects, and summarizes State, local, and tribal government input. The analysis identifies the selected
regulatory option as the least costly, most cost-effective, and least burdensome that accomplishes the
objectives of the IESWTR..
IES *TR Final PJA 1-8 November 12. 1998
-------
1.7 Regulator-v Fle’ibrlity Analysis
A Re ulatorv Flexibihtv Analysis was not prepared for this analysis or rule, since the rule only applies to
systems serving 10,000 or more people EPA has defined srna vstems under the Regulatory Flexibility
Act (R.FA) as utilities that serve fewer than 10,000 people This latter set of systems will be addressed as
part of the upcoming LTI Rule. An RFA will be developed as part of that rule. Although the sanitary
survey requirement in the IESWTR applies to all systems, regardless of size, costs are incurred by States
and are riot cons Ldered under the RFA.
IESW7’R Final JA 1-9 November 12, 1998
-------
2: Consideration of Regulatory Alternatives
2.1 Chronological Review of Regulatory Options Considered
2.1.1 Alternative Development Process
As discussed in Chapter 1, the 1994 Interim Enhanced Surface Water Treatment Rule (IESWTR)
proposal was developed as the result of a Federal Advisory Committee Act (FACA)-chartered Microbial
Disinfectants/Disinfection Byproducts (M-DBP) regulatory negotiation in 1992 and 1993. In response to
expedited regulatory deadlines established by Congress under the 1996 Safe Drinking Water Act
(SDWA), a FACA committee was rechartered in 1997 to develop recommendations on key IESWTR
issues based on new information obtained since the 1994 proposal. As Committee members reviewed
data, regulatory scenarios were forwarded to the Technologies Working Group (TWG) for detailed
analysis and cost estimation. Consensus TecommendationS were developed over the course of the
Committee’s deliberations.
The M-DBP Committee and TWG used a modified “Delphi” expert process in developing consensus
approaches. A “Delphi” analytical process uses teams or groups of experts to reach independent
understandings of technical problems. A modification to this process was used by the M-DBP Committee
in their deliberations. In general, the TWG provided guidance on the specific regulatory alternatives.
Analysts then prepared the cost estimates based on agreed upon assumptions and provided the estimates
to the TWG and Committee for review and feedback. Often, the cost estimates provoked discussion and
debate, with the TWG and Committee members asking for further research and refinements of the
estimates, before reaching a consensus on the recommended approach.
At each phase of the M-DBP Committee process, the Committee reviewed the findings and analysis of
the TWG and further refined the approach. As a result, a variety of alternatives were discussed and
costed in a series of meetings from March to July, 1997. At the first meeting in March 1997, the
Committee discussed turbidity and the use of a disinfection benchmark. The meeting in April focused on
turbidity monitoring; this discussion continued in May with added review of the role of sanitary surveys,
the retention of the predisinfection credit, and a physical removal credit for Cryptosporidium. June and
July M-DBP Committee meetings focused on coming to a conclusion on these issues and capturing
consensus language in an Agreement-in-Principle.
The IESWTR was proposed to improve control of pathogens such as Cryptosporzdium, with the objective
of maintaining protection from microbial pathogens while systems act to meet the new disinfection
byproducts (DBPs) maximum contaminant levels (MCLs) under the Stage 1 DBPR.
Because Cryptosporidium is particularly resistant to inactivation using chlorine, physical removal by
filtration is extremely important in controlling this organism. Current filtration requirements under the
Surface Water Treatment Rule (SWTR) mandate achieving a 0.5 Nephelometric Turbidity Unit (NTU)
for combined filter effluent (CFE) in 95 percent of monthly samples, with Levels never exceeding 5 Nih.
To improve filtration performance, the M-DBP Committee assessed the tightening of these turbidity
performance criteria and monitoring individual filtration performance.
IFS WTR Final RIA 2-1 November 12, 1998
-------
An underL ’ in assumption in the rule is that improved turbidity performance teve s can be achieved
primarily by changes in operation and administrative practices Costly new treatment technologies, in
general. are not necessary The alternative development process did consider the use of membranes. r
nanofiltration technology, if the 95 th percentile performance I c ls were set at 0.1 NTU. This RJA,
however, assumes a 0 3 NTU for CFE for in 95 percent of monthly samples, with levels never exceeding
I NTU, as described in the next section
2.2. Summary of Regulatory Alternatives Considered
2.2.1 Turbidity Treatment
Based on a review of new data, the M-DBP Committee considered tightened turbidity standards for CFE
at three levels: 0.1, 0.2, and 0.3 NTIJ. In general, the M-DBP Committee agreed that plants would
typically target their operations to achieve 0.2 Nih to ensure that they would consistently meet a 0.3
NTU standard. Similarly, plants that expect to meet a 0.2 Nih limit 95 percent of the time would
typically target their operations to achieve 0.1 Nih. At this initial stage of discussion, maximum
turbidity levels (currently set at S NTU) were assumed between 1.0 and 2.0 NTU. A mix of operational
improvements to comply with each level of CFE turbidity were varied at each level. The mix of
improvements selected by utilities primarily determines the costs of compliance.
Defining the types of operational improvements and the number of systems required to modify filtration
activities, and what those activities might be, are important parts of a regulatory discussion on changing
limits. As turbidity limits become increasingly rigorous, a Larger number of surface water systems would
need to modify their filter operations. With current filtration technologies, the difference in compliance
between 0.2 NTU and 0.3 NTU are those of degree. In general, more systems would be required to
modify operational practices, and apply improvements in greater percentages, at 0.2 NTU compliance
than at 0.3 Nih compliance.
Moving to the most stringent 0.1 NTIJ level, however, represents a shift in filtration operations that
differs substantively from the other Iwo possible turbidity limits. In addition, while turbidity
measurement has long been recognized as a means for evaluating treatment performance for removal of
particulate matter (which includes microorganisms), issues remain pertinent as to the accuracy and
precision of the measurements ( I c., turbidimeters with different designs, variations in calibration, and
measurement procedures). A major concern expressed by participants among the M-DBP Committee is
the ability to reliably measure low turbidity levels. The TWG assumed that if systems operated to
achieve a turbidity limit of less than 0.2 NTU 95 percent of the time (as an operating goal to consistently
meet a 0.3 NTh limit), this would provide an adequate margin of safety from variability in treatment
performance and turbidity measurement error. However, the TWG believed strongly that at 0.1 Nih and
below measurement variability became a much more significant issue of technical feasibility. Therefore,
compliance strategies at the 0.1 Nih limit represent a greater commitment of resources because of this
technology shift and issues related to turbidimeter accuracy precision at very low turbidity levels.
The M-DBP Committee explored two compliance options for the 0.1 NTU limit. One compliance option
includes the use of a barrier technology, such as a membrane filter, that could effectively reduce turbidity
levels to 0.1 NTh. Over 95 percent of all surface water systems using rapid granular filtration and
serving populations of 10,000 or more would need to make this technological upgrade.
lES t7 ’R Final PJA 2.2 Nowrnber 12. 1998
-------
Ozone v.as aluated as a second compliance option, to serve as a primary treatment choice as an
alternative to systems being required o achieve the 0 I NTU limit It was assumed that 85 percent of
systems would use ozone treatment technologies, and 10 percent of systems would need to use a barner
technology, such as a filter membrane. Five percent of systems are assumed to not need additional
technologies beyond those required for a 0 1 NT1J or 0 2 Nih limit
The M-DBP Committee discussed lowering the maximum CFE below S NTU. Technical analysis during
previous discussions on the rule had assumed a fixed maximum CFE level of between 1.0 and 2.0 NTU
for the previous options Three options for maximum CFE levels (2.0 Nih, I NTU, and 1.0 NTU) were
analyzed and were the subject of cost modeling. The actual values represented by these proposed CFE
levels were the subject of some discussion. While the 2.0 and 1.0 maximum levels refer to those specific
values, the I NTU maximum is best understood as representing values less than 1.5 NTU, due to
significant figures.
The data available to the TWG demonstrated that 80 to 90 percent of systems already achieve 1.0 NTIJ,
and the cost 4nalysis revealed that the primary issue is compliance with CFE turbidity. The Committee
agreed as part of this discussion that CFE turbidities of 0.3 NTU were achievable by systems with
current filtration technologies and focused their deliberations on this level.
2.2.1 Turbidity Monitoring
Concurrent with the discussion on CFE turbidity levels, the M-DBP Committee reviewed individual
filter monitoring requirements as implemented in several States. Individual filter monitoring is intended
to supplement CFE monitoring by providing a method to identi ’ problems with individual filter
operations that might otherwise be masked in CFE turbidity levels:
The State of California has individual filter monitoring requirements. Filter readings are taken at least
once every 15 minutes, and exceedances are reported to the State in monthly reporting forms. Filter
ripening, the period immediately after cleaning the filter and the start of a filter run, is monitored to
ensure that turbidity levels, commonly elevated during ripening, lower to normal levels within 4 hours.
Under the approach considered by the Committee, California served as a point of departure for
discussions on possible configurations for monitoring provisions. The M-DBP Committee made the
following recommendations regarding individual filter effluent turbidity levels:
Individual filter turbidinieters continuously record turbidity levels with exceedance reporting to
the State;
Different t es of exceedances would compel different responses;
Any exceedance of 1.0 Nit from any filter would be reported in an end-of-the-month
exceedance report ,
If there is no apparent reason for the abnormal filter performance, the system shall conduct a
filter assessment;
If readings of 1.0 NTU are recorded in three consecutive months for any filter, an assessment of
that filter by the utility is conducted;
!ESWTR Final RIA 2-3 November /2, 1998
-------
If readin s of 20 NTU are recorded in two consecutive months, a State or third-parry
Corn prehensive Performance Evaluation (CPE) would be conducted: and,
In all cases, exceedance readings are based on two consecutive filter readings, IS minutes apart,
to provide for instrument error
2.2.3 Disinfection Benchmarking
One of the underlying premises of the M.DBP Committee deliberations was that existing microbial
protection must not be significantly reduced as a result of utilities taking the necessary steps to comply
with provisions of the Stage I DBPR. A key recommendation from the Committee was that EPA include
a provision for disinfection benchmarking in the IESWTR. Benchrnarking will allow systems to more
precisely identify current levels of disinfection inactivation (i.e., disinfection profiles) and evaluate
potential changes that may possibly occur as the result of changes in disinfection practices to meet new
DBP MCLs.
Only certain utilities would be required to develop a disinfection profile and keep it on file for State
review during sanitary surveys (i.e., systems with TTHM/HAA5 levels exceeding 80 percent of the Stage
I DBPR MCLs). Of these systems, only a subset would be required to submit the profile to the State as
part of a package submitted for review (i.e., if the system is intending to make significant changes to its
disinfection practice).
In general, utilities that meet the criteria for preparing a profile may either create the profile by
conducting new daily monitoring or by using “grandfathered” data. A disinfection profile consists of a
compilation of daily Giardia lamblia log inactivation measurements computed over the period of a year.
If new data are required, systems must begin a I-year monitoring effort, to start no later than 15 months
after IESWTR promulgation. Profiles can span I to 3 years depending upon the information currently
available. The State will review disinfection profiles as part of its sanitary survey.
2.2.4 Sanitary Surveys
Sanitary surveys are used as a preventive tool to identify water system deficiencies that could pose a
threat to public health. The July 1994 Federal Register proposed that all systems that use surface water,
or groundwater under the influence of surt ce water, have a periodic sanitary survey regardless of
whether they filter. Prior to the 1ESWTR, the only sanitary survey requirements at the Federal level have
been those specified in the 1989 Total Coliform Rule. Beyond requiring sanitary surveys for systems
collecting less than 5 total coliform samples each month and specifying frequency, the Total Coliform
Rule does not specify what must be addressed in a sanitary survey or how such a survey should be
conducted. The SWTR does not specifically require water systems to undergo a sanitary survey;
however, unfiltered water systems, as one criteria to remain unfiltered, have an annual on-site inspection
to assess the system’s watershed control program and disinfection treaunent process.
Since the publication of the proposed IESWTR in 1994, EPA and the States have issued joint guidance
on sanitary surveys. The guidance outlines the following elements as integral components of a
comprehensive sanitary survey: source; treatment; distribution system; finished water storage;
pumps/pump facilities and controls; monitoring, reporting, and data verification; water systems
management and operations; and operators compliance with State requirements. The M-DBP Committee
JE.SWTR Final RJA 2-4 November /2, 1998
-------
recommended that sur,e s must be conducted for all surface water systems (including ground ater
under the influence of surface water) no less frequently than every three years for community systems
and no less frequently than every five years for noncommunity systems Any sanitary survey conducted
after December 1995 that addresses the eight sanitary survey components of the guidance document may
be counied or “grandfathered” for purposes of completing the first round of surveys
2.3 Cost and Benefit Analyses Conducted
National compliance costs and projected benefits were estimated for all elements of the IESWTR with
cost implications. These benefit and cost projections follow in Chapters 4 and 5
The largest and most complex national compliance cost estimates are associated with compliance under
the strengthened turbidity treatment provisions. Compliance cost estimates were based on the followingS
Determining the number df utilities currently meeting the requirement;
1dentif ’ing filtration improvement activities for those that do not currently meet the requirement;
Assessing the number of systems that would engage in improved operational practices and how
often those practices would be implemented; and,
Determining unit costs of improved operational practices.
Comparative analyses of options were developed during the M-DBP Committee technical discussions.
Each stage of the rule development was accompanied by a comparison in cost among the alternatives. As
the approach was refined, final sets of assumptions and models were reviewed until the national
compliance cost estimates presented in this RIA were established.
For cost estimates for both turbidity monitoring and disinfection benchmarking, the M-DBP Committee
identified the activities associated with each recommendation and then obtained unit costs for the
activities to estimate total costs. A discussion of system and treatment baselines, including source data,
follows in Chapter 3. Chapter 5 details the manner in which the cost estimates were generated, and
provides a summary of total costs.
IESWTR Final RJJ4 2-5 November 12. 1998
-------
3: Baseline Analysis
3.1 Industry Profile
Data on utilities and their capacity to achieve treatment levels were analyzed to develop the national
compliance cost model. Data Inputs include the total number of systems to which the provisions would
apply, households and populations served by these systems, average and maximum system flow rates,
and applicable costs of capital, operations and maintenance, and labor. Utilities are characterized as to
whether they are able to achieve compliance with the recommended provisions, and if not, which
practices they will need to modify in order to comply.
3.1.1 Total Number of Systems
The rule includes treatment provisions for surface water systems serving populations of 10,000 or more.
Systems serving less than 10,000 people will be covered in the Long-Term I Enhanced Surface Water
Treatment Rule (LTI).
The number of systems is derived from preliminary data collected as a result of the 1996 Information
Collection Rule (ICR) data collection effort and from the EPA’s Safe Drinking Water Information
System (SDWIS). SDWIS includes a registry of water systems, self-reported violations of water quality
regulations, and numbers of significantly non-compliant water systems, among other data. Unfiltered
systems and systems that include softenmg plants are not included in the total number of systems.
Preliminary information collected through the ICR provides a more recent and accurate picture of system
numbers and characteristics. Under the ICR, data are available for systems that serve populations of
100,000 or more. For systems serving less than 100,000, the analysis uses the SDWIS database. In
combination, the two data sources provide a reasonable accounting of water systems.
Analysis using these two data sources identified 1,381 surface water systems meeting the turbidity
treatment criteria established for the rule (i.e., serving 10,000 or more and using rapid granular
filtration). This estimate of the number of surface waler systems closely matches system estimates used
in the 1994 analysis of the proposed IESWTR, which identified 1,363 systems, a difference of slightly
more than 1 percent. Differences in the number of systems between 1994 and 1997 are explained by the
great variability inherent in the databases. The number of systems as reported in SDWIS changes
frequently, even daily, and reflects, among a number of factors, changing ownership and the continuous
establishment and dis olution of systems.
Although the turbidity and disinfection benchmarking requirements are applicable only to systems that
serve 10,000 or more people, the sanitary survey provisions in the rule will apply to all systems,
including those serving fewer than 10,000 people. Of these smaller systems, 5,165 surface water systems
were identified and used in the cost estimation procedures for sanitary surveys.
IFS WTR Final RJA 3-1 November 12. 1998
-------
\umber of Plants per System Size
To develop costs for turbidity requirements, the analysis included the total number of plants serving
populations of 10.000 or more (Exhibit 31) The total number of systems under 10,000 was used to
calculate the costs associated with sanitary surveys. For smaller systems (those serving less than 75,000
people), the Technologies Working Group (TWG) assumed I plant per system For systems serving more
than 75,000 people, 2 plants per system were assumed. A total number of 1,728 plants was used for the
analysis.
Population Served
Number of Systems
Average Number of
Plants per System
Total Number
of Plants
10,000-25,000
594
I
594
25,000-50,000
316
I
316
50,000-75,000
124
1
124
75,000-100,000
52
2
104
100,000-500,000
259
2
518
500,000-1,000,000
26
2
52
>1,000,000
10
2
20
Total
1,381
1,728
3.1.2 Treatment Characteristics in Systems Serving 10000 or More People
Once the universe of surface waler systems and plants was established, the current treatment
characteristics were profiled to determine the methods utilities are using to meet the current standards
and how utilities will have to modify their practices to comply with the IESWTR.
Four databases that summarize the historical turbidity of various fi1 ation plants were evaluated to assess
the national impact of modifying turbidity limits. The databases depicted turbidity information from the
American Water Works Service Company (AWWSC0), two multi-state surveys, and a survey of plants
participating in the Partnership for Safe Water program. Only turbidity data from plants serving 10,000
or more people were used. The analyses also included only plants that meet the current 95 percentile
turbidity standard, 0.5 NTU, and the current maximum turbidity standard, 5 NTU, in all months. Each of
the databases was analyzed to assess the current performance of plants with respect to the number of
months in which selected 95 percentile and maximum turbidity levels were exceeded.
The AWWSCo database included annual data for plants-operated by the company in 10 states. EPA
analyzed composite filtered effluent turbidity data obtained from the AWWSCo plants.
The multi-state survey data, which were divided into two databases (State I arid State 2), included
turbidity data from 86 plants in ii states. The plants in the State I database were expected to provide a
Exhibit 3.1 Systems and Plants using Rapid Granular Filtration
per Size of Population Served
In gencnl. the IESWTR do net apply to 1 ri T in this size catego y However, the smiWy swvey ov ioin do apply
o systems serving under 10.000 people
fES tTR Final PJA
3-2
Novemberi2. 1998
-------
more representati\e sample afr .pical plant performance among the plants for which data v ere available
The Siaie 2 database increased re2iorial representation that reflects geographic variations that may not
have been captured in the State I database
The last database included in the turbidity analysis wa.s from plants participating in the Partnership for
Safe Water, a joint venture of several public and private organizations, including the American Water
Works Association and EPA, among others. At the time of analysis, the Partnership membership
included 199 utilities serving approximately 80 million people. The data used were derived from the
Partnership’s 1997 report
These databases provide baseline data to determine the number of systems and plants that would be
expected to modify their treatment practices in order to comply with the turbidity treatment requirements
of the rule The analysis primarily used data from the State 2 and Partnership for Safe Water databases,
as their data were considered more complete and provided a better cross-section of utilities
Each data set contains a subset of systems that do not presently comply with the IESWTR The State 2
database provides a broad geographical distribution of the nation’s systems. The Partnership database is
more representative of larger, more professionally managed systems. The analysis captured these
differences by assuming that the State 2 database most accurately reflects the status of systems serving
populations below 100,000 people, and that the Partnership database accurately reflects systems serving
populations above 500,000 people. For those systems serving between 100,000 and 500,000 people, an
average of the two databases was used. It was not possible to merge the data from the three databases, as
methods of collection, population studied, and data compatibiLity differed.
The data indicated that, in general, as the turbidity limits became more stringent, fewer systems were
able to meet the Limits with their current operational practices. For each of the alternatives discussed
during the regulation development process, different compliance figures were used.
Systems presently unable to comply with the recommended turbidity limits are described as TM occurrence”
systems. The percentages of occurrence systems for each system size category determine the total
number of systems for which the national costs of compliance were calculated.
3.2 Cost Analysis
3.2.1 System Population Size Categories and Total Population
System population characteristics are important to this analysis in several ways. First, all utilities are
categorized by the size of the population served. For this RIA, only systems serving 10,000 or more
people were included (except in the case of sanitary surveys). These systems are divided into the seven
size categories used throughout the analysis and consistent with industry definitions of system size
categories
Household costs, however, did not use population data. Instead, for each system size category, average
flow in millions of gallons per day (MGD) was converted to an annual flow and then divided by a
number representing annual household use. The result was multiplied by the number of systems in the
size category to provii e the total number of households for each size category For further explanation
of the derivation of household costs, refer to Appendix F.
