UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON D.C. 20460
OFFICE OF THE ADMINISTRATOR
SCIENCE ADVISORY BOARD
March 19, 2010
EPA-SAB-10-004
The Honorable Lisa P. Jackson
Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, N.W.
Washington, DC 20460
Subject: Review of EPA's Supporting Analysis for Revisions of the Total Coliform Rule
Dear Administrator Jackson:
The Total Coliform Rule (TCR) was published in 1989 and became effective in 1990.
The TCR requires routine sampling by all public water systems (PWSs) to monitor for the
presence of total coliforms in the distribution system. The monitoring of total coliforms is used
as an indicator of the adequacy of water treatment and the integrity of the distribution system.
The TCR is mandated by law to be reviewed every six years as part of the existing drinking
water regulations requiring revisions to improve/maintain public health protection. In
preparation for the TCR revisions, EPA assembled a committee of stakeholders that met over
several months to address the TCR objectives and enhance the approach for protecting public
health. As a result of these efforts, an Agreement in Principal (AIP) which takes a more
proactive approach to identifying and fixing problems was developed. The TCR revisions are
intended to improve the effectiveness of monitoring requirements by balancing public health
gains from proactive find-and-fix requirements with reduced sampling requirements. Under the
revised TCR, TC-positive sampling results would prompt an assessment to identify sanitary
defects instead of having a Maximum Contaminant Level Goal/Maximum Contaminant Level
(MCLG/MCL) for TC.
In response to a request from EPA's Office of Water, the Science Advisory Board (SAB)
convened the Drinking Water Committee with additional experts from the Board to conduct a
review of EPA's proposed revision of the Total Coliform Rule (rTCR). The SAB committee
during their deliberations examined the supporting analysis, AIP and draft chapters of the
proposed rTCR. Specifically the SAB was asked to comment on:
(1) the statistical analysis of the monitoring data which were used to estimate the positive result
rates for total coliform (TC) and E. coli (EC) samples and predict the underlying baseline;
-------
(2) the list of corrective actions as described by the EPA that PWSs will implement and the
percentage of systems that will implement certain types of corrective actions;
(3) the methodology and assumptions used to estimate the net regulatory impacts of the current
TCR (with and without the effects of other EPA water regulations) as compared to the
recommendations contained in the Agreement in Principle (AIP) or the Alternative Analysis
(AA) which has more frequent routine monitoring requirements than the AIP; and
(4) the use of the reduction mE. coli- and TC-positive sample results and acute violations as an
indicator of benefits.
In general, the Committee was impressed by the work the Agency undertook. The
Agency obviously did a great deal of work and put a significant amount of thought into making
use of the limited amount of data. The Agency's engagement with stakeholders and
development of the AIP which addresses goals for improvement and research needs was laudable.
The Committee, however, also found that there were several areas that can be enhanced;
therefore, in their report, they provide a number of recommendations that are aimed at making
the rTCR more transparent and improve the scientific basis for the proposed revisions. While a
more detailed description of the technical recommendations is contained in the body of the report,
the key points and recommendations are highlighted below.
In response to the first charge question, the Committee found that while the Agency's
statistical analysis of the TCR data from monitoring samples used to estimate the underlying
baseline of TC- and EC- positive result rates was reasonable, it would benefit from some
refinements and further explanation. The Committee suggests that some of the key issues should
be further explored through sensitivity analyses. In addition to the Agency's current analysis,
these sensitivity analyses would provide a more robust basis for predicting and describing TC-
and EC- positive result rates.
In response to the second charge question, the Committee found the types of corrective
actions that systems would implement was reasonable as described by the Agency. The
Committee recommends that the Agency explain how this list of corrective actions was compiled.
The Agency, through its discussion with state representatives, has estimated that only 10% of the
small systems are able to find the cause of TC- or EC- positive samples and as a result will take
corrective action. The Committee believes that the rTCR should ideally be revised and designed,
so that a larger percentage of small systems would take corrective actions that result in long-term
benefits. Since systems may take corrective actions without knowing the exact cause, such as
implementation of a flushing program, the Committee recommends that the sensitivity analysis
use a wider range of implementation of corrective actions up to around 50%, especially after a
more rigorous examination of the system. Conversely, the Committee questions whether the
benefits of flushing will result in reduced risks for multiple years as used in the model, since
PWSs may take actions to correct sanitary defects, yet still continue to have TC-positive results
for a period of time.
In response to the third charge question, the Committee believes that a more robust
database than the current six-year data, that was specifically and voluntarily obtained and used in
the impact analysis by the Agency, is required to address the long list of assumptions. The
Committee recommends that the impact analysis and discussion should include identifying and
11
-------
summarizing all the variables with known modeling distributions. It should also be noted that
until more data are available, the Agency's model may be used to establish a reference baseline,
but should not be used as a predictive tool. The Committee also noted that an increased risk of
waterborne disease outbreaks may occur as a result of the proposed reduced monitoring strategy
and therefore, the anticipated savings in the annual national costs for the detection assays may
not be very significant. Furthermore, once a PWS is placed in the reduced monitoring regimen,
it is highly unlikely, unless the sanitary defect is egregious, that the system will be triggered into
more frequent sampling. The Committee is concerned that TC- or EC- positive sample results
will not be detected in a timely fashion by a quarterly or annual sampling plan. Overall, the AA
with its more frequent monitoring requirements appears to address public health protection
sooner in time than the AIP proposed implementation.
In responding to the forth charge question, the Committee felt that because of the
difficulties in linking TC- and EC- positive results to adverse human health outcomes, it is
equally difficult to determine whether measurable health-related benefits could be attributable to
the reductions in monitoring violations (positive results). TC monitoring results are not effective
sole endpoints for measuring benefits. The Agency's proposal of using TC as part of an overall
treatment technique evaluation where sanitary defects are identified and corrected as result of
TC-positive sampling results instead of having an MCLG/MCL for TC, is seen as a positive step
forward. The Committee recommends that other indicators should also be considered since TC
is not an adequate measure of health risk. While E.coli is viewed as a more appropriate measure
of risk of enteric disease, it does not capture the health risks from other pathogens, such as
Legionella, for example. Other measures, including structural and hydraulic integrity of
distribution systems, have been recently considered in a report by the National Research Council
and may provide valuable supplemental information on health risk associated with distributed
water.
The SAB appreciates the opportunity to provide EPA with advice on this important
subject. We look forward to receiving the Agency's response.
Sincerely,
/Signed/
Dr. Deborah L. Swackhamer, Chair
EPA Science Advisory Board
/Signed/
Dr. Joan Rose, Chair
SAB Drinking Water Committee
in
-------
NOTICE
This report has been written as part of the activities of the EPA Science Advisory Board, a public
advisory committee providing extramural scientific information and advice to the Administrator and
other officials of the Environmental Protection Agency. The Board is structured to provide balanced,
expert assessment of scientific matters related to problems facing the Agency. This report has not
been reviewed for approval by the Agency and, hence, the contents of this report do not necessarily
represent the views and policies of the Environmental Protection Agency, nor of other agencies in the
Executive Branch of the Federal government, nor does mention of trade names or commercial
products constitute a recommendation for use. Reports of the EPA Science Advisory Board are
posted on the EPA Web site at: http://www.epa.gov/sab.
IV
-------
U.S. Environmental Protection Agency
Science Advisory Board
Drinking Water Committee
Augmented for the Review of
EPA's Supporting Analysis for Revisions of the Total Coliform Rule
CHAIR
Dr. Joan B. Rose, Professor and Homer Nowlin Chair for Water Research, Department of
Fisheries and Wildlife, Michigan State University, East Lansing, MI
MEMBERS
Dr. Mark Borchardt*, Director, Public Health Microbiology Laboratory, National Farm
Medicine Center, Marshfield Clinic Research Foundation, Marshfield, WI
Dr. John (Jack) Colford, Professor, Division of Public Health, Biology & Epidemiology,
School of Public Health, University of California, Berkeley, CA
Dr. Penelope Fenner-Crisp, Independent Consultant, North Garden, VA
Dr. Joseph R. Landolph, Jr., Associate Professor, Molecular Microbiology and Immunology
and Pathology, Keck School of Medicine and Associate Professor of Molecular Pharmacology
and Pharmaceutical Science, School of Pharmacy, and Member, USC/Norris Comprehensive
Cancer Center, University of Southern California, Los Angeles, CA
Dr. Desmond F. Lawler, Bob R. Dorsey Professor of Engineering, Department of Civil,
Architectural and Environmental Engineering, University of Texas, Austin, TX
Dr. Christine Moe, Eugene J. Gangarosa Professor of Safe Water and Sanitation, Hubert
Department of Global Health, Rollins School of Public Health, Emory University, Atlanta, GA
Dr. Christine Owen, Water Quality Assurance Officer, Tampa Bay Water, Clearwater, FL
Dr. Richard Sakaji, Manager, Planning and Analysis for Water Quality, East Bay Municipal
Utility District, Oakland, CA
Dr. Gary Sayler, Beaman Distinguished Professor of Microbiology, and Ecology and
Evolutionary Biology, Director of the Joint Institute for Biological Sciences and Director of
the Center for Environmental Biotechnology, Oak Ridge National Laboratory, University of
Tennessee
Dr. Gina Solomon, Senior Scientist, Health and Environment Program, Natural Resources
Defense Council, San Francisco, CA
-------
Ms. Susan Teefy, Principle Engineer, Water Quality and Treatment Solutions, Inc., Canoga Park,
CA
Dr. Lauren Zeise, Chief, Reproductive and Cancer Hazard Assessment Branch, Office of
Environmental Health Hazard Assessment, California Environmental Protection Agency,
Oakland, CA
* provided written comments but did not attend the face-to-face meeting
SCIENCE ADVISORY BOARD STAFF
Dr. Sue Shallal, Designated Federal Officer, 1200 Pennsylvania Avenue, Washington, DC,
Phone: 202-343-9977, Fax: 202-233-0643, (shallal.suhair@epa.gov)
VI
-------
U.S. Environmental Protection Agency
Science Advisory Board
CHAIR
Dr. Deborah L. Swackhamer, Professor and Charles M. Denny, Jr., Chair in Science,
Technology and Public Policy and Co-Director of the Water Resources Center, Hubert H.
Humphrey Institute of Public Affairs, University of Minnesota, St. Paul, MN
SAB MEMBERS
Dr. David T. Allen, Professor, Department of Chemical Engineering, University of Texas,
Austin, TX
Dr. Claudia Benitez-Nelson, Associate Professor, Department of Earth and Ocean Sciences and
Marine Science Program, University of South Carolina, Columbia, SC
Dr. Timothy Buckley, Associate Professor and Chair, Division of Environmental Health
Sciences, College of Public Health, The Ohio State University, Columbus, OH
Dr. Thomas Burke, Professor, Department of Health Policy and Management, Johns Hopkins
Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD
Dr. Deborah Cory-Slechta, Professor, Department of Environmental Medicine, School of
Medicine and Dentistry, University of Rochester, Rochester, NY
Dr. Terry Daniel, Professor of Psychology and Natural Resources, Department of Psychology,
School of Natural Resources, University of Arizona, Tucson, AZ
Dr. George Daston, Victor Mills Society Research Fellow, Product Safety and Regulatory
Affairs, Procter & Gamble, Cincinnati, OH
Dr. Costel Denson, Managing Member, Costech Technologies, LLC, Newark, DE
Dr. Otto C. Doering III, Professor, Department of Agricultural Economics, Purdue University,
W. Lafayette, IN
Dr. David A. Dzombak, Walter J. Blenko Sr. Professor , Department of Civil and
Environmental Engineering, College of Engineering, Carnegie Mellon University, Pittsburgh,
PA
Dr. T. Taylor Eighmy, Vice President for Research, Office of the Vice President for Research,
Texas Tech University, Lubbock, TX
Dr. Elaine Faustman, Professor, Department of Environmental and Occupational Health
Sciences, School of Public Health and Community Medicine, University of Washington, Seattle,
WA
vn
-------
Dr. John P. Giesy, Professor and Canada Research Chair, Veterinary Biomedical Sciences and
Toxicology Centre, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
Dr. Jeffrey Griffiths, Associate Professor, Department of Public Health and Community
Medicine, School of Medicine, Tufts University, Boston, MA
Dr. James K. Hammitt, Professor, Center for Risk Analysis, Harvard University, Boston, MA
Dr. Rogene Henderson, Senior Scientist Emeritus, Lovelace Respiratory Research Institute,
Albuquerque, NM
Dr. Bernd Kahn, Professor Emeritus and Associate Director, Environmental Radiation Center,
School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA
Dr. Agnes Kane, Professor and Chair, Department of Pathology and Laboratory Medicine,
Brown University, Providence, RI
Dr. Nancy K. Kim, Senior Executive, New York State Department of Health, Troy, NY
Dr. Catherine Kling, Professor, Department of Economics, Iowa State University, Ames, IA
Dr. Kai Lee, Program Officer, Conservation and Science Program, David & Lucile Packard
Foundation, Los Altos, CA
Dr. Cecil Lue-Hing, President, Cecil Lue-Hing & Assoc. Inc., Burr Ridge, IL
Dr. Floyd Malveaux, Executive Director, Merck Childhood Asthma Network, Inc., Washington,
DC