IFS WTR Final R 1A 3-3 November 12. 1998
-------
3.2.2 Average S stem Flow Rates
Average system flow rates are integrated into the national compliance cost model in determining
household costs Average and maximum system flows, expressed in millions of gallons per day (MGD),’
were developed separately’ from the cost model but are key components in generating unit costs (EPA,
July 1998b)
The I 99 I Water Industry Database (W1DB) contains a higher value for the largest (greater than I
million) system size category (350 MGD versus 270 MGD) than the data sources used for the bulk of the
cost estimation under this analysis This higher.flow rate is calculated and displayed in the cost
appendices, where appropriate
3.2.3 Cost of Capitat
A cost of capital rate of 7 percent was used to calculate the unit costs for the national compliance cost
model. This rate represents the standard social discount rate preferred by the Office of Management and
Budget (0MB) for benefit/cost analyses of government programs and regulations.
En addition to the 7 percent rate, unit costs were generated using both a 10 percent and 3 percent rate and
evaluated using the national cost model. The 10 percent cost of capital rate provides a link to the 1994
IESWTR cost analyses and is assumed to be a reasonable estimate of the cost to utilities to finance
capital purchases that may be required under the recommended provisions.
The exhibits of cost estimates presented in Chapter 5 reflect the 7 percent rate. The 10 and 3 percent rates
are presented in the cost summary exhibit (Exhibit 5.1) for purposes of comparison. Costs presented in
the RIA are expressed in 1998 constant dollars.
3.2.4 Unit Costs
Unit cost estimates are an integral part of the calculation of national compliance costs for the turbidity
treatment feature of the rule. Both capital and operating and maintenance costs for each treatment
activity have been estimated (EPA, July 1998b). Unit costs were calculated at 3, 7, and 10 percent costs
of capital. Unit costs estimates, including revised flows, are included in Appendices B through D.
3.2.5 Costs of Labor
Labor rates in the national compliance cost model are used primarily to estimate costs to utilities and
States of the turbidity monitoring and disinfection benchmarking elements of the rule. Both of these
elements include a detailed cost model. Labor rates for these cost models were developed through the
TWO process detailed elsewhere and through a limited survey of plant and system operators conducted
by the American Water Works Association (AWWA)
Labor rates are calculated for three categories: management., technical, and clerical. Management are
those individuals with overall responsibility for the functioning of a plant or system. Technical engineers
are those individuals who operate a plant and would be expected to perform most of the functions
described in the labor burden and cost model. Clerical staff are those individuals primarily involved in
administrative office functions.
IESW7’R Final k/A 3 4 November 12, 1998
-------
•\ labor load rate. representing fringe payments. indirect costs. and general and administrative costs. was
multiplied b the direct labor rate This rate was originally estimated at 150 percent of the direct labor
rate (1 5 load), though current Department of Labor statistics indicate that a lower, 140 percent. rate (1 4
load) is more accurate The I 4 load rate was used in the final calculations
3.3 Benefit Analysis
Estimating the benefits of reducing exposure to Cryp:osporzd:um requires performing a risk assessment
to determine the number of illnesses reduced by the rule and then assigning a value to those reductions
Risk assessments require information on health effects, toxicity, and exposure Benefits analysis requires
information on the value of reducing health and other potential damages. Data to estimate the benefits
associated with reducing health damages (cost-of-illnesses avoided) were derived from previous survey
research on the costs for a giardiasis outbreak (Harrington, et al , 1985 and 1989).
3.3.1 Health Effects and Toxicity
Several sources were used to assess the health effects and hazards posed by Cryptosporidium in drinking
water. Data from the Center for Disease Control (CDC) provided the number of reported outbreaks and
resulting cases of cryptospondiosis (Center for Disease Control, 1996). Other publications provided
information on symptoms and the incidence of hospitalization and fatalities for the Milwaukee outbreak
(Mackenzie, et al, 1994). Information on the tqxicity, dose-response relationship, and ingestion
assumptions were derived from recent peer-reviewed articles (see Chapter 4). These sources described
recent studies on the infection and illness in human volunteers subjected to controlled exposure to
oocysts of Cryptosporidiurn to arrive at an estimate of the risk and toxicity of Crypsosporidium.
The analysis described in Chapter 4 on the characterization of national finished water Crypzosporzdium
distribution was used to assess the population exposure to Crypiosporidium in finished water supplies
(EPA, July l998a).
IESWTR Final RJA 3-5 November 12, 1998
-------
4: Benefits Analysis
4.1 Introduction
The health benefit derived from the promulgation of a drinking water standard is typicaUy thought to be
represented by the health damages that will be avoided as a result of compliance with the standard This
is. however, an incomplete concept The complete concept of the economic benefit of improved drinking
water standards consists of the total value of benefits to the consumer. These benefits include reducing
the probability of suffering health damage and other losses of utility captured in the consumer’s
“willingness-to-pay (WTP)” for the change (Freeman, 1979). To the extent possible, the analysis
presented in this chapter focuses on quantifying and valuing the WTP to avoid health damages, using
out-of-pocket costs only as a substitute measure tithe more complete value is not available
The economic benefits of the Interim Enhanced Surface Water Treatment Rule (IESWTR) derive from
the increased level of ‘protection to public health. Reducing turbidity is indicative of a more efficient
filtration process (Rose 1997). As the efficacy of the filtration process improves, a reduction in
waterborne pathogens, particularly Crypzosporzdium, is likely to be achieved (EPA, November 3, 1997).
In this analysis, the benefits of improved filtration are assumed to be entirely due to the decreased
probability of cryptosporidiosis, the infection caused by Cryptosporidium, and the avoidance of resulting
health costs. Exposure to other pathogenic protozoa, such as Giardia, or other waterborne bacterial or
viral pathogens, are almost certainly reduced by the recommended turbidity provisions but are not
quantified. Also, reduction in waterborne disease outbreaks, which involve societal costs other than costs
of illness (e g., loss of business due to a boiled water advisory, purchase of bottled water or the action of
boiling water) were not included because of difficulties in making such assessments.
Section 4 2 explains the analysis of the economic benefits principaUy associated with health damages
and resulting costs due to cryptosporidiosis avoided under the IESWTR. Section 4.3 discusses, but does
riot quantify, other economic benefits that may result from the rule, including reduced costs to sensitive
subpopulations, reduced or avoided costs of averting behavior, and enhanced aesthetic water quality
4.2 Health Benefits from Reducing Exposure to Cryptosporidiam
4.2.1 Exposure Assessment
Drinking water supplies can be contaminated by a number of pathogens that have been identified as the
cause of waterborne disease outbreaks (Center for Disease Control, 1996). In particular, the
contamination of dnnking water supplies with the parasite Cryptasporidium poses a health risk to the
public because the parasite is highly infectious, resistant to inactivation by chlorine, and small in size and
consequently difficult to filter (Guerrant, 1997). This analysis of benefits for the IESWTR focuses on the
reduction of exposure to Cryptosporidiwn in drinking water supplies through filtration and improved
operation and performance of the filtration process.
Cryptosporidiosis is an acute, self-limiting illness lasting 7 to 14 days with symptoms that include
diarrhea, abdominal cramping, nausea, vomiting, and fever (Juranek, 1995). Exhibit 4 1 contains
information on the symptoms of patients with cryptosporidiosis observed during a major outbreak in
Milwaukee.
IESWTR Final RIA 4-1 November 12. 1998
-------
Exhibit 4.1 S mptoms of 205 Patients with Confirmed Cases
of Cr ptosporidiosis during the Milwaukee Outbreak
S%mptom
Percent Mean
Range
Water Diarrhea
bdomirial Cramps
Weight Loss
Fever
vomiting
93 Duration 12 days
84 N/A
75 10 pounds
57 1009°F
48 . N/A
I - 55 days
N/A
1-40 pounds
990°- 104.9°F
N/A
Source Iackcnzic ci al 1994
Several subpopulations are more sensitive to cryptosporidiosis, including the young, elderly,
malnourished, disease impaired (especially those with diabetes), and a broad category of those with
compromised immune systems, such as AIDS patients, those with Lupus or cystic fibrosis, transplant
recipients, and those on chemotherapy (Rose, 1997). Symptoms in the Immunocompromised
subpopulations are much more severe, including debilitating voluminous diarrhea that may be
accompanied by severe abdominal cramps, weight loss, malaise, and lowgrade fever (Juranek, 1995).
Mortality is a substantial threat to the immunocompromised infected with C’yptaspor:dium:
The duration and severity of the disease are significant: whereas 1 percent of the
iinmunocompetent population may be hospitalized with very little risk of mortality (<
0001), Crypiosporidium infections are associated with a high rate of mortality in the
Lmmunocompromised (5C percent) (Rose, 1997).
There is no effective treatment for cryptosporidiosis (Guerrant, 1997).
According to waterborne disease outbreak data for 1993-1994, the Centers for Disease Control (CDC)
estimate that Crypto.sporsdium was responsible for over 400,000 cases of gastrointestinal infection
(Exhibit 4.2) (EPA, November 3, 1997). The vast majority of these cases occurred in one outbreak in
Milwaukee, Wisconsin, the largest recorded outbreak of waterborne disease in the United States. Of the
approximately 800,000 persons served by the water system, it was estimated using standard
epidemiological methods for estimating cases of illness that over 400,000 (50 percent) became ill. Of
those, 4,000 required hospitalization (approximately percent of those becoming ill), with at least 50
additional cryptosporidiosis-associated deaths among immunocompromised individuals (as reported on
death certificates) (Mackenzie, et aL, 1994; Hoxie, et al., 1996).
IESWTR Final RJA
44
November /2, /998
-------
Exhibit 4.2 %aterborne Crvptospor:dium Outbreaks in the LS.
Associated with Drinking Water by Type of Water Source
(1984-1995)
Date
Location
Water Source
Number of Cases
1984
1986
1987
1992
1991
1993
1993
1993
1993
1994
Braun Station, San Antonio, TX
Albuquerque, NM
Carrollton, GA
Jackson County, OR
Reading, PA
Milwaukee, WI
Yakima, WA
CookCounty,MN
Las Vegas, NV
Walls Walls, WA
Well
Lake
River
Springs and River
Well
Lake
Well
Lake
Lake
Well
2,000
56
13,000
15,000
551
403,000
7
27
78
104
Source Modified from Rose, 1997
The incidence of cryptosporidiosis indicated by the outbreak data presents a dilemma of interpretation.
On one hand, the Milwaukee outbreak is an anomaly in its magnitude of incidence relative to the
incidence historically reported in other outbreaks. On the other hand, the Milwaukee outbreak was
detected late, at about the time when the peak amount of cryptosporidiosis occurred, suggesting that
there may be other such incidences that were unrecorded. Only large outbreaks of cryptosporidiosis eases
concentrated in a specific location have a chance of being detected and reported. Isolated cases
(endemic) are much less likely to be reported. Many, perhaps most, infected individuals may not seek
medical treatment for their symptoms. If the infected individuals do seek medical treatment, primary care
physicians may not be able to isolate Crypiosporidium as the cause of the illness. If diagnosed,
physicians may not report the information to the CDC. These compounded impacts could lead to gross
under-reporting and under-estimating of cryptosporidiosis cases (Okun, et al., 1997).
In addition, the presence of Crypsosporidium in surface water sources is relatively common. Exhibit 4 3
summarizes the level and occurrence, where available, of Cryptosporidium in surface water sources and
in finished drinking water.
IESWTR Final RJA
4-3
November 12, /998
-------
Exhibit 4.3 Summary of Surface Water Monitoring Data
for Cryptosporidium Oocysts
Sample
Source
Number
of
Samples
Positive
Samples
(percent)
Range or Oocyst
Concentration
(Oocysts/L)
Mean
(Oocysts/L)
Reference
Stream
19
73 7
0—240
I 09
Rose, et al, 1988b
Stream! River
58
77 6
0.04— 18
094
Ongerth and Stibbs, 1987
Surface
Water
III
51 4
0 02 — I 3
Rose, 1988
Stream! River
38
73 7
<0001 —44
0 66
LeChevallier, et al , 1991 a
Impacted
River
II
100
2 — 112
25
Rose, et al., I 988a
Reservoir
inlet
l’O
30
0 007 —0 024
0012
4orton and LeChevallier,
1997
Reservoir
outlet
10
70
0.017—031
0081
4orton arid LeChevallier,
1997
Raw Water
85
87
0.07 —484
2.7
LeChevallier, et al, 199 Ia
Lake
20
707
0—22
058
Rosc,etal., 1988b
Lakel
Reservoir
32
75
I 1 —8 9
091
Ongerth and Stibbs, 1987
River
(pristine)
6
NA
0.8 — 5800
1920
Madore, et al., 1987
Riven Lake
262
51.5
0.065 —65.1
2.4
LeChevallier and Norton,
1995
Lakes.’ Rivers
147
20
0.3 —9.8
2 0
Atherholt, LeChevallier and
lorton, 1995
Lakes
179
5.6
0— 22.4
0.333
(median)
Archer, et al, 1995
Streams
210
6.2
0—200
0.07 (median)
rcher, et a!., 1995
Filtered
Water
82
26.8
0.001 —0.48
0.015
LeChevallier, c i al, 1991b
.
Finished
Water
(unfiltered)
6
33.3
0.00 1 — 0.017
0.002
LeChevallier, et aL, 1992
Finished .
Water
262
13.4
0.0029 —0.57
. 0.033
LeChevallier and Norton,
1995
Sourcc EPA, June 24, 1998a.
Because Cryptosporidium is exceptionally resistant to inactivation using chlorine, physital removal by
filtration is extremely important in controlling this organism. Based on the turbidity provisions in the
rule, many water systems would be expected to place an increased emphasis on improving overall
filtration performance. In addition to improving overall filter performance, the monitoring requirements
for individual filters in the proposed rule will improve the performance of individual filters at the water
treatment plants.
/ESWTR Final RJA 4-4 November 12. /998
-------
T i l) v. n b ,neiii anal sis is based on the assumption that irnpro ed overall Filtration performance
and ticlner ..ontrol er nd; dual Filter operations .i.iU lead to fe er Crtpiosporidiiim oocvsts in
Thished drinking ater supplies This is expected to reduce the incidence of crvptosporidiosis in two
‘a s First. the endemic risk (isolated cases that are not reported) is assumed to be reduced because the
rripro ed overall filtration performance will remove a greater portion of oocysts from the finished ‘‘ . tier
supply on a regular basis Second. the risk of outbreaks (large numbers of reported cases) may also be
reduced as the enhanced monitoring and tighter control over individual filter operations allow operators
to detect and prevent breaches in treatment (EPA, November 3, 1997) The risk assessment described in
the next section quantifies the expected reduced endemic risk, with a discussion in Section 4 3 2 of the
expected reduced risk of outbreaks
4.2.2 Risk Assessment Methodology
Risk assessment is an analytical tool that can be used to characterize and estimate the potentially adverse
health effects associated with exposure to an environmental hazard, in this case C yptosporzdzum (Rose,
1997) This risk assessment—used to estimate and understand potential benefits—follows a standard
methodology employed within EPA and the Federal government (National Research Council, 1983).
Risk assessment requires the use of scientific data and, if data are not available, reasonable assumptions,
to produce estimates when there is considerable uncertainty about the exact nature, extent, and degree of
the risk This risk assessment makes use of ranges and probability distributions to take into account
scientific uncertainty.
Risk assessment generally involves three basic steps: identi ing the type of health effect and magnitude
of danger from the substance in question (hazard identification), estimating the exposure (exposure
assessment), and then combining the two to characterize the overall risk (risk characterization) (National
Research Council, 1983). There are three possible endpoints to risk characterization: infection, illness
(morbidity), and deaths (mortality). This analysis calculates the number of illnesses and the associated
number of premature deaths attributable to infection from Cryptospor dium. Exhibit 4.4 displays the
steps in the risk assessment process for characterizing the endemic risk of morbidity and mortality from
Cryptosporidium in drinking water.
IESWTR Final Ri 4 4-5 November 12. 1998
-------
E hibii 4.4 Steps in the Risk . ssessment Process for Cripiosporsdium
H AZ RD
IDE\TIFIC ATION
EXPOSURE ASSESS,\IENT
RJsl
CHARACTERIZATION
Exposure
1 i
Populauon Size and Distribution
(by age, sensitive subpopulations, region)
1 i
Health effects
(# or illnesses, # oldeaths)
Source National Re eaich Council, 1983
Ingestion/Dose Human intake Factors
In order to quantify the health effects due to Cryptosporidium in drinking water, the following input
variables are necessary:
The dose-response function (relation of ingestion to infection);
The viability of oocysts;
The rate of morbidity (illness) given infection; and
The ingested dose (concentration of oocysts in finished water x daily ingestion of water).
The following sections describe the assumptions and derivation of these variables used in the risk
assessment.
Pathogenicity
(dose.response
re I at ions hips)
x
Concentration of Cryp osporidium in
Finished Drmking Water Supply and
Available for Human Consumption
1
Concentration of Crypiospond:um
Removed or Inactivated Durmg Treatment
Concentration of C,ypeospor:dwm in
Source Water
1E.S ’7R Final RIA
4-6
November 12, 1998
-------
. l-1 .sz . rd ldentific.itiofl
A ke step in the risk assessment is to characterize the incidence of the health effect in relationship to the
dose administered (dose-response relationship) Dose-response information for Crvpiospor:dium is
represented b the follo ing general model defining the probability of infection given a dose of
Cr tptosporid:um (Haas. et al, 1996).
[ i 1-exp(-d/k) j
Where fl = probability of infection
d = ingested dose
k = slope parameter (relation of ingestion to infection)
Using data from human ingestion trials of CrypiosporidiumparvUm (C. parvum) (DuPont, et al., 1995),
the best fit value for k (i.e., the number of infections grven ingestion), is estimated at 238.6, with a 95
percent confidence interval of 132.0 to 465.4 (Haas, et al., 1996). These trials were conducted with
healthy, medically screened individuals; as a result, the slope parameter k may be different for sensitive
subpopulations; i.e., a lower dose may induce a response in sensitive individuals equivalent to what a
higher dose induces in healthy individuals.
Not all infections will result in illness and observable symptoms The proportion of all infections that
result in illness is rçferred to as the morbidity ratio. Based on human ingestion trials, a constant
morbidity ratio of 0.39 (i.e., 39 percent of infections result in illness) was estimated, with upper and
lower 95 percent confidence limits of 0.62 and 0.19 (Haas, et al., 1996, DuPont, et al., 1995). The human
ingestion trials (DuPont, et al., 1995) assume no pre-existing immunity. Recently, however, it was found
that after repeated exposure to C. par ’um the rate of illness was the same as the initial exposure, but the
symptoms were less severe and fewer oocystS were shed by re-infected subjects (Okhuysen, et al., 1998).
Different strains of Cryptosporidium may produce different dose-response relationships and morbidity
ratios Preliminary results of human ingestion trials indicate that one strain results in approximately the
same dose-response relationship as C. parvum, while another strain is one log more infectious than C.
parvum but is encountered less frequently (DuPont, 1997). Until more complete experimental data are
available, the dose/response relationship for C. parvum will be used as a proxy for all species of
Cryptosporidium. Additionally, the dose-response relationship and morbidity rate could be different for
sensitive subpopulations. The analysis uses a lognormal distribution for the dose-response relationship
that runs from a low value of 78 to a high value of 782 (mean of 238.6), a one order of magnitude spread.
This distribution should adequately characterize the potential variability of the dose-response
relationship across sensitive populations.
1ESH 7R Final RIA 4-7 November 12, /998
-------
Risk . ssessment . ssumptions:
Dose/Response Relationship fl = I - exp (-dl k)
k value mean = 238.6, 5 th percentile = 132 0, 95 th percentile = 465 4 (data fit to
log normal distribution )
Morbidity mean = 5 ’ percentile = 0.19, 95 th percentile = 062 (assumed
triangular distribution)
Source Haas et ai. i996
4.2.4 Baseline Exposure Assessment
Estimating the exposure to Cryptosporidsum requires four basic pieces of information used to develop
two baseline assumptions (described below): the concentration of Crypiosporidium in source water; the
concentration of Cryptospor:dium removed or inactivated during treatment; the concentration of
Cryptosporidium remaining in finish d water supplies; and the amount of drinking water consumed on a
daily basis.