Dr. Lee D. McMullen, Water Resources Practice Leader, Snyder & Associates, Inc., Ankeny,
IA
Dr. Judith L. Meyer, Distinguished Research Professor Emeritus, Odum School of Ecology,
University of Georgia, Lopez Island, WA
Dr. Jana Milford, Professor, Department of Mechanical Engineering, University of Colorado,
Boulder, CO
Dr. Christine Moe, Eugene J. Gangarosa Professor, Hubert Department of Global Health,
Rollins School of Public Health, Emory University, Atlanta, GA
Dr. Eileen Murphy, Manager, Division of Water Supply, New Jersey Department of
Environmental Protection, Trenton, NJ
Vlll
-------
Dr. Duncan Patten, Research Professor , Department of Land Resources and Environmental
Sciences, Montana State University, Bozeman, MT
Dr. Stephen Polasky, Fesler-Lampert Professor of Ecological/Environmental Economics,
Department of Applied Economics, University of Minnesota, St. Paul, MN
Dr. Stephen M. Roberts, Professor, Department of Physiological Sciences, Director, Center for
Environmental and Human Toxicology, University of Florida, Gainesville, FL
Dr. Amanda Rodewald, Associate Professor, School of Environment and Natural Resources,
The Ohio State University, Columbus, OH
Dr. Joan B. Rose, Professor and Homer Nowlin Chair for Water Research, Department of
Fisheries and Wildlife, Michigan State University, East Lansing, MI
Dr. Jonathan M. Samet, Professor and Flora L. Thornton Chair, Department of Preventive
Medicine, University of Southern California, Los Angeles, CA
Dr. James Sanders, Director and Professor, Skidaway Institute of Oceanography, Savannah,
GA
Dr. Jerald Schnoor, Allen S. Henry Chair Professor, Department of Civil and Environmental
Engineering, Co-Director, Center for Global and Regional Environmental Research, University
of Iowa, Iowa City, IA
Dr. Kathleen Segerson, Professor, Department of Economics, University of Connecticut, Storrs,
CT
Dr. V. Kerry Smith, W.P. Carey Professor of Economics , Department of Economics , W.P
Carey School of Business , Arizona State University, Tempe, AZ
Dr. Herman Taylor, Professor, School of Medicine, University of Mississippi Medical Center,
Jackson, MS
Dr. Barton H. (Buzz) Thompson, Jr., Robert E. Paradise Professor of Natural Resources Law
at the Stanford Law School and Perry L. McCarty Director, Woods Institute for the Environment,
Stanford University, Stanford, CA
Dr. Paige Tolbert, Associate Professor, Department of Environmental and Occupational Health,
Rollins School of Public Health, Emory University, Atlanta, GA
Dr. Thomas S. Wallsten, Professor and Chair, Department of Psychology, University of
Maryland, College Park, MD
Dr. Robert Watts, Professor of Mechanical Engineering Emeritus, Tulane University,
Annapolis, MD
IX
-------
SCIENCE ADVISORY BOARD STAFF
Mr. Thomas Miller, Designated Federal Officer, 1200 Pennsylvania Avenue, NW
1400F, Washington, DC
-------
ACRONYMS
AA
AIP
CWS
DWC
EA
EC
EPA
GW
GWDR
HPC
MCL
MCLG
PCR
PWS
rTCR
SAB
SDWIS/FED
SW
SWTR
TC
TCR
TCRDSAC
TT
TWG
UA
Alternative Analysis
Agreement in Principle
Community Water System
Drinking Water Committee
Economic Assessment
Escherichia coli (E. Coli)
Environmental Protection Agency
Ground Water
Ground Water Disinfection Rule
Heterotrophic Plate Count
Maximum Contaminant Level
Maximum Contaminant Level Goal
Polymerase Chain Reaction
Public Water System
revised Total Coliform Rule
Science Advisory Board
Safe Drinking Water Information System /Federal Version
Surface Water
Surface Water Treatment Rule
Total Coliform
Total Coliform Rule
Total Coliform Rule / Distribution System Advisory Committee
Treatment Technique
Technical Working Group
Uncertainty Analysis
XI
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY 1
BACKGROUND ON THE REVISIONS TO THE TOTAL COLIFORM RULE (TCR) 4
CHARGE QUESTIONS TO THE COMMITTEE 5
RESPONSE TO THE CHARGE 6
DWC's Overall Critique of Revisions to the Total Coliform Rule (TCR) 6
Charge Question 1 8
Assumptions to Justify and Explain 9
Explaining and Disentangling Uncertainty and Variability 10
Criterion for Significance and Combining Data Sets 11
Charge Question 2 11
Analysis of Percentage of Systems Implementing Corrective Action 13
Charge Question 3 15
Safe Drinking Water Information System /Federal Version and 6-year Review Databases 16
Corrective Actions 18
Alternatives and Suggestions 20
Charge Question 4 22
TC as an appropriate endpoint 22
Reductions in acute violations as appropriate endpoints for informing benefits 23
Added value of the proposed revisions 23
Other analyses or endpoints that might be considered 24
xn
-------
EXECUTIVE SUMMARY
EPA's Office of Water (OW) requested that the Science Advisory Board (SAB) Drinking
Water Committee (DWC) review its draft supporting analysis in preparation for revising the
Total Coliform Rule (TCR). The Total Coliform Rule (TCR) was established in 1989 and is one
of the primary national regulations governing the microbiological quality of treated drinking
water in the US. The Total Coliform Rule addresses the monitoring of coliform bacteria and
E.coli, and is an important element in the protection against waterborne diseases via
demonstration of appropriate disinfection and groundwater protection. As the Agency embarks
on revising the TCR, it has expended a great amount of time and effort to include stakeholder
inputs, setting up meetings with state regulators, water utilities, local governments,
environmental advocates, public health professionals, consumer advocates, Indian tribes, and
others through a FACA process. As a result of these discussions, an Agreement in Principle
(AIP) was developed that is being used as the foundation for the proposed Revised Total
Coliform Rule (rTCR). This has included a proposed new, more proactive approach for
identifying sanitary defects and incorrect monitoring practices.
The DWC of the EPA Science Advisory Board met in June 2009 to deliberate on four
charge questions raised by OW to address the draft supporting analysis. These questions focused
on: (1) the underlying statistical analysis of the TCR monitoring data used to inform the
prediction of the underlying baseline total coliform andE. coli occurrence and violation rates, (2)
the characterization of the types of corrective actions that systems will implement and the
percentages of systems that will implement certain corrective actions, (3) the methodology and
assumptions used to predict the net impacts in total coliform-positive (TC+) samples, E. coli-
positive (EC+) samples, acute violations, assessments, and corrective actions between the current
TCR (with and without the effects of the Ground Water Rule), the AIP, and the Alternative
Analysis, and (4) the use of reduction inE. coli and TC occurrence and acute violations as
endpoints for informing benefits.