Exhibit 4.5 displays the national distribution of expected raw (source) and finished water
Crypiosporidium concentrations. The largest survey of Cryptosporidium oocyst occurrence in source
water, using currently available methods, is LeChevallier and Norton (1995), which was analyzed by
EPA in 1996. The mean concentration at the 69 sites from the eastern and central U.S. seems to be
represented by a lognormal distribution. Although limited by the small number of samples per site (one
to sixteen samples; most sites were sampled five times), variation within each site appears to be
described by the lognormal distribution. The quartiles, 90 th, and 95 percentiles for these occurrence data
are presented in Exhibit 4.5.
EPA assumes two potential sources of error in the LeChevallier and Norton (1995) data: 1)
measurements from the eastern and central U.S. may not be representative of the U.S. as a whole; and 2)
the existing analytical method provides poor Cryptosporidiwn recovery. EPA assumes that the
magnitude of the error from each source is approximately equal (about 0.5 logs) but opposite in sign.
Thus, the two error sources act to reinforce the original distribution derived from the LeChevallier and
Norton (1995) data. The poor recovery acts to produce a measured distribution lower than expected,
whereas the possibly poorer quality source water sampled by LeChevallier and Norton (than the U.S. as a
whole) acts to produce a measured distribution higher than expected. Insufficient data exist to further
evaluate these assumptions using quantitative statistical methods.
Assumptions were made about the performance of current treatment in removing or inactivating oocystS
to estimate the fmished water Cryptosporidium concentrations. EPA based these assumptions on
historical studies of Cryptosporidium and Giardia lamblia removal efficiencies by rapid granular
filtration as discussed in the IESWTR Notice of Data Availability (EPA, November 3, 1997). In
summary, a range of 2 to 6 log removal of Cryptosporidium oocysts was observed in several studies
conducted over a decade, depending on source water quality and treatment plant efficiency.
IESW7’R Final RJA 4-8 November 12, 1998
-------
remo’al in a to 6 lo removal range is based on p lo t piant studies that ma ’ be more
a urate t r measuring ooc ’ st removal since, in general, enough cysts were present in the source water to
detect cysts in the finished water Log removal at the low end of the range is primarily based on data
Irom lull-scale plants, some of which may not have been well operated during the evaluation period
These full-scale data were collected before the Surface Water Treatment Rule (SWTR) became effective
and, as such, were collected from full-scale plants some of whose operation may have been deficient as
compared with more recent operation Also, removal data indicated for full-scale plants are probably
biased to the low side because many of the measurements in the finished water are below detection levels
and in such cases finished water values were assumed to equal detection values. Current performance
among treatment plants is likely to be better than chat reflected in the data sets for full-scale plants that
had been collected before the effective date of the SWTR, due to improvements resulting from volunteer
partnership programs (i e , the Partnership for Safe Water) that improve treatment efficiency itt addition
to the SWTR
EPA believes that the SWTR and the Partnership for Safe Water have influenced the removal range of
typical plant performance upward from 2.0-2.5 log removal to 2.5-3.0 log removal, recognizing that
some plants fall above and below this range. Based on this information, the following two assumptions
were made about the performance of current treatment in removing or inactivating occysts so as to
estimate finished water Cryp:ospor:diwn concentrations. The standard assumption is that current
physical removal treatment of oocysts results in a normal distribution mean of 2.5 logs (and a standard
deviation of 0 63 logs). Because the finished water oocyst concentration represents the baseline against
which improved removal resulting from the rule is compared, variations in the baseline log removal
assumption could have considerable impact on the risk assessment. To evaluate the impadt of the
removal assumptions on the baseline and resulting improvements, an alternative normal distribution
mean of 3.0 logs (and a standard deviation of 0.63 logs) was also used to calculate finished water
concentrations of Cryptosporidium as a sensitivity analysis (Exhibit 4.5).
Exhibit 4.5 Baseline Expected National Source Water and Finished Water
Cryptosporidium Distributions, Based on Current Treatment
(oocysts/IOOL) -
Percentile
Source Water
Concentrations
Finished Water Concentrations (oocysts/IOOL)
Assuming Current Removal Equals:
2.5 Logs
3.0 Logs
25
50
75
90
95
103
23!
516
1064
1641
0.20
0.73
2.59
8.10
1604
0.07
0.23
0.82
2.56
5.07
Mean 4.24 1.35
Standard DevIation 25.43 7.76
The concentration of oocysls in finished water refers to a count of the total number of oocysts in the
water and does not take into account whether the oocysts are viable and potentially infectious. The
viability of oocysts after treatment is an area of scientific uncertainty. One study (LeChevallier, Norton,
and Lee, 1991) found that one tenth to one third of occysts in untreated water are viable and potentially
IFS WTR Final RIA 4.9 November 12. 1998
-------
nr ious ba ed on internal morpholo2ical structures To take into account the Impact of the lack of
specificir for species detection (many of which may not be infectious) and inability of methods to
distinguish berv een a live and dead oocyst, EPA chose the low end of the range for this analysis and
assumed that 10 percent ofoocysts in finished water are potentially viable and infectious The
viability/infectivity is modeled in this analysis as a uniform d -tribution with a mean of 10 percent, a ow
‘value of 5 percent. and a high value of 15 percent
Risk Assessment Assumptions:
Viability/infectivity (assumed uniform distribution):
low = 5 percent
average = 10 percent
high = 15 percent
The daily water ingestion of healthy adults was assumed to be lognormally distributed with a mean of
1 948 liters per person and a standard deviation of 0.827 liters. The distribution was truncated, or capped,
at three liters per day (Haas and Rose, 1995).
Risk Assessment Assumptions:
Daily ingestion of water:
1.948 liters per person (data fit to a lognormal distribution with
standard deviation of 0.82 7 liters and capped at 3 liters/day)
Sourcc H s dRosc,i995
4.2.5 Risk Characterization
The above assumptions and factors were used as inputs to a model that calculates the annual number of
infections. The calculations include determining mean daily individual exposure to oocysts from
drinking water by multiplying the concentration of oocysts per liter in finished water supplies by the
amount of water ingested per day. The daily exposure to risk is then multipLied by the viability/
infectivity factor to calculate the number of viable and potentially infectious oocysts. The daily risk of
infection for an individual resulting from the exposure to viable oocysts is calculated by applying the
dose-response relationship, adjusted for the morbidity rate, to the daily ingestion of viable oocysts. The
individual annual risk of infection is calculated by taking the daily risk to the 365 exponent. The
individual risk is converted to a total number of infections in the exposed population by multiplying the
individual risk by the total populatior .
l&SWTR Final RL4 4-10 November 12. 1998
-------
f I =Px(l -exp( V/k)) 6 xM
Where I = total number of illnesses
P = population exposed
M = morbidity rate
V = viability of oocysts
d ingested dose (concentration of oocysts in finished water x daily
ingestion of water)
k = slope parameter (relation of ingestion to infection)
In summary, EPA used the following assumptions in developing the risk characterization.
An exponenti al dose-response function for estimating infection rates with a lognormal
distribution (fit from data) and a mean k of 238.6 (Haas, et al., 1996);
Two liters per person daily water consumption with a lognormal distribution (fit from data)
(Haas and Rose, 1995);
A national surface water lognormal distribution of oocysts modified from data collected by
LeChevallier and Norton (EPA, June 24, 199k);
An assumed uniform distribution of percentage of oocysts that would be infectious with a mean
value of 10 percent (LeChevallier and Norton 1991); and,
An estimated 39 percent (assumed triangular distribution) mean value for people who are
infected and become ill (Haas, et aL, 1996, Dupont et al., 1995).
A Monte Carlo simulation was performed to estimate the mean number of infections, as well as the
distribution of infections around the mean for each of the assumptions about current treatment
performance. The simulation treats each variable (ingestion rate, dose-response, morbidity, and viability)
as a probability distribution, rather than as a single point estimate, to attempt to take into account the
potential uncertainty of the assumptions. The Monte Carlo technique allows repetitive calculations using
values from each distribution according to its probability. The result is a probability distribution of the
estimated number of infections and morbidity that allows the characterization of the mean (expected
value) and range of the risk. Exhibits 4.6 and 4.7 display the distribution of the estimated number of
infections annually as calculated in the Monte Carlo simulation at 50,000 trials. Appendices G and H
contain a summary of the resulting distributions for 2.5 and 3.0 log removal, respectively.
Assuming the standard 2.5 log removal performance for current treatment, the model estimates an
expected value (mean) of 1,503,000 Cryptosporidium infections per year resulting in 643,000 illnesses
from exposure to drinking water supplies in the water systems that will require changes under the rule.
Estimates, at the 90 percent confidence limit, range from a low of 8,000 to a high of 1,24 1,000 illaesses
per year. The values for the 90 percent confidence limit represent the notion that there is a 10 percent
chance that the number of illnesses could be as low or lowerthan 8,000 or as high or higher than
IESW7’R Final AdA 4-11 November 12. /998
-------
•Jt) L nde he m ris n assumptic n cf a 3.0 c remo a!. the icce estimates an a erace of
208.500 ! nes e . th a 90 percent confidence range of 2.500 to 384.500.
A recentl published study (Perz. et at.. 1998) performed a risk assessment for Crypiosporulitim. The
stud used slightl different input assumptions. but the annual predicted infection risk (infections per
person per veañ estimated are comparable at the median: 0.0009 from the studY compared v.ith 0.00 18 at
the 2.5 baseline log removal and 0.0006 at the 3.0 baseline log remo al derived from the current risk
assessment. The current risk assessment results in a much wider spread of values at the 90 percent
confidence interval because of a greater use of distributions in the Monte Carlo simulation.
Exhibit 4.6
Frequency Distribution of Annual Illnesses (Morbidity)
Current Treatment Assumption of 2.5 Log Cryptosporidium Removal, without IESWTR
Forecast: Baseline Total Annual Illnesses
50,000 Trials Frequency Chart 2,512 Outliers
245 12239
184
122
oe1
000 .. 0
0 825.000
4
1.250.000 2.500,000
1,875.000
IFS WTR Final RIA
4-12
November 12, 1998
-------
Exhibit 4.
Frequency Distribution of Annual Illnesses iMorbiditv
(urrent Treatment Assumption of 3.0 Log Cnptosporidium Removal, without IESWTR
Forecast: Baseline Total Annual Illnesses
50,000 Trials
C
178
133
089
044
.000
Frequency Chart
3.936 Outliers
C,
4.2.6 Risk Under the IESW R Provisions
As stated earlier, it is assumed that the turbidity provisions or the rule will result in lower endemic
exposure to Cryptosporidiwn, reflecting improvements in overall and individual filter performance (see
Section 4.2.4). The following assumptions were made to estimate the additional removal of
Crypiosporidium resulting from the turbidity provisions.
Exhibit 4.8 gives the total number of systems, population served, and the number of systems expected to
need additional removal due to the new treatment standard. The source for the number of systems and the
number expected to need additional treatment are described in Chapter 3. The number of systems in the
last three columns are based on results from the State 2 database for populations less than 100,000, the
Partnership for Safe Water data for popu’ations greater than 500,000, and both the State 2 and the
Partnership data for the population category between 100,000 and 500,000. Approximately 50 percent of
systems failed to meet the 0.2 NTLJ standard in at least one month of the year for the State 2 data.
8879
0 125000 250000 375000 500000
!FS*7R Final RIA
4-13
November /2. 1998
-------
Exhibit 4.S Expected \umber of Systems Requiring Additional Treatment if \lonthly
Turbidity Standard is Reduced to 0.2 NTU
System Size
(population served)
Number
of
Systems
Population
Served
(millionS)
Number
Expected
to Need
Additional
Treatment
303
— Systems Failing to Meet 0.2 NTU
Number
Number
Meeting 0.3
NTU
97
Number
Meeting 0.4
NTU
145
Failing to
Meet 0.4
NTU
62
10,000-25,000
25.000-50,000
50,000-75,000
75,000-100,000
100,000-500,000
500,000.1,000,000
>1 Million
594
316
124
52
259
26
10
15686
10202
10 100
39951
22 675
28.240
161
63
27
122
9
3 —
52
20
8,
49
4
2 —
232
77
30
13
47
2
I
315
33
13
5
26
2
I
142
Total 1,381 139.217 688
The assumed finished water Cryp:ospor:dium distributions that would result from additional log removal
with the rule were derived assuming that additional log removal was dependent on current removal, i.e.,
that sites currently achieving the worst filtered water turbidity performance levels would show the largest
improvements or high improved removal assumption (for example, plants now failing to meet a 0.4 NTU
limit would show greater removal improvements than plants now meeting a 0.3 NTU limit). The analysis
also assumes independence between the distribution of Crypwsporzdiwfl and turbidity level. Exhibit 4.8
contains the assumptions used to generate data on improved turbidity plant performance as a result of the
IESWTR.
Exhibit 4.9 is based on a study by Patania, et al., 1995, and shows the relationship between C. parvum
and removal efficiencies by rapid granular filtration as discussed in the IESWTR Notice of Data
Availability. This study showed that, a filter effluent turbidity of 0.1 NTh or less resulted in the most
effective cyst removal (Exhibit 4.9). The improved removal shown under the high removal assumptions
in Exhibit 4 10 are based upon this observed level of cyst removal. An incremental decrease in filter
effluent turbidity from 0.3 to 0.1 NTU reduced cyst removal by up to one log. This oocyst removal range
is the basis for the mid- and low- removal assumptions. Exhibit 4.10 contains the assumptions used to
generate the new treatment distribution for a low-, mid-, and high-log removal assumptionS.
The resulting improved removal assuming current log removal of 2.5 and 3.0 is displayed in Exhibit
4.11.
IE.SWTR Final R1A 4-14 November 12, 1998
-------
Exhibit 4.
Cumulative Probability Distribution of Aggregate Pilot Plant Data for C parvum Removal
0
E
1
x
II
.01 .1 1 510203050 70809095
Exhibit 4.10 Improved Cryptosporidlum Removal Assumptions
Additional Cryptosporidlum Log Removal with IESWIR
Low
Mid
High
Plants now meeting 0.2 Nih Standard
None
None
None
Plants now meeting 0.3 NTU Standard
0.15
0.2.5
0.3
Plants now meeting 0.4 NTU Standard
0.35
0.5
0.6
Plants now failing to meet 0.4 NTU Standard
0.5
0.75
0.9
80
C. pw -v ivt
C
4
0
3
0
0
c
cTurbidity0.l N-ru
•Turbidity0.1 NTU
(C. pai isi remOval > ffldic d value)
cT jty >0.1 NTU
•Tnrbidity>O.1 NTU
(C. parwsi removal> iodlcd value)
Percent of S npIes Less Than Equal to aGi Value
99 99.9 99.99
!FS*TR Final RIA
4.-li
November 12, 1998
-------
E’chibit 4.11 Expected National Source Water and Finished Water
Crvptosporidium Distributions with Imprcwed Removal
Assuming Current Log Removal of 2.5
Percentile
Source Water
Concentrations
(oocvsts/IOOL)
Finished Water Concentration (oocysts/1 001)
—
Current
Treatment
Improved Removal
Low
Mid
High
25
51
75
90
95
103
231
516
1064
1641
020
0.73
259
8.10
16.04
017
052
1.55
4 15
749
015
042
117
2.94
511
014
0.38
103
2.51
427
1.12
Mean
Standard DeviatIon
4.26
24.53
1.94
6.99
1.33
4.01
3.09
Assuming Current
Log Removal of 3.0
25
51
75
90
95
103
231
516
. 1064
1641
0.07
0.23
0.82
2.56
5.07
0.05
0.16
049
1.31
2.37
0.05
0.13
0.37
0.93
1.62
0 05
012
0.33
0.79
1.35
0.35
Mean
Standard DevIation
1.35
7.76
0.61
2.21
0.42
1.27
0.98
Using the assumption of a 2.5 current log removal and mid-case (from Exhibit 4.11) improvement in
removal, the turbidity provisions are estimated to reduce the mean concentration of oocysts from 4.26
oocysts per 100 liters to 1.33 oocysts per 100 liters, a reduction of 69 percent. Using the assumption of a
3.0 current log removal and mid-case improvement in removal, the turbidity provisions are estimated to
reduce the mean concentration of oocysts from 1.35 oocysts per 100 liters to 0.42 oocysts per 100 liters,
also a reduction of 69 percent.
The improved Crypeosporidiwn log removal values were input to the Monte Carlo model simulation.
(See Appendices G and H for distributions.) Exhibit 4.12 summarizes the calculated infections and
illnesses reduced (difference between the baseline and improved removal scenarios as modeled in the
Monte Carlo simulation) for each of the two curTent log removal assumptions under low-, mid-, and
high-case improved removal scenarios. The mean value presented in the tables represents the statistical
expected value of the distribution. The 10th and 90th percentiles implies that there is a 10 percent chance
that the estimated value could be as low as the 10th percentile and that there is a 10 percent chance that
the estimated value could be as high as the 90th percentile.
IESWTR Final RJA 4-16 November /2. 1998
-------
E hibut 4.12 umber of Infections and Illnesses
Baseline Crypiosporsdwm Removal Assumption
2.5 Logs 3.0 Logs
Current Treatment/Baseline
Annual Infections—Mean
Annual Illnesses—Mean
Annual lilnesses—lOth percentile
Annual lllriesses—9Oth Percentile
1,503,000
643,000
8,000
1,241,000
511,500
208,500
2,500
384,500
Low Improved Cryptosporiduam Removal
Annual I nfections—Nlean
Annual Illnesses—Mean
Annual lilnesses—lOth Percentile
Annual Iflnesses—9Oth Percentile
743,500
304,000
7,500
635,500
245,000
99,000
2,400
205,000
Illnesses Avoided with Low Improved Ciyptosporldium Removal Assumption
Annual Illnesses—Mean
Annual Illnesses—lOth PercenWe
Annual Illnesses—90th Percentile
338,000
0
1.029,000
110,000
0
322,500
Mid Improved Cryptosporidlum Removal
Annual lnfectioc,—Mean
Annual Illnesses—Mean
Annual Illnesses—I 0th Percentile
Annual lllnesses .—9Otb Percentile
521,000
210,000
6,900
456,500
168,000
67,000
2,200
144,000
Illncsscs Avoided with Mid improved Crypwsporid1. iva Removal Assumption
Annual Illnesses—Mean
Annual Ulnesses—lOth Percentile
Annual Illnesses—90th Percendle
432,000
0
1,074,000
141,000
- 0
333,000
High Improved Crjpzosporidlum Removal
Annual InfectIons—Mean
Annual Illnesses—Mean
Annual Ilinesses—lOth Percentile
Annual Illneues—9Otb Percentile
445,000
180,000
6,600
391,000
140,000
56.000
2,100
123,000
Illnesses Avoided with H b Improved C ypwspoddIum Removal Assumption
Annual Iilncsses .—Mean
Annual ulnessb—IOth Percentile
Annual Ulncnaes—90th Percentile
463,000
. 0
1,080,000
152,000
0
338,000
ote: Mean values derived from Monte Carlo sunulation may not precisely thatch values derived arithmetically.
4.2.7 Benefits from Reducing Endemic Risk from Cryptosporidium
The health benefits of the rule can be evaluated in terms of two valuation measures: l)cost-of-illness
(COl) avoided, and 2) willingness-to-pay (WTP) to reduce the probability of suffering an adverse health
effect (Freeman, 1979). COT avoided due to adverse health effects includes medical costs, lost income,
!&S VTR Final RJA
4-17
November 12. 1998
-------
reduced produ ti ir and a erting expenditures These are goods for chich there are obser\ able market
values and are therefore, easier to quantify than WTP values
The V TP concept goes beyond the expected value of avoided COl, to include the total value of health
benei ts In principle. WTP for reduced health risk is likely to exceed the market value of avoided co ’st of
illness WTP includes the intuitive notion that illness is, after all, disagreeable arid that one would be
willing to pay to avoid the pain and suffering associated with an adverse health effect beyond the cost of
the illness Since there are no markets for avoided pain and suffering, there are no observable market
transactions by which their value can be measured.
Another reason that the WTP for reduced health risk is likely to exceed the expected value of avoided
CO [ springs from risk aversion Essentially, uncertainty about future damages is unsettling, and there
seems to be an economic premium attached to these kinds of damages. Because it assumes a neutral
attitude towards risk, the use of expected COl (instead of WTP) will tend to understate the economic
value of risk reduction.
Expenditures on averting behavior also comprise a part of WTP In the context of reducing endemic
Ciyp:ospor:dzum risk, averting behaviors involve the day-to-day, routine activities that consumers
undertake with respect to drinking water, including consumption of bottled water or use of individual
filtration devices The reasons for undertaking these behaviors are numerous (i.e., taste, odor, reducing
exposure to chemical contaminants) with the motivation of reducing specifically the nsk from
Crypiosporidvum a minor factor. Expenditures on averting behaviors during outbreaks are discussed in
Section 4.3.4.