The text of the complete rTCR was not available to the Committee, and the focus was on
the supporting analyses and drafts of key chapters. The members relied on and examined the
text of the AIP and several of the rTCR Chapters for their evaluation and to address the charge
questions. This Executive Summary highlights the outcome of the Committee's deliberations
and the recommendations.
1. The Underlying Statistical Analysis
The Committee commends the Agency for all the work and analyses undertaken. The
Agency performed a partial sensitivity analysis which increases confidence in the assessment.
There was only a limited amount of data from the Safe Drinking Water Information System -
Federal Version (SDWIS/FED) and in the six-year review databases used in the Economic
Analysis that was deemed useful for the rTCR analyses and the Agency did a good deal of work
and put a significant amount of thought and effort into addressing those segments of the database
that would be useful. The work of the Agency toward engagement with stakeholders and
-------
development of the Agreement in Principle which addresses goals for improvement and research
needs was laudable.
The Committee found that the Agency's statistical analysis of the TCR monitoring data
used to inform the prediction of the underlying baseline total coliform and E. coli occurrence and
violation rates is reasonable but would benefit from some refinements and further explanation.
The Agency is commended for the development and application of a statistical approach to a
large, complex dataset, thereby enabling the analysis of a difficult and diverse problem. The
Committee had considered recommending various improvements on the analytical work done,
but noted that the proposed changes were unlikely to show a substantially different result and
would involve a significant effort over a protracted period. Instead, the Committee suggested
that some of the key issues (including seasonality and systems within specific classes and the
difference in their average characteristics regarding positive TC rates) should be further explored
through limited sensitivity analyses. The Committee believes that these additional analyses,
together with the work already performed, will provide a sufficient basis for adequate predictions
and descriptions of total coliform andE. coli occurrence and violation rates.
The Committee, after lengthy discussions with Agency representatives, concluded that
while the analysis was well done, there was, however, insufficient background explanation of the
assumptions and methods used. The Committee was unable to ascertain whether the analysis
was reasonable solely by reading the materials provided. For example, the Data Quality Report
which was referenced through out the document was not available to the Committee. Several
assumptions were not stated or were not clearly explained. The approach was not completely
described, thus the analysis and critical assumptions are in need of further justification and this
needs to be included in the rTCR as it moves forward.
2. The Characterization of the Types of Corrective Actions
The Committee agreed that the Agency's characterization of the types of corrective
actions that systems would implement was reasonable and complete. Corrective actions are
closely aligned with the assessments that must follow positive TC/EC test results during routine
system monitoring. The Committee recommends the Agency explain how it compiled this list of
corrective actions, and where the information came from (i.e., from expert discussions in the
Technical Workgroup of the TCRDSAC, and from information provided by utilities). In
working with states, the Agency found that 10% or less of the utilities were able to find the cause
of the coliform or E.coli occurrence and as a result take corrective actions. Thus, the Agency
estimated that only 10% of the small systems will take corrective action which represents only a
modest benefit from the current rule. This is without the impact of the Ground Water Rule.
However, there are corrective actions that may be undertaken without knowing the exact cause,
such as implementation of a flushing program. Therefore, the Committee recommends that the
sensitivity analysis use a wider range of corrective action implementation including corrective
actions up to and including 50%, especially after a Level 2 assessment. Disinfection seemed to
be the main corrective action considered. If this is correct, then, this fact should be stated, and if
it is not, then the other corrective action options, their costs, and the results of sensitivity
analyses should be clearly described. The Committee also questioned whether the benefits of
flushing would result in reduced risks for as long as several years (as used in the model) unless a
-------
routine flushing program is implemented, if so- this should be clarified. Such assumptions
should also have justification.
3. The Methodology and Assumptions Used to Predict the Net Impacts as part of the Economic
Analysis
The framework and methodology of the Economic Analysis (EA) appear to be consistent
with the USEPA Guidelines for Economic Analysis (2000). The USEPA guideline should be
referenced so that readers understand where the framework for this EA originates and better
understand the reasoning behind the rTCR EA structure. The Committee recommends that an
analysis and discussion in the EA would better inform policy makers by identifying and
summarizing all the variables with known distributions. Further, it should also be clearly noted
that the EA model was developed as a reference baseline and should not be used as a predictive
tool. Ultimately, the Agency will need a far more robust database than the six-year data review
database used in the EA to improve the reliability of the model. The number of assumptions can
be reduced by enhancing the data collection process.
While a reduction in the annual national costs is likely to result from the reduced
monitoring strategy, it is unclear from the analysis whether this may be offset by an increased
risk of waterborne disease outbreaks or endemic waterborne disease. The DWC is skeptical that
the monitoring frequencies for small systems (< 1,000 people) will provide a change to the TCR
that improves public health protection. The DWC is not generally supportive of decreased
monitoring. Overall, the Alternative Analysis (AA) appears to address and protect public health
sooner in time than the AIP proposed implementation.
4. The Use of Reduction in E. coli and TC Occurrence and Acute Violations as Appropriate
Endpoints
The Committee found it difficult to identify specific measurable health-related benefits
attributable to the suggested revisions which will be part of the Total Coliform Rule. From a
public health standpoint, the Committee felt that measuring reductions in total coliforms (TC)
and E. coli (EC) occurrences are not effective sole endpoints for informing benefits because of
the difficulties in linking these indicators to adverse human health outcomes. The Committee
agrees with the Agency's proposal to use TC as part of an overall treatment technique evaluaton
and sees this as a positive step forward. Instead of having an MCLG/MCL for TC, TC-positive
sampling results would trigger an assessment to identify sanitary defects. Implementation of
appropriate corrective action is expected to decrease waterborne disease and occurrence of acute
violations. However, several other indicators could be considered, as TC, in itself, is not an
adequate measure of health risk. While E.coli is viewed as more appropriate measure of risk of
enteric disease, it does not capture the health risks from other pathogens (e.g., Legionelld).
Other measures may provide valuable supplemental information on health risks from distributed
water including structural and hydraulic integrity of distribution systems.
-------
BACKGROUND ON THE SUPPLEMENTS TO FUTURE REVISIONS OF
THE TOTAL COLIFORM RULE (TCR)
The Total Coliform Rule (TCR), established in 1989, is one of the primary national
regulations governing the microbiological quality of treated drinking water in the US. As such,
it is an important element for protecting against waterborne disease. Each standard is mandated
by law to be reviewed and revised, if appropriate, every six years as part of the existing drinking
water regulations requiring revisions to improve/maintain public health protection, including the
TCR. As part of the review and in order to revise the TCR, the Agency and its advisory
committees conducted a substantial amount of work, which included the evaluation of available
data and research on aspects of distribution systems that may create risk to public health. The
Agency also began working with stakeholders to address controls for cross connections and
requirements for backflow prevention in distribution systems, as these are known to be
associated with significant risks, but in some cases, may be out of the water system's direct line
of control. (C. Rodgers-Jenkins, USEPA, Office of Water, SAB teleconference presentation,
5/20/09).