Information is not available on the direct measurement of either COl or WTP to reduce risk specifically
for Cryptosporadium. For the purposes of this analysis, estimates for the COt associated with giardiasis
will be used as a proxy for the cost of illness of cryptosporidiosis. The costs incurred during an outbreak
of waterborne giardiasis in 1983 in Pennsylvania were based on a survey of 370 people who had
“confirmed” cases of giardiasis, i.e., a positive stool sample. The study estimated direct medical costs
(paid for either by the victim or insurance company) including the costs of doctor visits, emergency room
visits, hospital visits, laboratory fees, and medication. The study also estimated other costs, including
time costs for medical care, value of work loss days, loss of productivity, and loss of leisure time
(Harrington, et al., 1989). The study did not value the “pain, suffering, stress, and anxiety, or any other
psychological or resulting physiological consequences of the outbreak.” (Harrington et aL, 1985).
Exhibit 4.13 contains a summary of the average losses for confirmed cases of giardiasis in 1984 dollars
and updated using the Consumer Price Index for all Urban Consumers to a 1998 price level.
The average losses per case of giardiasis reported in the survey are approximately $3,100 at the current
price level. The average losses per case of cryptosporidiosis could be less because cryptosporidiosis is
self-limiting in immunocompetent subjects, with infections lasting a shorter duration (7 to 14 days) than
giardiasis infections (30 days median length-of-illness in sample). To take into account the shorter
duration of cryptospondiosis, the estimates for non-direct medical costs of giardiasis are adjusted by the
ratio of the duration of cryptosporidiosis over the duration of giardiasis. The ratio and adjusted costs are
estimated using a Monte Carlo simulation Co model the distribution of potential duration for each illness.
Data from the Milwaukee outbreak indicate that the duration of cryptospondiosis is lognormally
distributed, with a range of I to 55 days, a mean of 1 days, and a median of 9 days (Mackenzie, çt al.,
1994). Data from the Pennsylvania outbreak indicate that the duration of giardiasis is lognormally
distributed, with a mean of 41.6 days and a standard deviation of 45 days (Harrington, et al, 1985). The
IESW7’R FrnaIR!A 4-la November /2, i998
-------
i CL sxiaecj v.ith r .p c por ce r e e “ e tr:curcn no ’ n
E hot 4 L .ith a mean of $2,000 and a median ot $L400 (90 percent ont dence nterva! $SuO-
$3 800)
Exhibit 4.13 Losses per Case of Giardiasis by Category
Loss Category
— Merage Losses
(1984 S)
(Harringion, et al.,
CPI Up
date Factor
Average Losses
(1998 5)
1985)
Direct Medical Costs:
.
Doctor visits
$ 36
2 27
S 82
Hospital visits
100
2.27
227
Emergency room visits
27
2.27
61
Laboratory tests
63
2.27
143
Medication
28
2.27
Subtotal
$ 254
$ 577
Indirect Medical Costs:
Time costs for medical care
$ 18
1
58
S 28
Value of work loss days
359 5* 1
1.58
567
Loss of work productivity
‘
37j***
1.58
586
Loss of leisure time
876’’
1.58
1,384
Subtotal
$1,624
$2,565
Total
.
$1,878
$3,142
• Consumer Pnce Index. All Urban Conswncvs. US City A eisgc. Medical Care: 242.0 Oune 1998 1O6 S (1984 a engc)
Consumer Prire Index, All Urban Consume , ,. US City Avcra c, All ILcn . 1630 (June 1998y103.9 (1984 ivcrqc)
‘ Based on the atsumpuon that the wage saw (or the unemployed. homemakers, and retirees equals the wage rate for employed persons
in the sample Use of an alternative sssumpdon or labor raze will result in different indirect costs.
It is important to note that the values in the above distribution reflect the potential COl avoided, not the
full WTP to reduce the probability of suffering a cryptosporidiosis infection. The estimates do not take
into account the value of avoiding pain and suffering, the economic premium associated with risk
aversion, or the costs of averting behaviors. Therefore the full value of the economic benefit to reduce
cryptosporidiosis may be higher than the $2,000 COl avoided per case mean estimate. Exhibit 4.15
contains the values of annual illnesses avoided, using the distribution of adjusted COl estimates.
To compare these results against previous studies, one study (Mauskopf and French, 1991) estimated the
WTP to avoid foodborne illnesses based on the nature and length of the illness, integrated with the value
of a statistical life and indices of self-reported health status to value the losses in quality and length of
life. The WTP estimates for illnesses similar to cryptosporidiosis range from $156 to $8,004 for mild to
moderate cases of botulism (5 to 21 days of weakness, vomiting, and nausea) and $266 to $2,484 for
salmonellosis (3 to 7 days of similar symptoms). Using these estimates, the value fbr cryptosporidiosis (7
to 14 day duration) could range from $218 ($3 1.20/day for 7 days) to $5,335 ($381/day for 14 days). The
cost of illness estimates (with a mean of $2,000) fall within this range and are a reasonable
approximation of the value to avoid health damages associated with cryptosporidiosis, recognizing that
some costs (such as averting expenditures) have not been monetized.
IFS tTR F,nalRlA
4-19
November /2. 1998
-------
Ethibit 4.14
Frequency Distribution of Adjusted Cost of Illness Estimate
Forecast: Cry ptospondiosis Cost of Illness
50,000 Trials Frequency Chart 348 Outhers
0
061
04 ’
020
000
C,
4.2.8 Benefits from Reducing Mortalities Due to Endemic Risk from Cryplosporidium
Cryptosporidiosis poses a serious risk of death in sensitive subpopulations, such as those with
compromised immune systems. Based on data from the Milwaukee outbreak, the fatality rate can be
estimated at approximately 0.0125 percent (0.0 125 percent of all illnesses would result in a fatality—50
fatalities/400,000 cases) in a mixed population of exposed persons. This figure was derived based on
death certificate reporting (50 additional deaths associated with cryptosporidiosis as reported on the
death certificate, of which 46 had AIDS as the underlying cause of death) and should be regarded as a
minimum estimate (Hoxie, ci al., 1996).
The fatality rate from the Milwaukee outbreak may not be reflective of overall mortality rates from low-
level endemic exposure. The estimated levels of Cryptosporidium in the finished water supplies during
the Milwaukee outbreak were much higher than the levels expected in systems complying with the
existing SWTR. Thus, the higher level of Cryptosporidium in the water supply could have resulted in a
higher fatality rate if more significant symptomatic response were associated with infection influenced
by higher ingested dosages. However, there is no data yet available to support this hypothesis; data is
only available to indicate higher probability of infection resulting from higher ingested dose levels.
There is some evidence that the fatality rate among susceptible subpopulations may not be linked to
community-wide exposure levels (Rose, 1997). The majority of fatalities identified from the Milwaukee
outbreak (46 of 50) were among individuals with AIDS (Hoxie, et aL, 1996). In another outbreak in Las
Vegas, similar mortality rates were observed in AIDS patients (52.6 percent among AIDS patients in Las
Vegas compared with 6 percent among AIDS patients in Milwaukee), although it was hypothesized that
the drinking water had been contaminated over an extended period of time with intermittent low levels of
oocysts, unlike Milwaukee’s massive contamination (Rose, 1997).
- 4083
0
J82
$0.00 $2500.00 $5,000.00 $7500.00 $10000.00
IESWTR Final RL4
4-20
November 12. 1998
-------
Exhibit 4.15 a1ue of Illnesses A%oided • nnualI
Baseline Cryptosporidium Removal Assumption
2.5 Logs 3.0 Logs
Illnesses Avoided with Low Improved Cryptospor:diu
m Removal Assumption
Mean
338,000
110,000
10th Percentile
0
0
90th Percentile
1,029,000
322 500
COl Avoided with Low Improved Cryptosporidium
Removal Assumption
Mean
$950,469,000
$262,876,000
10th Percentile
$0
$0
90th Percentile
$1,883,000,000
$584,500,000
Illnesses Avoided with Mid Improved Crypvosporidiu
m Removal Assumption
Mean
432,000
141,000
10th Percentile
0
0
90th Percentile
1,074,000
333,000
COl Avoided with Mid Improved Crypiosporidium
Removal Assumption
Mean
$1,172,000,000
$327,137,000
10th Percentile
SO
SO
90th Percentile
$1,960,000,000
$607,800,000
Illnesses Avoided with High Improved Cryptosporid
lum Removal Assumption
Mean
. 463,000
152,000
10th Percentile
0
0
90th Percentile
1,080,000
338,000
COl Avoided with High Improved Cr,ptosporidium
Removal Assumption
Mean
$1,240,000,000
$358,900,000
10th Percentile
SO
SO
90th PercentIle
I S1,999,000,000
$619,700,000
The Milwaukee fatality rate might also not be representative of the national fatality rate if there are
larger sensitive subpopulations in Milwaukee than nationally. In fact, sensitive subpopulations may be
under-represented in Milwaukee. According to Hoxie, et al. (1996), “Indeed, in 1992, just prior to the
outbreak, the annual reported AIDS case rate in the Milwaukee metropolitan area ranked 78 th among 98
metropolitan areas in the United States with populations 500,000 or more.” Thus, the greater presence of
sensitive subpopulations in some areas might indicate a greater susceptibility to cryptosporidiosis At this
time, there is no basis for adjusting the Milwaukee outbreak fatality rate to the general population
Assuming the Milwaukee fatality rate of 0.0125 percent, Exhibit 4.16 displays the estimated range of
fatalities prevented as modeled in the Monte Carlo simulation.
!ESW7’R Final PJA 4-21 November /2. 1998
-------
Exhibit 4.16 Number of Mortalities among E osed Population
Baseline Cryptosporidium Removal Assumption
2.5 Logs 3.0 Logs
Current Treatment/Baseline
Annual Mortalities—Mean
nnual Mortalities.—IOth Percentile
Annual Mortalities—90th Percentile
87
1
156
27
0
48
Low improved Cryptosporidium Removal
Annual Mortalities-—Mean
Annual Mortal ities —I 0th Percentile
Annual Mortalitles—901h Percentile
39
1
80
12
0
26
Mortalities Avoided with Low Improved Cryp:osporidüun Removal Assumption
Annual Mortalities—Mean
Annual Mortailties—lOth Percentile
Annual Mortalitfes—90th PercentIle
48
0
129
14
0
40
Mid Improved CrypvosporldI m Removal
Annual Mortalitles—Mean
Annual Mortailties—IOth Percentile
Annual Mortailtles—901h Percentile
27
1
57
8
0
18
Mortalities Avoided with Mid Improved Crypiospoiidium Removal Assumption
Annual Mortalities—Mean
Annual Mortalides-—IOth Percentile
Annual Mortalltlm—90th Percentile
60
. 0
135
- - 18
0
42
Hugh Improved Cryptospo,idJaim Removal
Annual MortalitIes—Mean
Annual Mortallties— .-lOth Percentile
Annual Mortallties—90th PercentiLe
23
1
49
7
0
15
Mortalities Avoided with High improved Cryptosporidhim Removal Assumption
Annual Mortailt lcs—Mean
Annual Mortalitlea—lOth PercentIle
Annual Mortalldes—9Oth Percentile
64
0
136
20
- 0
- 42
Studies that assess the value per statistical life (VSL) saved (i.e., reduced risk of premature death)
generally have central point estimates between $5 million and $8 million dollars with a range from $2
million to $14 million (Chestnut and Alberini, 1997). A recent EPA study characterized the VSL saved
as a lognorrnal distribution with a mean of $4.8 million with a standard deviation of $3.24 million
truncated at $13.5 million (in 1990 price level), based on 26 individual study estimates (EPA, 1997).
Updating the VSL for current price levels results in a distribution with a mean of $5.6 million and a
standard deviation of $3.16 million, truncated at $16.87 million ($13.5 X update factor).
IESWTR Final RJA
4-22
November 12. 1998
-------
-ic T r i’ e rnfl Dt Cr CR, . nmark in sens iti’ e
ubccpuiatioris ,here rna be some arguments tor adjusting the \ SL The r pical valuation methodok
used to dense the \ SL generally measure the individuals \\ TP to reduce the risk of a premature death
b a small amount The small reduction in risk is then spread across a broad population The mortalir ’.
risk associated ith cryptosporidiosis is different ri that a smaller sensitive subpopulation faces a higher
baseline rtsk The valuation literature is unclear on whether this type of a risk would have a higher or
lower WTP although one study found that respondents favored programs that affect smailer populations
facing higher baseline risks, assuming the same number of lives are saved (Van Houtven, 1997). A
review of existing empirical literature with respect to adjusting the VSL saved by drinking water
programs does not, however, provide a strong basis for spec ific adjustments (up or down) to the VSL
(Van Hourveri, et al, 1997)
An alternative method for valuing the increased mortality associated with cryptosporidiosis in sensitive
subpopulations, the quality-adjusted Life years (QALY) method, may be more appropriate than the
commonly-used VSL. The QALY method derives an estimate of the number and quality of life years
extended and then assigns a value to the additional life years. At the present time, there are several
limitations in applying the QALY method, including determining the increased life expectancy among
sensitive subpopulations and improved quality of life In addition, the empirical research to monetize
QALYs is ongoing and not sufficiently robust to use at this time (Chestnut and Alberini, 1997).
For the purposes of this RIA, Exhibit 4.17 displays the potential benefits for preventing fatalities using
the updated VSL distribution, recognizing the uncertainties inherent in this, or any available valuation
methodology.
4.3 Other Benefits
4.3.1 Reducing Health Effects to Sensitive Subpopulations
The health effect of cryptosporidiosis on sensitive subpopulations is much more severe and debilitating
than on the general population. The estimated COl avoided calculated earlier probably does not capture
the full value of costs to sensitive subpopu lations, since health trials were conducted with healthy
individuals and symptomatic responses are more severe in sensitive populations. For example, the
duration of ciyptosporidiosis in those with compromised immune systems is considerably longer than in
those with competent immune systems, with more severe symptoms often requiring lengthy hospital
stays. En those subpopulations, the CO1 from cryptospondiosis would be much larger than $2,000 per
case. During the Milwaukee Qutbreak, 33 AIDS patients with Crypicipondium accounted for 400
hospital days at an additional cost of nearly $760,000 (Rose, 1997). COl due to these hospital days alone
is estimated-at $23,000 per case ($760,000133 patients). Although the COt for sensitive populations is
expected to be greater than the general population, no attempt was made to quantify these effects for the
purposes of this regulatory impact analysis. Also, the cost of averting expenditures could be higher in
sensitive subpopulations.
4.3.2 Benefits From Reducing Risk of Outbreaks
Besides reducing the endemic risk of cryptosporidiosis, the turbidity provisions in the nile may also
reduce the likelihood of major outbreaks, such as the Milwaukee outbreak, from occurring. The
economic value of reducing the risk of outbreaks could be quite high when the magnitude of potential
costs is considered. For example, if the $2,000 per cryptosporidiosis infection estimate is applied to the
iFS WTR Final PJA 4-23 November /2, 1998
-------
4 3,3 Enhjnced . esthetic Water Quality
Economic theor.. suggests that improving the aesthetic quality of drinking water produces benefits
separate from improvements in health Consumers, presumably, would be willing to pay to protect the
aesthetic quahrv of drinking water from high turbidity levels. However, the improvements from the rule
may not be noticeable to the general public in terms of aesthetic water quality. These benefits are not,
therefore, quantified for this analysis.
4.3.4 Avoided Costs of Averting Behavior
During outbreaks or periods of high turbidity, 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. During an outbreak of giardiasis, expenditures on
averting behaviors, such as hauling in safe water, boiling water, and purchasing bottled water, were
estimated at between $1 .74 to $5 53 per person per day during the outbreak (Harringtort, et al., 1989). If
these expenditure figutes are applied to the Milwaukee outbreak, total expenditures on averting behavior
would lie between $19,448,000 ($1.74 X 14 days X 800,000 persons) to $61,936,000 ($5.53 X 14 days X
800,000 persons). Determining the precise reduction in outbreak risk and resulting benefits due to
reduced or avoided averting behavior is not possible given current information, but potential benefits
could be substantial.
IESWTR Fanal RIA 4-2.5 November 12, 1998
-------
5: Cost Analysis
5.1 Introduction
This chapter estimates total national costs of complying with the Interim Enhanced Surface Water
Treatment Rule (IESWTR) It discusses which elements of the rule incur costs, on what basis those costs
are estimated, and how they are aggregated. Chapter 6 compares the cost estimates with potential benefits
of the rule
The cost estimation for the IESWTR combines information from existing data sources with technical
assumptions based on expertise developed by the Microbial-DisinfectantslDiSiflfeCtiOfl Byproducts (M-
DBP) Advisory Committee and its Technologies Working Group (TWG). These estimates are the result
of an iterative process that was continually updated by new data and modified assumptions. Where
necessary, a chronology of the decisions that formed a particular estimate is discussed.
5.1.1 How this Chapter is Organized
Each section of this chapter addresses a particular provision of the IESWTR and its estimated cost. First,
a summary of the estimated national costs of compliance is presented. Subsequent sections discuss each
pros .sion that incurs a cost, assumptions and data elements used in the analysis, how the costs were
estimated, and results. The six provisions described include the following—
I) Turbidity treatment;
2) Turbidity monitoring;
3.) Exception reporting;
4) Disinfection profiling;
5) Sanitary surveys; and,
6) Covered finished water reservoirs.
The cost of regulating utilities is commonly passed to consumers; therefore, an estimate of annual costs
per household of the IESWTR concludes this chapter. Additional documentation on the analyses and cost
estimates in this chapter are documented in Appendices A through F.
5.2 Total National Costs of Compliance
Exhibit S I summarizes the estimate of total national costs of compliance for the IESWTR. The exhibit is
divided into two major groupings; the first grouping displays the final cost estimates (“Final Rule (1998
$s)’) of the IESWTR, the second displays compares this to an earlier estimation developed in 1994
(“1994 Proposal”).
The first column of the exhibit displays the total cost of compliance at a 3 percent cost of capital to
reflect a sensitivity analysis. The second column contains the total cost of compliance at a 7 percent cost
of capital, in keeping with the Office of Management and Budget (0MB) guidance on discounting. The
third column presents the total cost using a 10 percent cost of capital to maintain continuity with the rate
at which costs were calculated in 1994. The fourth and fifTh columns are the total costs of compliance as
computed in 1994 The iourth contains costs of the 19’)4 proposed rule in 1992 dollars. There is some
IESWTR Final RIA 5-1 November 12, 1998
-------
difference in ho’ monitoring and start-up costs for both utilities and States are calculated. based on
re’.ised methods of annualizing these COStS The fifth column updates the 1994 costs with an inflation
factor to 1998 dollars
Differences in cost between the 1994 proposal and the final IESWTR are accounted for primarily by
changes in the level of disinfection required and restoration of disinfection credit prior to disinfection
byproduct (DBP) precursor removal. This results in fewer systems needing to install additional
disinfection contact basins, relative to the costs in the 1994 proposal. The utility treatment options
proposed in 1994 totaled $467 million in annualized costs, compared with $209 million (10 percent cost
of capital) in 1998. a difference of $258 million
5.2.1 How Costs Were Developed
Cost estimates presented in this chapter are based on available data, assumptions, and decisions
developed by EPA and reviewed and confirmed through a modified “Delphi” process. A “Delphi”
analytical process uses teams or groups of experts to reach independent understandings of technical
problems. A modification to this process was used by the M-DBP Committee in its deliberations. A
TWG was formed by the M-DBP Committee early in the rule development process. On any particular
topic or question, the members of the TWG shared their knowledge, experience, and judgment. The
TWG fully discussed cost estimates and then used its collective judgment to reach consensus on whether
the estimate was appropriate for use in the analysis or whether further research was needed. Assumptions
or estimates generated through this process were presented to a full meeting of the TWG and further
presented to the M-DBP Committee.
5.3 Turbidity Treatment
53.1 Overview
The cost of adopting alternative treatment practices to meet the [ ES WTR’s turbidity treatment standards
represents the major portion of costs associated with the rule. Turbidity treatment rule provisions were
the subject of extensive discussion during the M-DBP Committee deliberations. An understanding of the
current configuration of water treatment plants, of how many would be projected to change their
treatment practices based on the rule, and of which alternatives would be implemented (if any) is needed
to accurately estimate turbidity treatment cosis.