The Agency began the review of the 1989 TCR in 2003. Based on stakeholder comments,
the Agency prepared a series of white papers on the TCR and distribution system issues1 (2006).
In July 2007, the Agency established the Total Coliform Rule/Distribution System Advisory
Committee (TCRDSAC). The TCRDSAC consisted of fifteen members representing the Agency
and state regulators, water utilities, local governments, environmental advocates, public health
professionals, consumer advocates, and Indian tribes. The Agency charged the TCRDSAC with
two major tasks; 1) provide the Agency with recommendations on how best to revise the TCR,
and 2) develop information and research needs to improve the understanding of the risks posed
by distribution system issues. TCRDSAC then formed a technical work group (TWG) to provide
data analysis and information to contribute to the discussion of the Committee. An Agreement in
Principle (AIP) was developed as a result of these discussions, and the AIP is being used as the
foundation for the proposed Revised Total Coliform rule (rTCR). TCRDSAC met thirteen times
between 2007 and 2008. Each representative on the TCRDSAC agreed to support the proposed
rTCR components that reflect the elements of the AIP.
Highlights of TCRDSAC recommendations are: 1) There will no longer be an
MCLG/MCL for total coliforms (TC); 2) TC- and E. coli (EC) positive sampling results will
trigger investigation (assessment), leading water managers to find the problem and fix it; 3)
Monitoring on a quarterly and annual basis may be allowed for some small ground water
systems; 4) Distribution system research and information collection will need to be a priority; 5)
There will be an overall shift in focus from monitoring results that lead to public notification to
monitoring results that trigger an assessment and corrective action. Benefits from these
recommendations will lead to a more proactive approach to public health protection which can
be instituted which should reduce confusion associated with the actions needed for TC violations.
The Agency stated that it was committed to proposing a rule (rTCR) consistent with TCRDSAC
1 See USEPA, Office of Water, Total Colform Rule (TCR) and Distribution System Issue Papers Overview at the
following URL: http://www.epa.gov/7safewater/disinfection/tcr/pdfs/issuepaper_tcr_overview.pdf
-------
recommendations. The initial rule will be proposed in 2011. The final rule will be proposed in
2012, and compliance with the final rule is expected in 2015.
The Agency's recommendations toward the rTCR are to use TC as part of an overall
treatment technique, where there is no MCLG/MCL for TC, but where TC-positive sampling
results would trigger an assessment to identify sanitary defects. A sanitary defect finding would
then trigger corrective action. A treatment technique (TT) violation would be assessed if the
investigation or corrective action was not completed. The Agency would retain an MCLG = 0
for E. coli and the current MCL associated with presence of TC/EC. The Agency would not use
fecal coliforms as a water quality indicator in this Rule, and there would be public notification
for TT or acute MCL violations.
In the AIP, for systems serving < 1,000 people, there would be new criteria for both
increased and reduced monitoring, a transition from the existing monitoring frequency unless the
primacy agency determines otherwise and a decrease in the number of additional routine
monitoring and repeat monitoring samples. For systems serving > 1,000 people, there are no
changes in routine monitoring but a decrease in repeat samples and the elimination of additional
routine monitoring for systems serving < 4,100 people.
As a principle of the assessments, the goal is that the rTCR would pro-actively enhance
public health by identifying sanitary defects and incorrect monitoring practices. In this new
scenario, the public water system (PWS) is responsible for the investigation which should ensure
barriers are in place and effective. Two levels of assessment would be created based on the
severity of the trigger, Level 1 and Level 2. The systems would be required to correct all
sanitary defects found in the assessments. A TT violation would consist of: 1) failure to perform
a Level 1 or 2 assessment when triggered, 2) failure to correct all sanitary defects identified in
the assessment, and 3) failure to correct sanitary defects according to an agreed upon schedule.
Four types of violations that need public notification are: 1) an E. coli MCL violation (Tier 1); 2)
a treatment technique violation (Tier 2); 3) a routine monitoring violation (Tier 3); and 4) a
reporting violation (Tier 4).
It should be noted that the data analysis to support the revisions to the TCR are based on
current and historical water quality data and practices in the United States. The SAB DWC
comments are tailored specifically to those water quality data and practices.
CHARGE QUESTIONS TO THE COMMITTEE
The Agency requests that the SAB Drinking Water Committee review the materials provided and
provide recommendations in the areas specified in the charge questions. As the Committee
considers the specific charge questions that follow, it is asked to consider whether the overall
approach that the Agency has taken to assess the impacts of rTCR is appropriate, given the
availability of the information, and, if it is not, whether there are alternatives that might be
considered.
-------
1. Is the underlying statistical analysis of the TCR monitoring data used to inform the prediction
of the underlying baseline total coliform and E. coli occurrence and violation rates
reasonable? If not, what changes or refinements might be appropriate?
2. Is the characterization of the types of corrective actions that systems will implement and the
percentages of systems that will implement certain corrective actions reasonable? If not,
what else might be considered?
3. Are the methodology and assumptions used to predict the net impacts in total coliform-
positive (TC+) samples, E. co//'-positive (EC+) samples, acute violations, assessments, and
corrective actions between the current TCR (with and without the effects of the Ground
Water Rule), the AIP, and the Alternative Analysis reasonable? If not, what alternatives
might be considered?
4. Are reduction in E. coli and TC occurrence and acute violations appropriate endpoints for
informing benefits? Do they appropriately capture the added value of the proposed
revisions? If not, what other analyses or endpoints might be considered?
The following attachments were included to facilitate the SAB discussions:
• Draft Supporting Analyses
• Baseline Conditions (Chapter 4)
• Occurrence and Predictive Model (Chapter 5)
• Benefits Analysis (Chapter 6)
• Cost Analysis (Chapter 7)
• Draft Technology and Cost Document
• Agreement in Principle
• Background on Current TCR and Rule Revisions Development (presentation)
• Comparison of Current TCR Requirements with the AIP and Alternative Analysis (table)
RESPONSE TO THE CHARGE
DWC's Overall Critique of the Supporting Material, Analyses and Draft Chapters toward the
Revision of the Total Coliform Rule (TCR)
The DWC evaluated the five TCRDSAC recommendations and found them to be
appropriate overall. First, the total coliform (TC) has little relationship to fecal pollution and
public health risk, and thus it is appropriate that there is no longer an MCLG/MCL for total
coliforms (TC). The second recommendation, that TC- and E. coli- positive sampling results
will trigger investigation (assessment), is intended to lead water managers to find the problem
and fix it. This should place an emphasis on investigating and correcting the deficiencies, rather
than on just reporting them. The third recommendation, that monitoring on a quarterly and
annual basis may be allowed for some small ground water systems, is appropriate as long as
these systems have demonstrated the ability to maintain microbiological water quality and
groundwater protection, particularly during the rainy seasons. The fourth recommendation, that
distribution system research and information collection need to be a priority, is also appropriate.