5.3.2 Methodology
The baseline of current treatment practices discussed in Chapter 3 served as the basis for discussion on
the compliance forecast. The compliance forecast estimates the number of systems required to modify
their treatment practices to meet the turbidity requirements and the treatment alternatives they would
likely select. Unit costs for each alternative were developed during a companion analysis. The total costs
of the turbidity treatment requirement were then calculated by multiplying the estimated number of
systems modifying treatment, the treatments they would likely implement, and the unit cost of that
treatment.
JESWTR Final RIA 5-2 November 12. 1998
-------
Exhibit 5. 1
Summar of Costs 1 nder the Interim Enhanced Surface Water Treatment Rule (S000s)
Final Rule (1998 Ss)
1994 Proposal
3.1
Cost of
Capital
1/. T
Cost of
Capital
to’ ’.
Cost of
Capital
l0’ ’.Costof l0%Costof
Capital Capital
1992 Si 19985,
Utility Costs
Utility Treatment Capital
S 751,965
S 75L965 [
S 758,965
S 3,665.568 S 4,370,389
Annual Costs
Annualied Capual
Annual O&M
Total Treatment
Turbidity Moniionng
Turbidity Exccptions
Disinfection Bcnchmarking
SubwtaJ
63.999
105.943
171,942
95,924
195
2,841
270.902
85.61!
105.943
191,554
95,924
195
2,841
290.574
103.437
103.943
209,380
95,924
195
2,841
308.340
39 1,702 466,891
S 391,702 S 466.89!
Annualized One-Time Costs
Turbidity Moniconng Stan-Up
H Bcnchmarking
Swbtoal
289
175
464
405
246
651
504
306
810
Tota lAnnua lUt llityCoats
$271,366
S291j65
$309,150
State Coats
Annual Coats
Turbidity Monttonng
Turbidity Exccptions’’
Sanitasy Survey
Disinfection Bcnchmari rng
Subio ai
5,256
409
6,979
2,789
15,433
5,236
409
6,979
2,189
15,433
5,256
409
6,979
2,789
15.433
867 1,034
$ 867 $ 1.034
Annualized One-Time Costs
Turbidity Monnonng Sian-Up
Disinfection Benchmarkmg Start-Up
Sanitary Survey Sian-Up
Subiotal
27
22
39
88
38
30
55
123
48
38
69
lii
Total Annual StateCosts
$15,52 1
$15,556
$15,588
Total Annual Costs S 286,887 S 306,721 S 324,738( $ 392,569 $ 467,925
• Capitai cosU ate smiuthzcd over 20 yeats with the exeapuon of tuibidimcters and pio ss ccviuol m diflc ion equipmeni. which
atinuahzd over 7 yc i.
Costs essociated with Individual Filter Efllucni turbidity requucmenis focexcepuonsteporung and Individua l Filter A cssntents
• All onc -tJmc costs ate annualized over 20 y s.
e s• Cost, associated with Reporting Eacepiatos and Conprehaisive Pcifoimance Evaluations.
!E ø7R Final RIA
5-3
November 12, 1998
-------
Compliance Forecast
The compliance forecast is the heart of the turbidity treatment cost analysis To forecast utility
compliance with the IESWTR, EPA made assumptions regarding which turbidity treatment alternatives
might be implemented. Treatment alternatives could be implemented singly or in combination with any
number of other alternatives.
The compliance forecast is presented as a list of alternatives, with an estimate of the percentage of total
systems that would implement each of the alternatives. Alternatives are generally not exclusive;
therefore, the sum of percentages exceeds 100. The total number of systems forecast to modify their
treatment process is subdivided into different system size categories (defined by the population served),
often with different forecasts for each category.
Estimates of compliance are based on an understanding of current levels of turbidity (as measured in
nephelometric treatment units, or NTUs), and the requirements in the rule. Utilities generally measure
turbidity in two ways: as the output from an individual filter, and as a combined stream of all individual
filter outputs (combined filter effluent—CFE). The 1ESWTR requires continuous turbidity monitoring at
individual filters and requires CFE to be below 0.3 NTU in 95 percent of the monthly measurements, and
CFE is not to exceed I NTU at any time. The IESWTR also sets levels for monthly exceptions reporting
and follow-up activities (see Chapter 1 for discussion of specific activities that are triggered at different
individual filter effluent levels).
During the development of the rule, the TWO analyzed different individual and CFE maximum
turbidities arid reviewed costs associated with each. Analysis originally focused on reducing the existing
0.5 NTU standard to either 0.2 NTU or 0.3 NTU 95 percentiles. In each case, compliance was measured
as meeting the limit 95 percent of the time and not exceeding a CFE maximum between 1.0 NTU and 2.0
NTU. In general, plants that expect to meet a 03 NTU limit 95 percent of the time, in order to ensure that
they would consistently meet this level, would typically target operations to achieve 0.2 NTU. Similarly,
plants that expect to meet a 0.2 NTU limit 95 percent of the time would typically target operations to
achieve 0.1 NTh. In response to concerns that the 0.2 Nih and 0.3 NTU compliance forecasts did not
capture the full potential of turbidity treatment, costs for a more restrictive 0.1 Nih combined turbidity
limit were also analyzed.
The number of systems required to take some action vanes by the proposed regulatoty levels of turbidity
(Exhibit 5.2). At 0.3 NTU, 691 out of a possible 1,381 systems (50 percent) were projected to need to
modify their treatment to comply. At 0.2 NTU, it was assumed that an additional 404 systems (29
percent) would need to modify treatment to comply. For 0.1 NTU, however, rather than incrementally
increasing the overall number of systems, the number for 0.2 NTU (1,095 systems) was used. This
assumed that increased protection would be achieved through adoption of membrane technology rather
than other treatment practices to reduce turbidity (from 0.2 to 0.1 NTh) and that no additional systems
would be affected by the increased requirement.
IESIPTR fl,wJ RJA 5 - 4 November 12, 1998
-------
Exhibit 5.2
Number ofS tems Modifying Treatment Practices to Meet Limit
System Size
(population served)
Number of
Systems
(using rapid
granular filtration)
To Meet 0.3
NTU Limit
To Meet 0.2
NTU Limit
To Meet 0.1
NTU Limit
10,000-25,000
25,000-50,000
50,000-75,000
5,000-1O0,000
100,000-500,000
500,000-1 Million
>1 Million
594
316
124
52
259
26
10
303
161 .
63
27
122
11
4
475
253
99
42
202
18
7
475
253
99’
42
202
IS
7
Total
1.381
691
1,095
l,095
‘1.095 iystcnls nccd to moisfy trcaOncnt to meet the 02 NTU and the 0 I NTU standards Due to roundtn€. the numbet of systems eath of
thc e calcgories totals 1.096
Treatment Activities
Specific treatment activities to help utilities meet the turbidity treatment requirements were proposed by
experts in the technical aspects of treatments and later confirmed by the M-DBP Committee. Treatment
activities were grouped in ten categories. As a general rule, it is assumed that activities are not exclusive
of one another; rather, they can be combined with other activities to make a “treatment mix.” This
assumes that systems can implement more than one treatment activity in order to meet the turbidity
treatment levels. This precludes an analysis of hundreds of separate treatment scenarios. Descriptions of
the treatment activities are included in Technologies and Costs for the Interim Enhanced Surface Water
Treatment Rule (EPA, July, 1998b).
The compliance forecasts in Appendix A display the percentage of systems implementing a specific
treatment activity Appendix A contains all three compliance forecasts discussed dunng the development
of the (ES WTR (final level: 0.3 NTU; alternate levels: 0.1 and 02 Nih).
In general, costs are estimated by multiplying the compliance forecast for each treatment alternative by
the unit cost of the alternative then multiplying the result by the number of systems expected to modify
treatment. This is not applied universally, however. Certain variations in the compliance forecast capture
situations unique to specific treatment activities. These variations are included in the calculations that
generate the total cost of compliance. These variations include the followcng—
The four filtration improvements (filler media addition, filter media replacement wukow support
gravel, filter media and support gravel replacement, and filter media, support gravel, and
underdra:n replacement (Appendix A-I )‘ are not intended to be inclusive; instead, they represent
four separate activities that do not overlap.
The percentage provided in the table for individual filter turbidimerer installation applies to all
utilities for which these regulations are applicable, not just to those systems in need of modifying
their treatment to meet the turbidity levels in the rule. This assumes that all systems will be
required to install individual filter turbidimeters under the rule, regardless of current
IESWTR Frnal RL4
November /2, /998
-------
performance This 3ccounts for the approximately 20 percent of s>stems that already have
rurbidimeters in place
The first activity under Process Control Testing Modiuicaue- (mod 5’uimpIemen: turbidirneter
monitoring and recording) applies to all systems for which .. e rule applies. Eighty percent of all
systems are assumed to need Supervisory Control and Data Acquisition (SCADA) systems in
order to comply with the [ ESWTR. En this case, the SCADA system is expected to monitor and
record—to monitor system flows in th plant (flowmeters), to monitor and control chemical
additions in the plant, and to acquire and record data from turbidimeters.
The second activity under Procçss Control Testing Modification (modl)51/irnplemenl process
monitoring (other than turb:d: y)), refers only to those systems that have or will implement a
SCADA system for turbidity and includes the incremental activity of including a feedback
mechanism for other parameters to allow for continuing corrections to the water stream. In the
cost model, this percentage applies to those systems that need to modify treatment practices to
meet the recommended limits.
Calculating Total Treatment Costs
Units costs for each treatment activity were developed by EPA and are presented in dollars per thousand
gallons ($Jkgal) of average water flow per day. Total treatment costs were computed for each of the three
treatment approaches discussed during the development of the IESWTR (0.1, 0.2, 0.3 NTU). Total
tscatment costs were calculated using the following conversion calculation.
Unit Cost of Activity Conversion Equation
[ Unit Cost of Activity 1/ ($/kgal)] x
[ (1000 Gal/MGD) (Average Flow MGD)) x
(365 DayslYear) x
(Number of Systems Required to Modify Treatment) x
(Percent of Systems Needing a Treatment ActIvity I)
The total annual cost of treannemi is calculated by amortizing the total capital cost at different costs of
capital and adding to the annual operation and maintenance (O&M) cost. Cost amortization used three
different costs of capital (3, 7, and 10 percent).
5.3.3 Estimated Treatment Costa at 0.1,0.2, and 0.3 NTU Levels
This analysis was originally developed to support M-DBP Committee deliberations in 1997.The data
displayed in Exhibit 5 3 are, therefore, presented at the three levels under discussion at that time: 0.1, 0.2,
and 03 NTU. The TWO did not set a specific maximum CFE, so a 1.0 NTh to 2.0 NTU maximum CFE
was assumed for the p irposes of this analysis.
IESWTR Final RI.4 5-6 November /2, 1998
-------
The e iim3Les presented to the \I-DBP Comminee in 1997 sho clear distinctions among the
ditferent proposed re utator levels At 0 3 MTL. total annual costs for turbidity treatment ere
estimated to be S 74 million nnual household cost increases were estimated to be S6 35. or SO 53 per
household per month At 02 NflJ. total annual cost increases were estimated to be $317 million, with
average annual household costs of $6 62, or $1155 per househo’d per month. Household cost increases
remained somewhat stable between these two alternatives because although costs at 0.2 NTU rose
sharply, the total number of households also increased due to the larger number of systems affected. At
0 I NTU. the total annual cost of treatment was estimated to be $3,213 million, or roughly 10 times that
at 0 2 NTU and 20 times the 0 3 NTU scenario Average household cost increases under this scenario
equaled $67 I 7, or 55 60 per household per month
Exhibit 5,3 Cost Estimates for Alternative
Combined Filter Effluent (CFE) Turbidity Limits—May 1997 ($000/year)
System Size
(population served)
Number of
Systems
Mazimum 1.0 to 2.0 NTU CFE
O..3 NTU
0.2 NTU ’
0.1 NTU***
10,000-25 ,000
25,000-50,000
50,000-75,000
75 ,000-) 00,000
I0O,000-5 00 ,000 .
500,000-1 MillIon
>1 MillIon
594
316
124
52
259
26
10
$41,211
32,782
18,605
11,395
47,370
12,316
10,074
$64,640
53,307
27,720
14,326
105,373
29,882
21,511
$519,443
493,916
278.044
152,368
1,221,318
322,177
225,945
Total
1,381
S173,754
$316,759
$3,2 13,211
Average Annual Hou5ehold Cost Increase
$ 6.35
S 6.62
S 67.17
• The iurbrduy level of a syucms CFE 03 NT).) in at levi 95% of mcnthiy mcaturcmcnu
mc turbidity level ofaays*ein’sCFE 0.2 NT).) in at ieist95% of monthly ne wcmeius
The turbidity icvcl of a sy *cm s CFE s 0 1 NT).) in i icatl 95% of monthly neasw’cmcnn
Projected compliance activities most significantly affecting cost include changing primary coagulant
feed points, filter rate-of-flow controller replacement, individual filter turbidimeter installation,
accounting for recycle flow in process control decisions, and process control strategy facilitators.
Additionally, assumptions contained within the compliance forecast for 0.2 Nih and 0.3 NTU differed.
Twice as many systems would install coagulant aid polymer feed and filter aid polymer feed capabilities
in complying with the 0.2 NTU limit as compared with the 0.3 NTU limit.
The estimated total annual costs for systems to comply with the 0.1 Nih limit differed by almost a
factor of 10 from both the 0.2 NTh and 0.3 NTU alternatives. It was assumed that for a system to
comply with the 0.1 Nih turbidity limit, 9 percent of systems would need to install membrane
technology, an expensive alternative that accounts for most of the cost difference between 0.1 NTU and
0.2 NTh. Differences between the two alternatives are slightly moderated, however, by two other
assumptions.
IE.SWTR Final RJA 5-7 November 12, /998
-------
First, the compliance forecast assumes that. in general. moving to more restrictive limits implies that
more systems will have to modify turbidity treatment practices This is exhibited in the compliance
forecast for the 0 3 NIl) and 0 2 NTU limits The 0 I NTU limit, however, uses the compliance forecast
used in the 0 3 NTU analysis. with the exception of membrane technology Systems would not need the
additional improvement in turbidity treatment that movmg from 0.3 to 0 2 NT1J would imply. Therefore,
at 0 1 NTU the compliance forecast includes treatment modifications equivalent to those assumed
necessary to meet the 0.3 Nil) limit. Gains in treatment performance to reach the 0. 1 NTU limit are
achieved through use of membrane technology
Second, in the compliance forecast for both the 0.2 Nih and 0.3 NTU limits, 80 percent of all systems
are anticipated to install individual filter turbidimeters. With the 0.1 Nih option, the use of membrane
technology would effectively remove protozoa and other microbial pathogens. Therefore, no individual
filter turbidimeters would be needed.
5.3.5 Estimated Treatment Costs at 1, 1.0, 2, and 2.0 NTU CFE Maximums
In addition to establishing the levels of CFE monthly turbidity limits, the M-DBP Committee reviewed
the difference in cost at alternative CFE maximums. Prior to this rule-making, limits of CFE turbidity
were set at a maximum of 5 NTh When discussing where to set the monthly CFE limits, the M-DBP
Comm inee had assumed that CFE maximums would be between 1.0 and 2.0 NTU. This implied that if
systems were modifying their treatment at any time to meet the 3.0 CFE 95 percent of the time and their
CFE would not exceed 2.0 NTU at any time.
Three levels for CFE maximum were analyzed. Cost estimates were prepared for I NTU (essentially
levels that could be rounded to I, i.e., up to 1.5 NTU), 1.0 NTU, 2 NTU (essentially levels that could be
rounded to 2, i.e., up to 2.5 Nih), and 2.0 NTU CFE maximums.
New compliance forecasts and cost estimates for the 1, 1.0, and 2.0 NTU CFE maximum levels were
based on analysis conducted by EPA (EPA, June 24, 1998). System data from the Partnership for Safe
Water and State 2 databases served as the basis for the analysis (see Chapter 3). The analysis
conservatively assumed a 0.8 maximum CFE NTU target to meet a 1.0 maximum CFE NTU. The result
indicated that additional costs might be incurred to achieve the more stringent 1.0 and I maximum CFE
NTUs.
At the I maximum CFE NTU level, systems in population size categories below 50,000 were assumed to
perform additional treatment activities 20 percent of the time; for all other systems, 10 percent was
assumed. The slight difference between the 1.0 and the I maximum CFE NTU levels (approximately 3
percent of systems already adopted alternative treatment activities to meet the 0.3 NTU standard) led the
M-DBP Committee to review cost estimates only for the I NTU CFE maximum level. The 2.0 Nfl) CFE
maximum option, the Outer range of maximums assumed under the previous stage of alternative
development, was not explored further as costs for this option had been previously estimated.
To account for activities related specifically to meeting a 1.0 NTU maximum, the 03 NTU compliance
forecast was modified through the modified “Delphi” process and confirmed by the M-DBP Committee.
The CFE maximum level did not require new treatment; instead, increases in the percentages of some
treatment alternatives were presumed sufficient to meet the limit. In all but one case, these percentages
were added to the existing figures. Increases in staff training were assumed to apply to all systems, not
only those systems for which treatment changes were to be made.
IESWTR Final RL4 54 November 12, 1998
-------
:rir ’ ‘ ere estimated (using a 0 3 NT monthl ‘ 5 ’ percent Ie cornainea ; iter
er;luent turbiuir at and 2 0 NIl.. CFE maximum levels Costs at the I NT 1 .. le eI ere estimated at
S03 mil(ion. costs or the 2 0 NYU scenario at 5)99 million (Exhibit 5 3)
Exhib,t 5.4 National Cost Estimates for Alternative Maximum Combined Filter
Effluent Turbidity Limits—June 1997(10 Percent Cost of Capital Rate)
Size Caiegory
(Population Served)
Number of
Systems
Combined Filter Emueni s 0.3 NTU
a Least 95% of the Time ($0003)
2.0 Max 1 ’
I Max 1’1’
10,000 -25,000
25,000-50.000
50,000-75,000
75,000-100,000
100,000-500,000
500,00 0 -I Million
>1 Million
594
316
124
52
259
26
io
$41,211
34,388
17,973
9,309
6,7290
17,200
12,334
$42,706
35,303
18,190
9,407
67,849
17,287
— 12,189
Total 1,381 5 199,458 S 202,932
• CFE not to exceed 20 NTU I ny time
CFE not to exceed I NTU az any time (e 5sentjaily I S NTt))
5.3.6 Ozone for Cryptosporidium Inactivation
The M-DBP Committee conducted additional analysis on the use of ozone as an alternative to systems
required to achieve 0 1 NTIJ. In developing a compliance forecast for this alternative, 85 percent of all
systems required to modify treatment were assumed to install an ozone system and contactor. Of this
number, 30 percent would use Granular Activated Carbon (GACYBiologicaily Activated Filters (BAF),
and 2 percent would use ammonia for bromate control.
Of the remaining IS percent of systems that would not use ozone treatment, two-thirds (or 10 percent of
the total) would install membranes (due to the inability to adequately control for bromate if they were to
use ozone). The remaining third of systems (or S percent of the total) required to modify treatment were
able to use regular process controls to achieve compliance.
The total annual costs for Cryptasporidium inactivation through the primary application of ozone are
detailed in Exhibit 5.5. These costs were calculated by multiplying the unit costs for each option by the
annualized cost equation (shown earlier) and by inserting the appropriate compliance forecast percentage
as above
IESW7 ’R Final RU
5-9
November /2, 1998
-------
E%hlblt 5.5 Cost Estimates for Crvpiosporidiutn Inactivation by Ozone
System Size
(population
served)
Number
of
Systems
Systems
Required to
Modify
Treatment
Systems Using Ozone (85%)
Systems Not Using
Ozone (15%)
Total
Systems
*
GAC/BAF*
*
Ammonia
***
Membranes
•***
Regular
Process
Controls
****•
10.000.25.000
25,000.50.000
50,000- 15.000
75,000-100,000
100,000-500,000
500,000 -I M
> IM
594
316
124
52
259
26
10
475
253
99
42
202
IS
7
404
215
84
36
172
15
6
121
65
25
I I
52
5
2
8
4
2
I
3
1
•
48
25
10
4
20
2
1
24
13
5
2
10
I
0
Total
1,381
1,095
932
281
19
110
55
Annual Est. Cost (5000) S 351,200 S 237,100 $ 440 S 1,279,000 -
‘85% of systems required to modify treatment would have to inaLLi an ozone system conlactor
-. 30’!. of the plants unslailing an ozone system contactor would use GAC/BAF
2/. of the plants installing an ozone system would use wnmon*a lor brontaic conoo)
‘“‘Of the 15’!. of the sysicim that wou1 ioi use ozone treatment. 10% would install membranes
“ Of the total of a ll systcms 5%’re able to use regular process controls to achieve compliance
CalcuEations were based on ozone unit costs from Technologies and Costs for she Interim Enhanced
Surface Water Treatment Rule (EPA, July, 1998b) as well as through TWO deliberations (BAF and
ammonia) Because the figures for Cryptosporidium inactivation were developed with an earlier cost
model, subsequent methodological changes in the model have been made that could, if applied, reduce
the estimated costs of this option, but the magnitude of this change is not known.