The fifth recommendation, of an overall shift in focus from monitoring information that only
result in public notification to monitoring information that trigger an assessment and corrective
-------
action, is a significant advancement if it is properly executed. It places the emphasis where it
should be, on corrective action. The DWC agrees that benefits from this more proactive
approach to public health protection should reduce confusion associated with what follow-up
actions are necessary and taken for TC violations. However the DWC notes that the EPA
analysis appears to assume a low rate of corrective actions (10% of TC detection events), which
appears to contradict these stated goals.
The Agency's recommendations of using TC as part of an overall treatment technique,
where TC-positive sampling results would trigger an assessment to identify sanitary defects
instead of having an MCLG/MCL for TC, is seen as a positive step forward in principle. The
fact that a sanitary defect finding would trigger corrective action is also very good and
appropriately places more emphasis on remedial action. The DWC agrees that a treatment
technique (TT) violation should be assessed if an investigation or corrective action was not
completed. The DWC also agrees that the Agency should retain an MCLG = 0 for E. coli, as
well as the current MCL-associated with the absence of TC/EC, and not use fecal coliforms as
indicators of water quality in this Rule. There should be public notification for any TT or acute
MCL violations.
For systems serving < 1,000 people, there are new criteria for increased and reduced
monitoring, a transition with existing monitoring frequency unless the primacy agency
determines otherwise and a decrease in number of additional routine monitoring and repeat
monitoring samples. For systems serving > 1,000 people, there are no changes in routine
monitoring but a decrease in repeat samples and elimination of additional routine monitoring for
systems serving < 4,100 people. The DWC is skeptical that the monitoring frequencies for small
systems (< 1,000 people) will provide a change to the TCR that improves public health
protection. The DWC is not generally supportive of decreased monitoring. According to the
Agency's document, the highest percentage of TCR violations occurs in the smaller systems;
hence, it appears as though small systems are more vulnerable and are more likely to experience
a waterborne outbreak. Thus, this vulnerability may be better captured with improved
comprehensive assessments. The problems in the sanitation of these small water systems,
which do not have substantial capital and personnel to monitor their systems comprehensively,
may be best controlled only through a required monitoring scheme.
The DWC agrees that in the scenario proposed for the rTCR, the PWS should be
responsible for assessment, and this should strengthen their capacity to ensure barriers are in
place and effective. DWC also agrees that there should be two levels of assessment, based on
the severity of the trigger, Level 1 and Level 2, and that the systems should correct all sanitary
defects found in the assessments. A TT violation would consist of failure to perform a Level 1
and/or 2 assessment when triggered, failure to correct all sanitary defects identified in the
assessment, and failure to correct sanitary defects according to an agreed upon schedule. The
public notification continues to be important and should include not only a description of the
violation but also the necessary follow-up remedial actions. This would assist in alleviating
public concerns that the PWS is not attending to the problems. The four types of violations that
need public notification are: 1) an E. coli MCL violation (Tier 1) which may be of immediate
public health concern; 2) a treatment technique violation (Tier 2); 3) routine monitoring violation
-------
(Tier 3); and 4) a reporting violation (Tier 4). DWC agrees with these revisions as they would
apply to the future TCR.
It appears from the table provided to the DWC, entitled, "Comparison of Current TCR
Requirements with the AIP and Alternative Analysis," that the PWS can still do its own
assessment (left to the discretion of the primacy agency; but the person doing the assessment is
required to have specific qualifications, unlike a Level 1 Assessment), even if the PWS's lack of
action results in frequent Level 1 triggers that then lead to a Level 2 assessment. This could
result in a public health risk that may be unaddressed in a timely fashion If the issue is raised to a
Level 2 Assessment, the primacy agency should be deeply involved.). The Drinking Water
Committee (DWC) generally agrees that the assessment concept will orient the PWS toward
action and that this will likely improve the microbiological quality of the drinking water
delivered.
Charge Question 1. Is the underlying statistical analysis of the TCR monitoring data used
to inform the prediction of the underlying baseline total coliform and E. coli occurrence
and violation rates reasonable? If not, what changes or refinements might be appropriate?
The statistical analysis of the TCR monitoring data is reasonable, and the Agency is
commended for its systematic and thoughtful analysis of such a large dataset. The
documentation of the statistical analysis in a draft Chapter 5, however, should be improved,
which will address transparency and clarity. The Committee recommends that Chapter 5 be
revised with details which allow the reader to understand better what was done, and all key
assumptions should be clearly stated and justified. The Committee also suggests that further
sensitivity analyses be undertaken.
The Agency did a good job analyzing a difficult and diverse problem. This involved
developing and applying a statistical approach to a large, complex dataset. Positive samples for
different classes of water systems were characterized and used for prediction. Water systems
within a class were allowed to vary in their characteristics. The approach was reasonable and
appears well executed. A significant amount of thought went into structuring the problem and
the analysis.
The Committee had considered recommending various improvements on the analytical
work done, but noted that the proposed changes were unlikely to show a substantially different
result and would involve a significant effort over a protracted period with little added benefit.
Instead, the Committee suggests that some of the key issues be further explored through a limited
sensitivity analyses. The Committee ultimately finds that these additional analyses, together
with the work already performed (with a few refinements discussed below), will provide a basis
for adequate descriptions of total coliform andE. coli occurrence and violation rates.
In making these findings, the Committee notes, however, that the explanation of the
analysis is deficient. The Committee was unable to ascertain whether the analysis was
reasonable solely by reading the materials provided, including the draft Chapter 5, where the
analysis is laid out. Several assumptions are not stated or are not clearly explained. The
-------
approach is not completely described, and the analysis and critical assumptions need further
justification.
The Committee was only able to understand what was done as a result of the discussion
during the meeting. The Committee members had to ask a number of clarifying questions to
understand the analyses performed. EPA staff explained what was done and why. After hearing
these explanations, the Committee was satisfied that the analyses performed were reasonable and,
with a few minor refinements noted below, adequate and, in fact, commendable.
The Committee recommends that Chapter 5 be revised to describe the analysis in
sufficient detail so that the reader can understand the basics of what was done. Assumptions
should be stated clearly and justified. The key assumptions that affect the entire analysis should
be stated early in the chapter and better justified. Implicit assumptions should be stated and
explained. The basic elements of the analysis should be clearly described. As the chapter
currently stands, the reader is left to use his or her judgment as to whether the assumptions are
justified and what was done.
The Committee does recognize that the analysis performed is complex and somewhat
difficult to explain, and that it may take extra effort than may typically be the case to make the
chapter clear and understandable. It may help to have a number of technically facile, non-expert
readers review the revised chapter to see whether they understand the analysis, whether they
could explain it to someone else based on what was written, and whether they find the analysis
including assumptions adequately justified. Specific comments, suggestions and
recommendations are given below.