5.3.7 Estimated Cost of Turbidity Treatment
Final estimated costs for turbidity treatment in this RIA differ from those presented to the M-DBP
Committee. For ease of comparison to earlier cost estimates, unit costs used by the M-DBP Committee
were generated using a tO percent cost of capital. Later analyses expanded this to include costs of capital
of 7 and 3 percent. As explained in Chapter 3, a 7 percent cost of capital is now used to calculate total
annualized costs. Final estimates of the cost of turbidity treatment are presented at all costs of capital in
Exhibit 5.6.
JESWTR Final RJA 5-10 November /2. /998
-------
Exhibit 5.6 Final - nnual Cost Estimates for Turbidity Treatment Requirements
(0.3 ‘ TU 95th Percentile, 1.0 NTU CFE Maximum) (1998 S000s)
System Size
(population served)
umber or
Systems
3 Percent
Cost of Capital
7 Percent
Cost of Capital
10 Percent
Cost of Capital
10,000-25,000
25,000-50,000
50,000-75,000
75,000-100,000
100,000-500,000
500,000-1 Million
>1 Million
594
316
124
52
259
26
10
S 33,946
29,316
15,450
7,958
56,895
16,3W
10,130
S 37,624
31,862
17,143
8,861
63,544
18,38 1
11,641
S 40,932
35,304
18,564
9,508
69,080
20,092
12,927
Total
1.381
$ 170,00
$ 189,056
S 206,407
5.4 Monitoring Individual Filter Turbidity
5.4.1 Overview
The IESWTR requires that all surface water systems that use rapid granular filtration and serve at least
10,000 people to monitor individual filter turbidimeters for each filter in their system. This section
discusses the model used to estimate costs and displays the result of the analysis. Costs for monitoring do
not include the capital costs of the nirbidimeters. These are included in the previous discussion on
turbidity treatment. This section provides separate and aggregated cost estimates to utilities and States.
A generalized turbidity monitoring model was developed to provide a framework for estimating costs
associated with the IESWrR. The model assumes the use of turbidimeters for each filter and an on-line
SCADA system. Filter readings would be taken at least once every 15 minutes and tabulated. The model
assumes that once during each work shift (8 hours) the turbidity data would be converted to a reviewable
form and would then be reviewed by a system manager. In cases where the monitoring recorded
exceedances, an exception report would be made to the State and, if warranted, an individual filter
assessment might occur.
Exceptions reporting to the State is warranted if any of the following occur
An individual filter has a turbidity level greater than 1.0 Nih for two consecutive measurements
15 minutes apart; and,
An individual filter has a turbidity level greaterthanO.5 NTU at the end of the first four hours of
filter operation after backwash for two consecutive measurements 15 minutes apart.
Requirements for additional triggers are discussed in subsequent sections.
IESWTR Final RLI
s_Il
Nuwmber 12. 1998
-------
5.4.2 Methodology
Costs of turbidity monitoring include both start-up and annual costs for utilities and States. In each case,
the underlying estimation methodology is the same. For both utilities and States, specific activities
associated with rnonitonng were identified, primanly through the use of the modified “Delphi” process
and subsequent confirmation by the M-DBP Committee.
Labor Rate Assumptions
Labor rates used to calculate the turbidity monitoring labor burden, are derived from a document
summarizing cost estimates put forth by plant operators (Via, 1997). Originally, a 1.5 load rate, or 150
percent of wages, (rate of fringe, overhead, and general and administrative costs used to calculate actual
total labor cost) was incorporated into the labor rates to account for the true labor costs. Current
Department of Labor statistics indicate that a load rate of 1.4 is more accurate (Bureau of Labor
Statistics, 1997). The labor rates in the cost model, therefore, reflect a load of 1.4. Unloaded labor costs
ranged from $15.00 per hour for technical engineers to $22.00 per hour for managers. Eighty percent of
the monitoring labor burden will be for technical workers and 20 percent of the labor buiden will be for
managers (EPA, June 24, 1998b).
5.4.3 Estimated Costs to Utilities for Turbidity Monitoring
Overview
Turbidity monitoring is required of all systems covered by the rule and using filtration. This section
estimates the costs to those utilities of monitoring and reporting results annually and the costs associated
with start-up of turbidity monitoring.
Monitoring and Reporting Coats
Utility monitoring activities at the plant level include data collection, data review, data reporting, and
monthly reporting to the State. Burden hours were derived from conference calls with EPA and plant
operators and were reviewed during M-DBP Committee deliberations.
The labor burden hours for data collection and review were calculated under the assumption that plants
are using on-line monitoring, in the form of a SCADA or other automated data collection system. The
data collection process requires that a plant engineer gather and organize turbid iineter readings from the
SCADA output and enter them into either a spreadsheet or a log once per 8-hour shift (three times per
day). Updating of system software was not included as a cost in the final analysis. Upgrading would
occur only if there were an equivalent or greater cost savings from labor reductions due to fewer
readings.
After data retrieval, the turbidity data from each turbidimeter will be reviewed by a plant engineer once
per 8-hour shift (three times per day) to ensure that the filters are functioning properly and are not
displaying erratic or exceptional patterns. A monthly summary data report would be prepared. This task
involves the review of daily spreadsheets and the compilation of a summary report. It is assumed to take
one employee 8 hours per month to prepare. Reicordkeeping is expected to take five hours per month.
Recordkeeping entails organizing daily monitoring spreadsheets and monthly summary reports.
!ESWTR Final RJA 5-12 November 12, 1998
-------
ill ls be re ie ed monthly at the s stem le el o ensure that each plant in a s stem is
in ompllanLe ith the rule A system-level manager or echnical worker ill re ie the daily
monironirw spreadsheets arid monthly summary reports that are generated at the plant level This task is
estimated to take about 4 hours per month Once the plant-level data have been reviewed, the system
manager or technical worker ill also compile a monthly system summary report. These reports are
estimated to take 4 hours each month to prepare.
Start-up Costs
A list of utility start-up activities was derived from “Delphi” discussions with a sample of plant
operators Utility start-up acnvities include reading and understanding the rule, mobilization and
planning, and employee training. System managers would review the rule in order to understand
provision and to determine how these standards will affect their operations. It is assumed that each plant
will need to complete some mobilization and planning in order to comply with the turbidity provisions.
This will require that system managers assess current plant operations and employee schedules in order
to implement a strategy for monitoring the turbidity data.
Total Estimated Cost to Utilities for Turbidity Monitoring
Annual costs to utilities for turbidity monitonng are estimated at $96 million (Exhibit 5.7). The total
utility labor burden of complying with monitoring and reporting requirements is estimated to be over 4
mill ion hours per year. This ecuals an average of 3,016 hours per system per year. The national utility
start-up and implementation costs are estimated at $4.5 million. This is annualized at 7 percent, with a
resulting annual cost of $0.4 million. The labor burden associated with utility stan-up and
implementation activities is estimated to be over 160,000 hours. Actual burdens and costs will vary from
system to system depending on the level of sophistication of the data management systems.
Compliance ACtivitIes
Respondents
Affected
Unit Costs
CF
Annual Costs
Utility Start-Up Costs
Utility Plant Monitoring Costs
Utility System Monitoring Costs
1,381 Systems
1,728 Plants
1,381 Systems
$ 3 l08
52,644
3,588
0.09439
$ 405,136
90,968,832
4,955,028
Total Annual Utility Costs forTurbidity Monitoring and Start-Up $ 96,328,996
5.4.4 Estimated Costs to States for Turbidity Monitoring
Overview
The State’s responsibility under the rule includes reviewing system data to ensure that all systems in the
State are in compliance with the IESWTR. State activities also include reviewing Statewide utility data,
recordkeeping, and determining compliance. State activities were identified through a process of
Exhibit 5.7 Utility Turbidity Start-Up and Monitoring Annual Costs
• Thc Capiiaiizanon Facior (CF)n cikuined uiing the si of cipuii (7%), the number of yean ofc thzavon (20 yevs), md the CWTVItt
value of money (SI).
Stan.up cons a t annualized ovei 2Oyemn with a çF of 0 09439
IESWTR Final PJA
5-13
November 12. /998
-------
iniervte s v.th State oii cials. review of similar regulatory requirements. and confirmation by the M-
DBP Comm irtee Annual State costs for review (nationwide) are estimated to be $5 3 million The annual
labor burden is esttmated to be 182.000 hours. or about 132 hours per system.
Start-Up Costs
One-time State start-up activities include the adoption of the rule and State regulation development. The
list of State start-up activities was derived from technical experts and State regulators.
Total Estimated Cost to States for Turbidity Monitoring
Exhibit 5 8 presents the estimated cost of implementing turbidity monitoring. The rule would collectively
cost States an estimated total of 5407,000 to implement. This is annualized at 7 percent, with a resulting
annual cost of $38,000. The national labor burden for the State program start-up is estimated to total
14,000 hours.
Exhibit 5.8 State Turbidity Start-Up and Monitoring Annual Costs
• The Ca itaiuanon Factoc (CF) caicuI ed tn the coa* of capttai (7%), the ctumbcr of years of capttthzation (20 yevs), and the current
value of monc , (SI)
Sian-up costs are annualized over 20 years with a CF of 0 09439
5.5 State and Utility Turbidity Exceptions Reporting Costs (Exception Reports, IFAs,
and CPEs)
5.5.1 Overview
The turbidity monitoring provisions, in tandem with existing CFE monitoring requirements, are designed
to provide utilities and States with a means o better assess effluent quality. The IESWTR sets new limits
for CFE levels; therefore, exceedance of the individual filter limits would trigger a variety of responses
as described below, depending on the limit exceeded.
Exceptioas Reporting
A monthly exception report must be filed by each utility for exceedances of the individual filter turbidity
limit. Two samples, taken 15 minutes apart, at which a plant exceeds either an individual filter turbidity
of 1.0 NTh or, after ripening, an individual filter turbidity of 0.5 NTU after the first four hours of filter
operation after backwash, constitute an exception. This precludes any anomalous readings by allowing
sufficient time for “bubbles” or other distortions to disperse.
State Start-Up Costs
State System Monitoring Costs
Total Annual State Costs for Turbidity Monitoring and Start-Up S 5,294,503
IESH’TR Final RIA
s-id
November 12. 1998
-------
IFAs and CPEs
In addition to the monthly exception report required for each exceedance. additional requirements are
triggered when exceedances persist. If a plant reports exceedances of 1 0 NTU at one filter for three
consecutive months. an individual filter assessment (IFA) is required. The 1FA will be performed by the
utility If a plant records exceedances of 2 0 NTU at one filter in two consecutive months, a
comprehensive performance evaluation (CPE) is required A State or third parry must perform the CPE.
The intent of this rule element is to provide an opportunity for utilities to correct filter problems after
being alerted to their presence. Thus, a utility can react to the preliminary readings with the exceptions
report arid begin corrective actions internally, thus possibly avoiding costs associated with the IFA or the
CPE.
5.5.2 Methodology
This analysis assumes exceedance rates for each category and the level of effort and cost to respond to
those exceedances based on previous experience and through the use of the modified “Delphi” process.
The incidence of individual filter turbidity readings that would trigger an exception report to the State is
estimated to occur at 10 percent of all systems each year. Compiling and submitting these reports to the
States is estimated to take 8 hours and cost a system $414 per report. The reporting process involves the
review of rnonitonng data spreadsheets and writing the report.
Two percent of all systems are estimated to exceed IFA thresholds of 1.0 NTU individual filter turbidity
in 3 consecutive months. At this percentage, approximately 28 IFM will be conducted each year, at an
estimated cost of $5,000 each. This cost assumes that each WA takes 50 hours to complete at a rate of
$100/hour.
One percent of all systems are estimated to exceed CPE thresholds of 2.0 NTU turbidity in two
consecutive months. At this percentage, approximately 14 CPEs will be conducted each year at a cost of
$25,000 each. This cost is based on the assumption that each State or third party CPE takes 250 hours to
complete at a rate of$ 100 /hour (Exhibit 5.9). For this analysis it is assumed that Stales will perform the
CPEs.
Estimated Cost of Exceptions for States and Utilities
Estimated annual costs for utilities filing exception reports and conducting IFAs total $195,173. States
are expected to incur annual costs of $64,000 to review the exception reports and $345,000 to perform
CPEs. Cumulative annual costs for exception reports, IFAS, and CPEs total $604,000.
IESWTR Final RL4 5-li November 12, 1998
-------
Exhibit 5.9 Utility and State Turbidity Exception Costs
Respondents Affected Cost per
Annually Occurrence
Annual Costs
Utility Costs
Annual Reporling Exceptions
Annual IFAs
Total Utility Exception Costs
138 Systems
28 Systems
S 414
5000
S 57.173
138,000
s 195,173
State Costs
Annual Reporting Exceptions
Annual CPEs
Total State Exception Costs
138 Systems
14 Systems
$ 461
25,000
S 63,664
345,000
$ 408,664
Total Annual Costs s 603437
5.6 Disinfection Profiling and Benchmarking
5.6.1 Overview
This section discusses the cost associated with generating data for and performing a one-time
disinfection profile of a utility’s microbial backstop’disinfection data—disinfection benchmarking. This
profile establishes a benchmark of the utility’s disinfection practices, providing regulators with data to
support their review of utility activities during a sanitary survey or when the utilily changes its
disinfection practices. Unlike turbidity monitoring, which must be done by the 1,381 large surface water
systems that employ rapid granular filtration, disinfection benchmarking requirements must be met by all
1,395 large surface water systems, a difference of 14 systems.
As described in the rule, a disinfection benchmark consists of a compilation of daily Gia?dia lamblia log
inactivations (plus virus inactivations for systems using either chioramines or ozone for pnmary
disinfection), computed over the period of a year, based on daily measurements of operationai data
(disinfectant residual concentration, contact time, temperature, and where necessary, pH). To establish
the disinfection benchmark, the utility will determine the lowest average month (critical period) for each
12-month period and average critical periods to create a benchmark reflecting a lower bound of the
utility’s current disinfection practice. Those utilities with necessary data to develop benchmarks, using
operational data collected prior to promulgation of the rule, may use up to 3 years of that data in
developing their benchmarks. The benchmark will be the average of log inactivations of the lowest
month each year for the 3-year period. Those utilities that do not have 3 years of relevant operational
data will have to begin a 1-year monitoring effort to develop a benchmark. This effort will begin no later
than 15 months after the IESWTR is promulgated.
Costs for benchmarking include costs of the I-year monitoring effort for developing the profik and
benchmark, as well as review of the benchmark when considering disinfection changes. The costs are
shared by both utilities and States. Where costs were unavailable, assumptions were provided through the
modified “Delphi’ process.
IESH’TR Final RJA 5-16 November 12, 1998
-------
5.6.2 \1eihudoIo
For each State or uulit acti ir\ identified. estimated burden hours were rnultiphed by labor rates Labor
caiegor subtotals ere totaled by activity and activities totaled b major benchmarking processes The
labor associated with benchmarking is conducted at the plant el, with the cost per system based on the
number of plants For this analysis it was assumed that smaller systems had one plant. larger systems had
two Again, this analysis includes 1,395 systems to include both filtering and unfiltering systems The
percentage of systems requiring benchmarkirig was determined using data from the 1996 Water Industry
Data Base (WIDB), and totals calculated by system size
Data Monitoring
Each system will review data for total trihalomethanes (TTHM) and S haloacetic acids (H.AA5) to
determine whether it must develop a year-long disinfection profile. Much of these data are already
available All systems over 10,000 already collect TTHM data for compliance with the 1979 Total
Trihalomethane Rule. Systems over 100,000 also collect HAAS data for the 1996 lnformati n Collection
Rule (ICR) To comply with the requirement of reviewing HAA5 data, only systems serving between
10,000 and 100,000 persons are expected to incur costs in collecting new data.
Costs for HAA5 data collection were estimated in the companion Stage 1 Disinfectants/Disinfectjon
Byproducts Rule (Stage I DBPR) regulatory impact analysis. Those costs are incorporated here. For each
collected sample, a 50-minute effort is required, costing $200. For a total of 12 months, this equals
$2,400 and 10 hours. Multiplying by the number of systems serving between 10,000 and 100,000 (1,086
systems) generates a cost of $2,606,400. In addition, it is estimated that I I systems will be required to
generate a public notification of failure to monitor for HAA5. Each public notice cost $210 for a total of
$2,300. Total start-up costs are $2,608,700. This one- time cost estimate annualizes at a 7 percent raze to
$246,000.
Percentage of Systems Needing to Develop a Benchmark
Three industry databases (1996 WIDB, the Partnership for Safe Water database, and the State 2 Survey
database) provided the number of plants per system and number of systems per each size category (1,381
total systems). In addition, 14 unfiltered systems were included in this analysis. In determining the
number of plants per size category, the number of plants per system was multiplied by the number of
systems in each of the population size cazegones. HAA5 and TTHN figures from the 1996 W1DB were
used to estimate the percentage of systems that would prepare a disinfection benchmark under the
IESWTR This analysis determined that 29 percent of the systems would need to develop a benchmark.
This percentage reflects the number of systems with data showing Tfl-!Ms or HAAS greater than or
equal to either 64 .ugfL or 48 big/L—80 percent of the maximum contaminant levels (MCLs) for 1 ’THMs
and HAA5, respectively.
Utilities must prepare a disinfection benchmark if they:
Measure TTHM levels of at least 80 percent of the MCL (64 gfL) as an annual average; or,
Measure HAA5 levels of at least 80 percent of the MCL (48 .ig/L)as an annual average.
(ES WTR Final RL4 5-17 November 12, 1998
-------
The IQ°6 \ 1DB includes annual a ’. erage TTH\4 and HAA5 figures from 57-t plants (comprising 399
s stemsj Anal sis of the 78 s stems in the 1996 WIDB for ‘ .‘.hich TTHM and HAAS data exist shows
that 29 percent had TTHM levels greater than 64 Lzg/L andior HAA levels greater than 48 j . g/L Under
the rule these systems would conduct a disinfection benchmark (Exhibit 5 10)
TTH v1 and l-IAA5 data exist only atthe distribution system level arid, therefore, only permit an analysis
at that level For example, if there were five plants in a system, the TTHI ’vl and 1-LAA5 data for these
plants were identical as all five feed into the same distribution system. To avoid double counting, only
one set of TTHM and HAA5 data were used as part of the analysis in this example The absence of plant
disinfection byproduct data limited an analysis to the system level.
Exhibit 5.10 TTHM and HAAS Data from the Water Industry Database (WIDB)
TTHM level
(jzgIL)
HAA5 level
( IL)
Number of WIDB Systems*
Systems wI TTHM data
on ly’
Systems wI both TTHM &
HAA5d2I2***
64
<64
0.048
<0.048
50(22%)
180 (78%)
22 (29%)
56 (71%)
Total 230(100%) 78(100%)
• Systems for which no data exists or for which oniy zero axial. were omitted from the data set.
Plants that aic calcgonzcd only cordmg to lThMs because HAA3 data are isol reported
• S Systems thai havc either TTHM levels that arc z 64 /L or HAA5 icvcts that are 48 t gfL
Utility Activities
Utility costs associated with benchmarking were divided into four activity components. These are cost
per system, cost per plant using paper data, cost per plant using mainframe data, and cost per plant using
Personal Computer (PC) data. Each component is made up of activities defined by the TWG, reflecting a
pLant’s method of collecting data. Plants with paper data were assumed to represent half of the number of
plants needing a disinfection benchmark, while plants with mainframe data and plants with PC data each
represent a quarter. The TWG assumed that all plants currently collect this data in either an electronic or
paper format and, therefore, would not incur additional data collection expenses due to microbial
profiling.
Average system costs were multiplied by the percent of systems needing a disinfection benchmark and
summed by system category. Each plant category was multiplied by the corresponding percentage (the
percentage of plants using either paper, mainframe, or PC data), with total plant costs representing the
sum of all types by system size category.
State Activities
Each State will review disinfection benchmarks as part of its sanitary survey process. Those utilities that
decide to make a significant change in disinfection practice (e.g., move point of dismfection, change the
type of disinfectant, change the disinfection process, or make other changes designated as significant by
ihe State) are required to develop a disinfection benchmark and must consult with the State prior to
implementing such a change. Supporting materials for such consultation must include a description of
IFS WTR Final RIA
5-18
November 12, 1998
-------
me proposed ‘ .han e :he disinfection benchmark, and an anaks s ot ho the proposed change atfe :
the etfecti eness of disinfection
State activities considered applicable to the disinfection benchmark process included reading and
understanding the rule changes. mobilization and planning. training of State staff, and providing training
in protocols for utilities and consultants These activities are complemented by additional tracking of
system compliance, review of data received, making regulatory determinations, meeting with utilities,
and recordkeeping
Estimates for State start-up and annual costs were totaled State start-up costs are derived by multiplying
State start-up costs by the number of States and terntories (56). These costs are then annualized at 7
percent. Total annual costs for all States are also derived by multiplying the annual costs per State by 56.
Labor Rate Assumptions
Labor costs and assumptions figure prominently in benchmarking cost estimates. Two classes of labor
comprise the work effort: management and technical. Management and technical positions are assumed
to divide labor hours at a 1.4 ratio. Clerical hours are incorporated into the cost, but are not part of the
burden needed to complete an activity.
Costs of labor were derived from the modified “Delphi” process, and are based on actual labor rates in
place in numerous systems(Via, 1997). Unloaded labor costs ranged from $15.00 per hour for technical
engineers to $22.00 for managers. Labor rates include a 1 4 load, representing fringe rates, provided by
the Bureau of Labor Statistics (BLS).
5.6.3 Estimated Costs of Disinfection Bencbmarking
Exhibit Si I displays the specific activities and costs pertinent to utilities required to benchmark under
the recommended provisions. Provided are the total costs incurred by systems and costs based on current
information collection techniques.
Annual Utility Disinfection Benchmarking Cost Estimates
Cost per Entity
Number of
Entities
Annual Costs
3,987
Paper Data 3,551
Mainframe Data 1,339
PC Data 1,283
405
252
126
126
S 1,616,517
894,701
168,638
161,582
$ 2,841,438
Exhibit S 12 displays State costs for reviewing utility disinfection benchmarks. This exhibit illustrates
the start-up costs per State, as well as the total annual costs for all States.
IESWTR Final PJA 5-19 November 12. /998
-------
E’thibit 5.12 4nnual State Disinfection Benchmarking Cost Estimates
Cost per State
Number or States
& Territories
Total Costs
Start-Up Cost (annualized)
Annual Cost
S 544
49,795
56
56
5 30,489
2,788,632
Total Annual Cost S 50,341 S 2,819,121
Labor costs are the primary factor in both State and utility total benchmarking costs. Where labor rates or
activity burdens are high, costs are high. This sensitivity to labor rates and burdens increases the need to
better understand the ratio between high- and low-burden activities. For example, modifying assumptions
on plant processes with high labor burdens (e.g., plants with paper 4 ta) could substantially alter the final
cost totals.
5.7 Sanitary Surveys
5.7.1 Overview
A sanitary survey is an onsite review of the water source, facilities, equipment, operation, maintenance,
and monitoring compliance of a utility. The survey evaluates the adequacy of the system. its sources and
operations, and the distribution of safe drinking water. The sanitary survey documents the capabilities of
a system to continually provide safe drinking water and identifies any deficiencies.
Elements of the rule, such as disinfection benchmarking, expand existing sanitary survey practices. For
example, disinfection benchmarking requirements may entail additional State review during a sanitary
survey to assess the results of the disinfection profile for microbial inactivation. lii addition, the IESWTR
also requires States, as part of the sanitary survey requirement, to work with utilities to overcome or
address significant deficiencies.
The IESWTR requires that the State, or third party approved by the State, conduct sanitary surveys for
all surface water systems (including both filtered and unfiltered systems) no less frequently than every 3
years for community systems and no less frequently than every 5 years for noncommunity systems. Any
sanitary survey conducted after December 1995 that addresses the eight elements outlined in the rule
(source; treatment; distribution system; finished water storage; pumps, pump facilities, and controls;
monitoring, reporting, and data verification; system management and operation, and operator compliance
with State requirements) may be considered “grandfathered” for purposes of completing the first round
of surveys. This approach also provides that for those community systems determined by the State to
have outstanding performance based on prior santtary surveys, successive sanitary surveys may be
conducted no less frequently than once every 5 years.
5.7.2 Methodology
States will perform start-up activities, such as planning and training, to.prepare for conducting sanitary
surveys. These costs are based on a per-state estimate of the. technical and managerial labor hours that
will be required. This annualized cost at 7 percent cost of capital is presented in Exhibit 5.14.
IESWTR Final RJA 5-20 November 12 1998
-------
-\nnua aniur . sur e . coSts are a function of plant and s stem size The ar er the plant or stem the
more e tensr.e the data gathering, data reviev . and data reporting effort Estimated costs per sur .e in
those s’siems serb trig less than 5.000 are rough v similar. larger size categories see progressively
higher sanitary survey costs (Exhibit 5 14, see Appendix E- 11 for details)
States are expected to conduct sanitary surveys on a rotating basis For this analysis, 80 percent of
surveys are assumed to have already been conducted. This analysis assumes that the remaining 20
percent to comply with the rule.
Unlike other elements of the rule, the sanitary surveys need to be conducted for all surface water
treatment systems, not just those serving more than 10,000 persons However, this does not trigger the
requirement for a regulatory flexibility analysis because the IESWTR requires States to cdnduct surveys,
as reflected in this cost analysis.
Sanitary surveys must also be conducted for systems that do not filter, unlike most of the IESWTR
requirements. The Impact of surveys for these systems on cost estimates is minor. Unfiltered systems
have a smaller treatment process to review but a more extensive source water review. The Surface Water
Treatment Rule (SWTR), however, addresses source water review in unfiltered systems, so these costs
are not counted in this estimate.
This analysis establishes a list of activities that are typically conducted during a sanitary survey This list
is derived from guidance on conducting surveys. Each activity has an estimated cost, computed as the
number of hours needed to complete the task by labor category. The total time needed to complete an
activity is longer in larger systems than in smaller.
Total sanitary survey costs were omputed for each size category for both filtered and unfiltered water
systems (6,560 systems), then multiplied by the percentage of plants needing to conduct a survey. It is
assumed that there are 5,165 small surface water systems. The baseline number of large systems (1,395)
to perform sanitary surveys is different than the baseline number of systems presented in Exhibit 3.1,
since insufficient data exist to determine how the 14 unfiltered systems that do not have to modify
treatment are categorIzed by system size. The baseline for this analysis was established from the 1994
Stage I DBPR RIA. While the number of systems reported in each size category differs from that
presented in the rest of this RIA, it provides a good estimation of the costs States will have to incur to
perform sanitary surveys.
5.7.3 Estimated Cost of Sanitary Surveys
Exhibit 5 13 displays the revised baseline estimated for this analysis. Exhibit 5.14 displays the total cost
to States per population size category based on this revised baseline. Appendix E-l I includes start-up
and annual activities and burdens for each size category, including distinctions between filtered and
unfiltered systems (Exhibit 5.14 may not match detail of Appendix E-11 due to independent rounding).
The cost estimates in this analysis include all costs of conducting sanitary surveys, not just the
incremental effort included in the rule. The costs presented here are, therefore, considered an
overestimate of the probable costs.
IESWTR Final RJ,4 5-21 November 12, 1998
-------
Exhibit 5.13 Revised Baseline of Systems
Based on 1994 Stage 1 DBPR RIA
Sbstem Size
(population
served)
Total Number
of Plants
Number of
Plants, Filtered
Filtered Plants
to Conduct
Survey
Number or
Plants,
Unfiltered
Unfiltereu
Plants to
Conduct Survey
< 10K
l OK.25K
25K-50K
SOK-75K
75K-lOOK
I OOK-500K
SOOK-IM
> IM
5J65
569
328
157
108
350
86
30
4,880
551
322
155
216
344
82
28
976
110
64
31
43
69
17
6
285
18
6
2
0
6
4
2
57
4
2
0
0
i
0
0
Total
6,560
6,578
1,316
323
64
Exhibit 5.14 Total Stan-Up and Annual Costs of Sanitary Surveys
Based on Revised Baseline of Systems
Start-Up Costs Annualized at 7% Cost of Capital
s 55,356
System Size
(population served)
Costs for Filtered Plants
Costs for Unfiltered
Total Costs
<10K
S 1.464,000
$ 55,575
S 1,519,575
10K-25K
165,000
3,900
168,900
25K-5OK
580,800
9,110
589,910
50K-75K
499,500
0
499,500
75K-lOOK
1,041,675
0
1,041,675
IOOK-500K
2,194,200
15,115
2,209,315
500K-IM
669,375
0
669,375
‘ IM
280,800
0
280,800
Subtotal
S 6,895,350
S 83,700
6,979,050
Total Annual Costs (including annualized costs)
S 7,034,406
IF . SWTR Final RL4 5-22 November 12, 1998
-------
5, Co erecJ Finished ater Reser oirs
5.8.1 Over iew
The IESWTR requires that systems cover all new finished water reservoirs, holding tanks, or other
storage facilities for finished water Finished water reservoirs open to the atmosphere are subject to the
same environmental factors as surface waters, depending on site-specific charactenstics and the degree
of protection provided These include contamination by persons swimming, by disposal of garbage into
the reservoir, by microbial and other organisms, and by small mammals, birds, fish, and the growth of
algae This contamination is marked by increases in algal cells, bacteria, turbidity, total and fecal
cokforrns (e.g., E coli), and pathogens.
5.8.2 Methodology
The analysis of costs for covering finished water reservoirs is complicated by the lack of data regarding
the construction of reservoirs. The precise number of systems constructing finished water reservoirs is
unknown. As the rule requires all systems constructing finished water reservoirs to cover them, its cost
impact is only on those who were not originally planning to construct covers. Furthermore, reservoirs are
not uniform in size, configuration, or depth, requiring the development of a range of unit costs to capture
the variability of the total costs.
To address these factors, several key assumptions were made. To estimate the number of reservoirs being
constructed, data on the number of systems serving at least 10,000 people were gathered from a 1987
ASDWA survey and compared with the base number of systems in this analysis, derived from the Safe
Drinking Water Information Systems (SDWIS). The difference is the total number of new systems for a
9-year period, which was then extrapolated for the full 20 years used elsewhere in this analysis.
EPA estimates that 10 percent or fewer of nevlly constructed systems use finished water reservoirs.
Although some States require that finished waler reservoirs be covered, many systems have already
covered their reservoirs as a response to the public health concerns raised by uncovered reservoirs. The
actual number of covered finished water reservoirs, however, is difficult to establish. This analysis is
based on the estimate that 10 percent of systems use finished water reservoirs, and that all covers for new
reservoirs are implemented in response to this rule. EPA believes this to be a conservatively high
estimate of the number of finished water reservoirs to be covered specifically in response to the !ESWTR
provisions.
The calculations for this rule element use a model finished water reservoir, assuming a 25-foot depth and
a reservoir storage volume equal to one day of average water flow capacity at each system size category.
Cover costs are estimated at $2.00 per square foot.
Estimated Cost of Covered Finished Water Reservoirs
The estimated costs of covered finished water reservoirs at 7 percent cost of capital are presented in
Exhibit 5.15
Based on the assumptions above, approximately 73 new systems will build covered finished water
reservoirs in the next 20 years. Annualized capita’ costs are estimated to be $2 million and annual O&M
JESWTR Final RJA 5-23 November 12, 1998
-------
is estimated to be SO 5 mi hon Approximately 65 percent of the covered finished water reservoirs are
estimated to be built in the r o smallest size categories
Exhibit 5.15
Annual Cost of Covered Finished Water Reservoirs (7 percent cost of capital)
System Size
(population
served)
Estimated
Number of New
Plants with
Covered Finished
Water Reservoirs
Total Annualized
Capital Costs
Total Annual
O&M
Total Annual Cost
IOK-25K
25K-SOK
50K -75K
75K-tOOK
IOOK-SOOK
500K-tM
> IM
30
17
8
6
9
2
1
S 146,141
193,049
151,209
150,244
477,542
483,162
345,115
$ 122,618
87,772
51,232
42,501
105,596
83,350
56,782
$268,759
280,821
202,441
192,745
583,138
566,512
401,897
Total
73
S 1,946,462
S 549,851
S 2,496,313
5.9 Household Costs
5.9.1 Overview
Household costs are the translation of the total cost to utilities to their customers. The previous sections
estimate total utility costs for the various elements of the 1ESWTR. mid section further refines the
analysis of the cost impacts of the rule by expressing utility costs as increases in annual costs to
individual households.
5.9.2 Methodology’
One estimated cost, turbidity treatment, complicates the calculation of househo’d costs because it is a
compilation of activities, each with a different cost to utilities. Two assumptions are made with respect to
turbidity treatment, therefore, in this analysis.
The underlying assumption that drives the turbidity treatment portion of the total cost analysis is that
compliance forecast activities to meet the rule requirements are more likely to occur together than not. In
other words, a utility is more likely to have to implement a group of activities rather than an individual
activity.
IFS WTR FnaI RL4
5-24
November 12, 1998
-------
A und assumption n oI’es an e\ceptlon to the tirst assumption The irst tour fUtration acti ines
Apperidi A are mutually exclust e and ‘ ouId not be implemented together Thus. a utilir’ outd
pertorm onl\ one olthe follo ing
Filter media addition.
Filter media entire replacement without support gravel,
Filter media and support gravel replacement; or
Filter media, support gravel, and underdrain replacement
The underlying assumption that activities are more likely to occur together than not has implications in
the choice of methodology. This assumption precludes the use of a simple average cost per household
(calculated by dividing the total turbidity treatment costs for each system size category by the number of
households) for several reasons. First, a simple average by size category underestimates the upper bound
of household costs Some systems within each category will be more likely to implement many, if riot all,
of the activities, thereby resulting in a much higher-than-average household cost for their rarepayers. An
alternative methodology is needed to capture the projected distribution of costs across treatment
alternatives,
The methodology used for this analysis assumes that a small percentage of systems within each size
..ategory will need to implement a/ I of the general treatment activities and one of the first four filtration
activities to comply. The next increment of systems are assumed to implement all bur one (the least
common) of the general treatment activities and one of the first four filtration activities to comply. The
process continues dropping out general treatment activities, until a final increment of systems only
implements the most common treatment activity. Once the range of activities was estimated, costs of
these activities were calculated. These system unit costs were then converted to household costs. The
final step repeats this process for each system size. The results are a list of the number of households at a
specific cost per household. These results are then graphed to display the cumulative distribution of
household costs. Detailed household cost estimates are presented in Appendix F.
5.9.3 Results of Household Cost Analysis
Under the IESWTR, households will face the increases in annual costs displayed in Exhibit 5.16. All
households served by large surface water systems will incur additional costs under the 1ESWTR since all
systems are required to perform turbidity monitonng activities. However, as shown in the cumulative
distribution of households affected by the rule, a large number (92 percent) of households will face a
maximum increase in cost of $12 per year ($1 per month). In other words, 60 million households will
incur no more than a $1 increase in their monthly costs. Five million households( percent) will face an
increase in cost of between $12 and $60 per year ($l-$5 per month). The highest cost faced by 23,000
households is approximately $100 per year ($8 per month).
The assumptions and structure of this analysis, in describing the curve, tend to overestimate the highest
costs. To be on the upper bound of the curve, a system would have to implement all, or almost all, of the
treatment activities. These systems, conversely, might seek less costly alternatives, such as connecting
into a larger regional water system. The TWG thought that this was an extreme situation, and the
resulting high values may occur only for a small number of households. In addition, even at the upper
IFS WTR Final RIA 5-25 November 12. 1998
-------
end the monthl cos per household is less than Sb per month. relati .elv small in comparison with other
common household expenditures
Exhibit 5.16
Cumulative Distribution of Annual Cost per Household of the IESWTR
100% _______________________________
90%
• 80% f $5 per month (99th percentile
70% ) $1 per month 92nd percentile )
60% _____________
•g 50% __________________________
40% ___________________________
30% _________________________
20% ________________________________________
c 10% _________________________________
0% _____________________
5- $20 $40 $60 $80 $100 $120
Annual Cost per Housihold
5.10 Combined Effect of the Stage 1 DBPR and the IESWFR
Because the IESWTR and Stage I DBPR were developed in tandem to address the risks of disinfection
byproducts while not compromising protection against microbial contaminants, it is important to
examine the combined effects of both rules as well as those rules expected to be implemented in the next
several years.
While the Stage I DBPR may impose additional costs to large surface water systems beyond those
described in this chapter for the IESWTR, these systems may see greater benefits as well. The
anticipated impact of both rules at a 7 percent cost of capital is summarized in Exhibit 5.17.
JESWTR Final PJA 5-26 November 12, 1998
-------
ExhibIt 5.1,
CosI Impact of Current and Expected Rule-Makings
System I pes
.
Current and Eapected Rules
D1DBP
Stage I ( OO0)
Interim
ESW’ 5000)
Other
Rule-makings Planned
Small Surface Water
$56 ,804
SO
STagc2DBPR
Long-term ESWTR I (LII)
Large Surface Water
278,321
291,165
Stage 2 DBPR
Long.cerm ESWT 2 (LT2)
Smail Ground Water
218.062
0
Stage 2 DBPR
Ground Water Disinfection
Large Ground Water
130,651
0
Stage 2 DBPR
Ground Water Disinfection
Subiotal
S6 3,83g
$291,165
Stares
17.342
15,556
Total.,
$701,180
S 306,721
IESJ+7R Final RJA
5-27
Noveffther 12, /998
-------
6: Net Benefits
This section provides a comparison of the benefit and cost outcomes with benefit/cost principles
Chapters 4 and 5 present quantitative summaries of the final benefit and cost impacts of the Interim
Enhanced Surface Water Treatment Rule.
The assessment of net benefits is always somewhat problematic due to the relative ease of quanti ing
compliance treatment costs versus the difficulty of assigning monetary values to the avoidance of health
damages and other benefits arising from the regulation. The challenge of assessing net benefits for the
IESWTR is compounded by thefact that there are areas of scientific uncertainty regarding the exposure
assessment and the risk assessment for Crypsosporzdium. Areas where important sources of uncertainty
enter the benefits assessment include the following.
Occurrence of Cryprosporidium cocysts in source waters;
Occurrence of Cryptosporidiurn oocysts in finished waters;
Reduction of Cryptospor dzum oocysts due to treatment, including filtration and disinfection;
Viability of C,ypzosporidium oocysts after treaunent;
Infectivity of Crypwsporid:um;
Incidence of infections (including impact of under reporting);
Characterization of the risk; and, -
Willingness-to-pay to reduce risk and avoid costs.
The benefits analysis attempts to take into account some of these uncertainties by estimating benefits
under two different current treatment assumptions and three improved removal assumptions. The
benefits analysis also used Monte Carlo simulations to derive a distribution of estimates, rather than a
single point estimate.
Exhibit 6.1 summarizes the annual cost of the rule at the 3, 7, and 10 percent costs of capital. Annual
utility costs at 7 percent are approximately S29 million and annual State costs are approximately $15
million.
Exhibit 6.2 summarizes the mean expected value of potential annual benefits expected to accrue to the
turbidity provisions under the six different scenarios, as well as the range. The range presented in the
exhibit represents the I 0 and 9O ’ percentiles of the calculated distribution of illnesses. Thus., the actual
number of illnesses ha a 10 percent probability of being as low or lower than the bottom end of the
range presented and as high or higher than the top of the range presented.
IES 4rTR Final RJA 6-1 November /2, 1998
-------
Exhibit 6.1
Summar’ of Costs under the Interim Enhanced Surface Water Treatment Rule (S000s )
Final Rule (1998 Ss)
3’!.
Cost of Capital
7/.