Suggestion 1. Assumptions to Justify and Explain
The Committee found that the assumptions in the statistical analysis were not clearly
presented and/or justified. Some examples of assumptions and procedures are provided below
that would benefit from better explanation. These are given only as examples, and we leave it to
the Agency staff to work through the document to make sure that all major assumptions are
clearly articulated and adequately explained. Again, the Committee is recommending that the
key assumptions be presented early, and that other minor ones are explained as they occur in the
description of the analysis and results.
• It is not clearly stated that a main goal of the model fitting exercise is to get a
characteristic distribution of probability of positive hits for each water system class, and
that the distribution is representing how systems within the class vary from one another in
that probability.
• There needs to be an explanation of the basic model structure. Exhibit 5.2 appears to be
an attempt at showing how different parameters in the beta distribution are related to one
another for different systems within a water system class. The title is very general, and
the exhibit has no legend. Subscripts are used to indicate the different water systems
within the class; however, the subscripts are not used anywhere else in the chapter. What
-------
would be a difficult concept to convey even with full annotation is totally
incomprehensible. The figure also needs a more explanatory title. If the figure is
retained, it needs to be explained in a figure legend and in the text. The Committee also
recommends the addition of another illustration to show data inputs and model outputs,
with full accompanying annotation and description.
The reason for assuming that seasonality does not impact the analysis should be
discussed and justified, perhaps through a sensitivity analysis.
There is a hierarchy to the analysis that is left unexplained. Each water system class has
its own characteristic beta distribution. Yet, alpha and beta in the figure and text are not
subscripted. In the text, the probability parameters, number of samples (binomial "N"s),
positive samples (binomial "K"s), all have no subscript. In the figure, for example,
subscripts were left out of equations (such as on p 5-7). This made it unclear that they
were relevant to each individual system. While the use of subscripts can be avoided, the
underlying assumptions need to be clear, e.g., be clearer about the fact that independent
analyses (probability distributions) of the routine and repeat samples were performed.
Suggestion 2. Explaining and Disentangling Uncertainty and Variability
The Committee noted that the discussion of uncertainty and variability was unclear. At
times, it appeared that the terms uncertainty and variability were being used interchangeably.
Both uncertainty and variability are an issue in the analysis as illustrated in the following
example. Within a given system classification, water systems can vary considerably in terms of
whether they will test positive and in terms of their violation rates because some of the causes for
TC positives are not under the control of the PWS. For some classes of systems, there may be a
large number of systems within the class and the certainty regarding how much they vary from
one another and the difference in their average characteristics may be high. For other classes,
there may be far fewer observations which may contribute to less certainty. The Committee
found that the document needed to be clearer in how these issues were handled and suggests that
Chapter 5 and the analysis address the issue through further sensitivity analysis and in graphical
presentations.
Chapter 5 could, for example, include a figure plotting positive occurrence rates for some
different prototypical classes of systems. This would provide the reader a visual representation
for how systems can differ within a class. Chapter 5 could also provide plots of the maximum
likelihood estimate for mean positive occurrence for a system class and display the uncertainty in
that estimate. Agency staff has developed some plots like this that could be annotated and
included with proper explanation in the document. Plots that disentangle uncertainty from
variability would be helpful. On the basis of the discussion with Agency staff, the Committee
believes that these plots have been done and that the analysis is fine in this regard, but the
presentation is somewhat lacking.
The Committee recognizes that there are data gaps and that assumptions based on best
professional judgment were made. Therefore, the Committee is not recommending a full Monte
10
-------
Carlo analysis of uncertainty; rather, we suggest the issue be explored in a limited way in a
sensitivity analysis. The analysis can be qualitative or semi-quantitative and should explain how
uncertainty in the beta distribution for classes, where uncertainties are relatively large, may
affect the ultimate result. The Agency stated that uncertainties are greatest for classes with few
observations, and that this is due to the fact that the actual number of systems in the class is small.
However, while this may be a reasonable rationale, it might also be due to one or two outliers
that skew the distribution. The Committee suggests that some simple calculations should
adequately illustrate that the overall result is robust and the rationale has been tested rather than
assumed.
Suggestion 3. Criterion for Significance and Combining Data Sets
The Committee believes that the criterion for testing significance and for deciding to
combine or not combine different data sets should be more explicitly stated and perhaps more
rigorously done, including a statistical power test. In a number of places, the chapter asserts that
systems do not significantly differ in the probability that they will test positive for total coliform.
That finding is used to justify combining data sets for different types of systems. Figure 5.3 is
the only attempt to explain this combining, and both the figure itself and the explanation of for
combining data sets are weak.
In the Agency's analysis, the criterion for deciding whether there is a significant
difference in probabilities is based solely on whether there is a difference in the mean
probabilities of a positive test result for the two systems being compared. However, some
system types can have the same mean, yet the systems may vary within that class. An analysis of
a class of non-disinfected water systems may, for example, include a large number of systems on
a pristine water source in a cold climate, and at the same time also include a large number of
systems with a not so pristine source. The variance for this class can have a wide frequency
distribution, but the mean may be the same as that for another class that is more homogeneous
but relatively pristine. Violations may be more common in the first class than the second. Thus,
in addition to the mean, the variance should also be considered. The beta distributions for these
two classes will differ. Therefore, an alternative criterion for combining could be based on the
hypothesis that both systems have the same beta distribution.
Charge Question 2. Is the characterization of the types of corrective actions that systems
will implement, and the percentages of systems that will implement certain corrective
actions, reasonable? If not, what else might be considered?
To answer charge question 2, the Committee reviewed Chapter 5, Appendix D, and the
supporting document Technology and Cost Document for the Revised Total Coliform Rule. The
Committee agrees that the Agency's characterization of the types of corrective actions that
systems will implement is reasonable and extensive. The types of corrective actions can be
ascertained from current practices. Corrective actions are closely aligned with Level 1 and 2
assessments following positive TC/EC testing during routine system monitoring. Both
assessments are intended to be part of a proactive approach to identify sanitary defects that may
11
-------
put public health at risk, due to the potential exposure to and consumption of contaminated
potable water. Such contamination can arise from source water and ineffective treatment (not a
sanitary defect, but the lack of a barrier or ineffective barrier), any practice or event that results
in intrusion into the distribution system, reservoirs such as biofilms, or stagnant and aged waters
within the system itself (for example as a result of poor operation and maintenance practices).
The proposed rTCR clearly recognizes these issues and uses an assessment process to
initiate analysis of system variables that trigger corrective actions. Assessments are intended to
aid in the following:
• Identifying inadequacies in sampling sites and protocols, sample processing and
presumably appropriateness and QA/QC of the TC/EC analytical methodology.