Cost of Capital
Cost of Capital
Utility Costs
S 8,965 S 758,965
Utility Treatment Capital
S 758.965]
Annual Costs
103.437
Anrn#.alized Capital ’
Annual O&M
Total Treatment
Turbidity Monitonng
Turbidity Excepuons’
Disinfection Benchmarking
Subtotal
65,999
105.943
171,942
95,924
195
2,841
270.902
85,611
105,943
l9 1,554
95,924
195
2,841
290,514
105.943
209,380
95,924
195
2,841
308.340
Annualized One-Time
Costs
405
504
Turbidity Monitoring Start-Up
HAA Bcnchmasting
Subtotal
289
175
464
246
651
306
810
S 309,150
Total Annual Utility Costs
S 271,366
S 291,165
State Coats
Annual Coats
5,256
Turbidity Monitonng
Turbidity Excepuons”’ ‘
Sanitary Survey
Disinfection Benclamarking
Subtotal
5,256
409
6,979
2,789
15,433
5,256
409
6,979
2,789
15,433
409
6,979
2,789
/5.433
Annualized One-Time
Costs***
48
Turbidity Monitoring Start-Up
Disinfection Bcnchmarkiflg Start-Up
Sanitary Survey Start-lip
Subtotal
27
22
39
88
30
- 55
/23
38
69
155
$ 15,588
Total Annual State Costs
S 15,521 S 15,556
Total Annual Coats S 286,387 S 306,721 $ 324,738
• Capital costs are annualized over 20 years with the exception of nubsdimetres and process control modific iofl etpupment, which age
annualized over 7 years ______
es Costs associated with Individual Filter Effluent Turbidity Rcquudnents foci ceptiOns reporting and Individual Filter Assessments
All one-time costs age annualized over 20 years
Costs associated with Reponing Exceptions and Comprehensive PerfonnanCe Evilugicits.
IESWTR Final RIA 6-2 November 12, 1998
-------
E’thibit 6.2 Summar of Potential Annual Benefits
.
Baseline Assumes...
2.5 Log Crypi’osporzdium
Removal
3.0 Log Ciyptosporidusm
Removal
Mean
Range
Mean
Range
Crylposporiduosis Illness Avoided Annually
LOW Number of Illnes ses Avoided
Cost of Illness Avoided
MID Number of illnesses Avoided
Cost of illness Avoided
HiGH Number of illnesses Avoided
Cost of Illness Avoided
338,000
SO 950 billion
432.000
S 1172 billion
463,000
51 240 billion
0- 1,029,000
SO - I 883 billion
0. 1.074,000
SO - 1 960 billion
0- 1,080.000
SO-I 999 billion
1 10.000
30 263 billion
141,000
SO 327 billion
152,000
SO359billion
0-322.500
SO - 0585 billion
0- 333,000
SO - 0608 billion
0-338,000
50-0620 billion
V.lue of Crytposperidiosis Mortalities
Avoided Annually
LOW Number of Mortalities Avoided
Value of Mortalities Avoided
MID Numbcrof Mortalities Avoided
Value of Mortalities Avoided
HIGH NumberofMortajitje sAyoided
Value of Mortalities Avoided
48
SO 272 billion
60
SO 341 billion
64
30 363 billion
0. 129
SO - 0 674 billion
0-135
SO -0 706 billion
0- 136
SO -0 708 billion
14
$0085 billion
IS
30.107 billion
20
30.115 billion
0 -40
30. 0 209 billion
0-42
SO. 0219 billion
0.42
30. 0 221 billion
Reduced Risk of Cryipospondlosu
Outbreaks
Cost of Illness Avoided
Emergency Expendrnircs
Liability Costs
Bcneliu not quantified, but could be substantial for large outbreak (SO 800
billion cost of ilIn avoided foes Milwaukee-level outbreak)
Reduced Risk front Other Psthogena
Benefits not quantified.
Enhanced Aesthetic Water Quality
Difference may not be noticeable to
Avertin Behavior
3
consumer
Bcnc&s not quantified, but could be substantial for large outbreak SO 020
billion to 50062 billion for a Milwaukee-level
Given the costs summartzed in Exhibit 6.1 and the benefits assuming a mean number of illnesses avoided
summanzed in Exhibit 6.2, the recommended rule results in positive net benefits under all three
improved removal scenarios (low, mid, and high) assuming that current treatment achieves a removal of
2.5 logs, taking into account only the value of cost of illness (CO!) avoidvd. Using a current treatment
removal assumption of 3.0 logs, net benefits are negative under the low improved removal assumption
using only the value of COt avoided. When the value of mortalities prevented is added into the benefits,
all baseline assumptions and removal scenarios have positive net benefits at the mean.
Thus, the monetized net benefits are positive across the range of current treatment assumptions,
improved log removal scenatios, and cost of capital rates. The benefits due to the illnesses avoided may
be slightly overstated because the mortalities were not netted out of the number of illnesses. This value is
minimal and would not be captured at the level of significance of the analysis. Several categories of
benefits, including reducing the risk of outbreaks, reducing exposure to other pathogens such as Gia?dia,
and avoiding the cost of averting behavior have not been quantified for this analysis, but could represent
substantial additional economic value. In addition, the estimates for avoided COl do not include the
value for pain and suffering or the risk premium.
/FS t7R Final &FA 6-3 November /2, /998
-------
7: The Economic Rationale for Regulation
7.1 Introduction
This section of the RJA discusses the statutory authority on the economic rationale for choosing a
regulatory approach to protect public health from drinking water contamination. The economic rationale
is provided in response to Executive Order Number 12866, Regulatory Planning and Review, 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 assessthe
significance of that problem (Sect. I b(1)).
In addition, 0MB Gu;dance dated January II, 1996, stales that “in order to establish the need for the
proposed action, the analysis should discuss whether the problem constitutes a significant market failure
(p 3).” Therefore, the economic rationale laid out in this section should not be interpreted as the
Agency’s approach to implementing the Safe Drinking Water Act. Instead, it is the Agency’s economic
analysis, as required by the Executive Order, to support a regulatory approach to the public health issue
at hand.
7.2 Statutory Authority for Promulgating the Rule
The 1996 reauthorization for the Safe Drinking Water Act (SDWA) mandated new drinking water
requirements. EPA’s general authority to set Maximum Contaminant Level Goals (MCLGs) and the
National Primary Drinking Water Rule (NPDWR) was modified to apply to contaminants that “may have
an adverse effect on the health of persons,” are “known to occur or there is a substantial likelihood that
the contaminant will occur in public water systems with a frequency and at levels of public health
concern,” and for which “in the sole judgment of the Administrator, regulation of such contaminant
presents a meaningful opportunity for health risk reductions for persons served by public water systems”
(1996 SDWA, as amended).
The 1996 Amendments also require the promulgation of the interim Enhanced Surface Water Treatment
Rule (IESWTR) and a Stage I Disinfectants/Disinfection Byproducts Rule (Stage I DBPR) by
November 1998. In addition, the 1996 Amendments require EPA to promulgate a Final Enhanced
Surface Water Treatment Rule and a Stage 2 DBPR by November 2000 and May 2002, respectively.
7.3 The Economic Rationale for Regulation
In addition to the statutory directive to regulate microbial contaminants, there is also economic rationale
for government regulation. The need for government regulation often results from an imperfection in the
market’s ability to provide safe water at price levels that efficiently satisfy consumer needs. in a
perfectly competitive market, market forces guide buyers and sellers to attain the best possib’e social
outcome. A perfectly competitive market occurs when there are many producers of a product selling to
!ESWTR Final )?JA 7-i November 12. 1998
-------
mans bu ers. and both producers and consumers have complete knowledge regarding the products of
each firm There must also be no barriers to entry in the tndustr . and firms in the industry must not have
any ad antage over potential new producers Several factors in the public water supply industry do not
satisf the requirements for a perfect market and lead to market failures that require regulation. -
First, the public water market is a very limited competitive market with monopolistic tendencies. These
monopolies tend to exist because it is not economically efficient to have multiple suppliers competing to
build multiple systems of pipelines, reservoirs, wells, and other facilities. Instead, a single firm or
government entity performs these functions under public control. Under monopolistic conditions,
consumers are provided only one level of service with respect to the quality attribute of the product., in
this case drinking water quality. If they do not believe the margin of safety in public health protection is
adequate, they cannot simply switch to another water utility
Second, there are high information and transaction costs that impede public understanding of the health
and safety issues concerning drinking water quality. The type of health risks potentially posed by trace
quantities of drinking water contaminants involve analysis and distillation of complex toxicological data
and health sciences. EPA is currently in the final stages of developing the Consumer Confidence Report
rule that will make water quality information more easily available to consumers. The Consumer
Confidence Report rule will require community water systems to mail their customers an annual report
on local drinking water quality. However, consumers would still have to analyze this information for its
health risk implications. Even if informed consumers are able to engage utilities regarding these health
issues, the costs of such engag rnent-transaCtiOfl costs (measured in personal time and commitment)
present another significant impediment to consumer expression of risk preference.
SDWA regulations are intended to provide a level of protection from exposure to drinking water
contaminants that would not otherwise occur in the existing market environment of public water supply.
The regulations set minimum performance requirements for all public water supplies in order to protect
all consumers from exposure to contaminants. SDWA regulations are not intended to restructure flawed
market mechanisms or to establish competition in supply. While these distortions are essential conditions
in weighing perceptions of benefits and costs, SDWA regulations do not attempt to correct market
imperfection directly. Rather, SDWA standards establish the level of service to be provided in order to
better reflect public preferences for safety. The Federal regulations remove the high information and
transaction costs by acting on behalf of all consumers in balancing the risk reduction and the social costs
of achieving this reduction.
1ESW ’TR Final RIA 7-2 November 12. 1998
-------
References
58 FR 51735 Executive Order 12866 “Regulatory Planning and Review” October 4, 1993
59 FR 7629 Executive Order 12898. “Federal Actions to Address Environmental Justice in Minority
Populations and Low income Populations.” February 16, 1994.
54 FR 27544 Total Coliforni Rule. June 29, 1989.
American Water Works Association. “Safe Drinking Water Act.” August 6, 1996.
Archer, J R., et al. 1995. “Cryptosporidium spp. Oocysts and Giardia spp. Cyst Occurrence:
Concentrations and Distribution in Wisconsin Waters.” Wisconsin Department of Natural Resources:
PUBL-WR420 -95
Atherholt, 1., M.W. LeChevallier and W.D. Norton. 1995. “Survey of Surface Waters for Giardia and
C ypiosporzd:um and Water Treatment Efficiency Evaluation.” Research Project Summary. Prepared for
New Jersey Department of Environmental Protection.
Casemore, D. P. and F. B. Jackson. 1984. “Hypothesis: Cryptosporidiosis in Human Beings is not
Primarily a Zoonosis.”J. Infect. 9: 153-156.
Centers for Disease Contro’. 1996. “Survesllance for Waterborne-Disease Outbreaks—United States,
1993-1994.” 45(SS-1).
Chestnut, L.G., A. Alberini, Ct al. 1997. “Monetary Valuation of Human Mortality Risks in Cost-Benefit
Analyses of Environmental Programs.”
Cordell, R.L and D.G. Addiss. 1994. “Cryptosporidiosis in Child Care Settings: A Review of the
Literature and Recommendations for Prevention and Contml.” Pediatrics Infectious Disease Journal: 13.
Current, W.L., N.C. Reese, J.V. Ernst, W.S. Bailey, M.B. Heyman and W.M. Weinstein. 1983. “Human
Cryptosporidiosis in lmmunocompetents and Immunodeficient Persons. Study of an Outbreak and
Expenmental Transmission.” New England Jownal of Medicine: 308.
D’Antonio, R.G., R .E. Winn, I.P. Taylor, et al. 1985. “Outbreak of Cryptosporidiosis in Normal Hosts.”
Ann. Intern. Med. 103:886-888.
DuPont, H.L., C. Chappell, C. Sterling, P. Okhuyscn. LB. Rose and W. iakubowski. 1995. “Infectivity of
Crypvospordiumpcirvum in Healthy Volunteers,” New England Journal of Medicine: 332.
DuPont, H.L. 1997. Presentation at EPA Workshop.
Fayer, R. and B.L. Ungar. 1986. “Cryptosporidiuin spp. And Cryptosporidiosis.” Microbial Review: 50.
IFS WTR Final RM References November /2. 1998
-------
References
Fox. K R and D. A. Lvtle 1996 “Milwaukee’s Crypto Outbr k Investigation and Recommendations.”
Journal of the American Water Works Association. 88(9) 87-94
Freeman, A.M. 1979. “The Benefits of Environmental Improvement: Theory and Practice.” Johns
Hopkins University Press Resources for the Future.
Gerba, C.P, J B. Rose and C.N. Haas. 1996. “Sensitive populations: who is at the greatest risk?”
International Journal of Food Microbiology: 30(1.2).
Grubbs, W D., B. Macler and S. Regli. 1992. “Modeling Giardia Occurrence and Risk.” EPA-811-B-92.
005. Washmgton, D.C.: Office of Water Resource Center.
Guerrant, R.L. 1997. “Cryptosporidiosis: An Emerging, Highly Infectious Threat.” Emerging Infectious
Diseases: 3(1).
Haas, C.N., C.S. Crockett, LB. Rose, C.P. Gerba and A.M. Fazil. 1996. “Assessing the Risk Posed by.
Oocysts in Drinking Water.” Journal of the American Waier Works Association.
Haas, C.N. and J.B. Rose. 1995. “Developing an Action Level for Crypsosporidiwn.” Journal of the
American Water Works Association: 89(9).
Harrington, W.; A.J. Krupnick and W.O. Spofford, Jr. 1985. “The Benefits of Preventing an Outbreak of
Giardiasis Due to Drinking Water Contamination.” EPA/Resources for the Future Report
Harrington, W., A.J. Krupnick and W.O. Spofford, Jr. 1989. “The Economic Loss of a Waterbome
Disease Outbreak.” Journal of Urban Economics: 25.
Hoxie, N.J., J.P. Davis and J.M. Vegeront. 1996. “Cryptosporidiosis-Associated Mortality Following a
Massive Outbieak in Milwaukee, Wisconsin.” American Journal of Public Heaiih: 87(12).
Juranek, D.D. 1995. “Cryptosporidiosis: Sources of Infection and Guidelines for Prevention.” Clinical
Infectious Diseases: 21(1).
LeChevallier, M.W. and W.D. Norton. 1995. “Giardia and Cryp:osporidivm in Raw and Finished
Water.” Journal of the American Water Works Association: 87.
LeChevallier, M.W. and W.D. Norton. 1992. “Examining Relationships Between Particle Counts and
Giard:a, Cryprosporidium, and Turbidity.” Journal of the American Water Works 4.rsocwi son. 87(9): 54..
68.
LeChevallier, M.W., W.D. Norton and R.G. Lee. 1991 a. “Giardia and Cryptasporidiun spp. in Filtered
Drinking Water Supplies.” Applied Environmental Microbioiogy 57(9).
/&SWTR Final P.M References November 12. 1998
-------
References
LeChevallier. M W. ‘l v D Norton and R G Lee 1991b “Occurrence of Giardia and Crvptosporidiuin
spp in Surface Water Supplies Applied Environmental M,crob,olo ir 57(9) 2617-2621
LeChevaiher. M W, W I) Norton and R. G Lee, 199 Ic. “Occurrence of Giard:a and Crypiosporidiwn
spp. in Surface Water Supplies.” Appi. Environ. Microbial. 57(9): 2610-2616.
Mackenzie, W , N Hoxie, M Proctor, M Gradus, K. Blari, D. Peterson, J. Kazmierczak, D Addiss, K.
Fox, J Rose and i. Davis 1994. “A Massive Outbreak in Milwaukee of Cryptosporidiurn infection
Transmitted Through the Public Water Supply.” New England Journal of Medicine: 331.
Madore, M.S., LB. Rose, C.P. Gerba, M.J. Arrowood and C.R. Sterling. 1987. “Occurrence of
Crypwspor:dsurn Oocysts in Sewage Effluents and Selected Surface Waters.” Journal Parassi: 73(4).
Mauskopf, J. and M. T. French. 1991. “Estimating the Value of Avoiding Morbidity and Mortality from
Foodborne Illnesses.” Risk Analysis. 11:619-631.
National Research Council. 1983. “Risk assessment in the Federal Government Managing the Process.”
Washington, DC.: National Academy Press.
Norton, W. D. and M.W LeChevallier. 1997. “Survey of Open Finished Water Reservoirs for Giardia
Cysts and Cryptosporidiwn Oocysts.” Prepared for the New Jersey Department of Environmental
Protection. September.
Okhuysen, P.C., C.L. Chappell, C.R. Sterling, W. Jakubauski and H.L. DuPont. 1998. “Susceptibility and
Serologic Response of Health Adults to Reinfection with Cryptospor:diumpar .” Infection and
Immunity. 66(2).
Okun, D.A., G.F. Craun, i.K. Edzwald , LB. Gilbert andi.B. Rose. 1997. “New York City: To Filter or
not to Filter?” Journal of the American Water Work, Association: 89(3).
Office of Management and Budget; Guidance. “Economic Analysis of Federal Regulations Under
Executive Order 12866.” Januaiy 11, 1996.
Ongerth, i.E. and H.H. Stibbs. 1987. identification of Cryprosporidium Oocysts in River Water.”
Applied Environmental Microbiology 53(4).
Patania, N.L., 1G. Jacangelo, L. Cummings, A. Wilczak, K. Riley and J. Oppenheimer. 1995.
“Optimization of Filtration for Cyst Removal.” AWWARF.
Perz, J F., F.K. Ennerver and S.M. LeBlancq. 1998. “Cryptospondium in Tap Water Comparison of
Predicted Risk with Observed Levels of Disease.” American Journal of Epidemsology 147(3).
iFS *TR Final RIA References November 12, 1998
-------
References
Regli S . B A v1acler. J E Cromwell. X Zhang, A B Gelderoos. W D Grubbs and F. Letkiewicz. 1993.
Frarnework for Decision Making EPA Perspective” In G F Craun, ed. Safety of Water Disinfection
Balancing Che,n,ca/ and Microbial Risk. Washington, DC: International Life Sciences Institute Press.
pp 487-538
Rose. J.B 997 “Environmental Ecology of Cryptosporidium and Public Health Implications.” Annual
Review of Public Health: 18.
Rose, J.B. 1988. “Occurrence and Significance of Cryptosporidium in Water.” Journal of the American
Water Works Association: 79(1).
Rose, J.B., H. Darbin and C. P. Gerba. 1988a. “Correlations of the Protozoa Crypiosporidium and
Giardia with Water Quality Variables in a Watershed.” Proc. mt. Conf. Water Wastewater Microbiol.
Newport Beach, CA.
Rose, J.B,, D. Kayed, M. S. Madore, C. P. Gerba, M.J. Arrowood and C. R. Sterling. 1988b. “Methods
for the Recovery of Giardia and Cr p:osporzdium from Environmental Waters and Their Comparative
Occurrence,” In: P. Wallis and B. Hammond, eds. Advance.c in Giardia Research. Calgary, Canada:
University of Calgary Press.
Unfunded Mandates Reform Act of 1995 (UMRA), P.L. 104-4. Title II. (Also Section 202).
U.S. Bureau of Labor Statistics. 1997. http://www.bls.gov/
U.S. Environmental Protection Agency. July 28, 1998. “Costs and Technology Document for the Interim
Enhanced Surface Water Treannent Rule.” Office of Ground Water and Drinking Water.
U.S. Environmental Protection Agency. July 1998a. “Occurrence Assessment of the Interim Enhanced
Surface Water Treatment Rule.”
U.S. Environmental Protection Agency. June 24, 199$a. “Occurrence Assessment for the Interim
Enhanced Surface Water Treatment Rule.” Prepared for Office of Ground Water and Drinking Water.
U.S. Environmental Protection Agency. July 1998b. “Technologies and Costs for the Interim Enhanced
Surface Water Treatment Rule.”
U.S. Environmental Protection Agency. November 3, 1997. IESWTR Notice of Data Availability.
Federal Register.
U.S. Environmental Protection Agency. 1997. “The Benefits and Costs of the Clean Air Act, 1970..
1990.” Prepared for U.S. Congress.
U.S. Environmental Protection Agency. 1996. “An Evaluation of the Statistical Performance of a Method
for Monitoring Protozoan Cysts in US Source Water.”
IESWTR Final PJA References November 12, 1998
-------
Refe re rices
U S Environmental Protection Agency 1995 “State Guidance on Sanitary Surveys.” December 1995
U S Environmental Protection Agency 1994 “National Primary Drinking Water Regulations: Enhanced
Surface Water Treatment Requirements, Proposed Rule.” 59 FR 38832. July 29, 1994.
U S. Environmental Protection Agency/Science.Advisory Board. 1990. “Reducing Risk: Setting
Priorities and Strategies for Environmental Protection.”
Van Hourven, G.L., J C. Whitehead, 1.14. Bingham and B. Depro. 1997. “Vaiuing Dnriking Water
Beneflts: Theory, Methods, and Research Needs.” Draft Report.
Van Houtven, G.L. 1997. ‘ Altruistic Preferences for Life-Saving Public Programs: Do Baseline Risks
Matter.” Risk Analysis. 17:85-92.
Via, S. 1997. “Estimate of Disinfection BenchmarkingfProfihing Cost.” American Water Works
Association memorandum.
!ESWTR Final RM References Novermber J2. 1998
------- |