• Identifying atypical events such as storm flows or construction breaks affecting or
indicating impaired water quality.
• Identifying changes in distribution system maintenance and operation that may affect
water quality.
• Identifying changes in source water quality and/or treatment resulting in the potential for
impaired water quality.
• Identifying inadequacies in the underlying WQ monitoring data itself.
These assessment elements and corrective actions are clearly identified in supporting
documentation for the rTCR, specifically in the Technology and Cost Document for the Revised
Total Coliform Rule. However, they could be more clearly listed in Chapter 5, so the reader does
not have to delve into supporting reports to find the information.
Cost estimates principally for salary and wages (including monitoring, reporting, and operations
and analytical analysis) are also provided. The corrective actions are summarized as follows:
1) Flushing
2) Sampler Training
3) Replacement and Repair
4) Pressure Maintenance
5) Hydraulic Residence Time Maintenance
6) Storage Facility Maintenance
7) Booster Disinfection
8) Cross Connection and Backflow Prevention
9) On-line Monitoring and Control
10) General Security Measures
11) Standard Operation Procedure Training and Implementation
The Committee recommends that the Agency explain how it compiled this list of
corrective actions and from where the information came (i.e., from expert discussions in the
Technical Workgroup of the TCRDSAC, and from information provided by utilities). It is
important for the Agency to acknowledge that this is not a definitive list of all actions that could
be taken and that it is possible that even if all these actions were implemented in a system, that
all total coliform positive observations may not be eliminated. Coliform-positive samples may
be the result of numerous factors in a water system, including cross connections, construction,
12
-------
sampling, etc. It is also noted that a Level 2 assessment mandates that a "certified" operator
must respond to violations. Certification is rather ill defined and is a responsibility of the
individual State. The expertise of the "certified" operator will be very important in determining
the efficacy of the investigations and the percentage of systems taking corrective action after
Level 2 assessments.
The list of corrective actions appears somewhat dismissive of investments in analytical
methods and monitoring, source water analysis, water quality databases and analysis resulting
from the Level 1 and 2 assessments. Real time microbial monitoring, ultimately capable of
targeting specific pathogens of direct public health significance, is on the horizon and real time
residual chlorine monitoring is only a limited surrogate approach.
Suggestion 1. Analysis of Percentage of Systems Implementing Corrective Action
The percentages of systems that will implement corrective actions are un-knowable, so
there is nothing the Agency or the DWC can do except to take a best guess. An estimate that is
based on expert judgment is a reasonable approach to this situation. Two things need to be done
when taking a best guess: (1) a careful justification of what expert information sources were
relied upon and (2) a sensitivity analysis. The Committee commends the Agency for doing the
latter, which greatly increases confidence in the assessment. However, the Agency could do a
much clearer job justifying and explaining how it arrived at these numbers. Specifically, there
should be some text explaining the tables in Appendix D, and there should be reference to
Appendix D and the list of corrective actions in Chapter 5. As the tables stand now, it appears
that large surface water-community water systems (SW CWS) would do nothing in response to a
TC positive; zero percent implementation is assigned for all Level 1 Corrective Actions for those
system categories. Presumably, this is because large systems are assumed to already be doing
assessments and corrective actions. The Agency should clearly document the logic for these
assigned values.
Based primarily on the results of a limited survey of current TCR requirements, the
Agency estimates that corrective actions will on average be implemented only 10% of the time
(section 6.2.7 and associated footnote 1 on page 6-8). It is surprising that such a relatively small
percentage of systems are projected to take some kind of corrective action - at least by flushing
their systems - after a Level 1 or Level 2 event. We expect that flushing would (and should) be a
more common practice. Therefore, we think it would be justifiable for the Agency to use a
greater than 10% likelihood of taking corrective action. In addition, after a Level 2 event, a
serious assessment, done by a certified operator, should find some type of corrective action that
can be taken in most small systems, especially if the menu of corrective actions includes a switch
to disinfection.
This entire discussion appears overly weighted on the influence of the ground water rule
(GWR) and instituting disinfection processes for GW. Does implementation of disinfection in
ground water (GW) systems bias the presumed 10% of corrective action? The sensitivity
analysis uses a range of 5-20% for corrective action implementation. We suggest a broader
range for analysis, including corrective actions up to around 50%, especially after a Level 2
assessment. In addition, if disinfection is the main corrective action, this should be stated. If it is
13
-------
not, then the sensitivity analysis results should be clearly discussed for other corrective action
alternatives and the costs associated with them.
Assessment and corrective actions are well described. However, the discussion of
corrective action appears to emphasize flushing in response to a TC/EC positive monitoring
result. While offering immediate public health protection, flushing alone may not provide
solutions to the underlying problem causing the positive TC/EC monitoring result. Therefore,
the Committee questions whether the benefits of flushing will result in decreased risks for
several years, as the Agency currently estimates in the model. The Committee recognizes that
from a utility perspective, flushing and disinfection is a cost effective solution, if the origins of
the TC/EC cannot not be adequately documented, thus avoiding potential costly repair and
replacement corrective actions that may not ultimately be warranted. It makes sense to switch
some proportion of GW systems to long-term disinfection in the model, and thus change their
attributed rates accordingly after a positive finding.
Overall, the challenge laid down to the Agency by the TCRDSAC was to come up with a
revised TCR that will foster continuous improvement in the industry by encouraging assessments
and corrective actions. It is discouraging to see that the Agency's projections do not fully realize
that vision. The rule should be designed so that a larger percentage of small systems would take
corrective actions that result in long-term benefits. The central estimate of 10%, although
perhaps realistic, represents only a modest benefit from the rule, although it may be better than
the status quo for small systems. It would be useful for the Agency to run the model with a
significantly higher percentage of corrective actions (i.e., 50%), to determine whether the
benefits would increase substantially if the rule can be designed such as to achieve these higher
rates.
Suggestion 2. Alternative monitoring and measures
The AIP recognizes the evolution of analytical capabilities (section 3.2) for TC and EC.
As part of Assessment and Corrective Action, best available methods are appropriate. As an
active area of research, future relevant analytical methodologies should be embraced as they
become available, and the rTCR should not be constrained to current best available technology.
It is important to note that while various methods may report TC, not all the TC's mean the same
thing. There is an ongoing study comparing the methods and the AIP contains a component that
addresses it. The Committee also suggests that the Agency consider moving to alternative
measures other than TC which they believe is not a reliable indicator from a public health
perspective.
Monitoring needs are also identified. As part of research and information needs (4.2.c),
Tier 1 should include molecular microbial methods, stable isotope ratio techniques, and mass
spectroscopic analytical methods for source and contaminant identification which need to be
developed to assist in focused corrective action, rather than corrective action in immediate
response to a proposed rTCR monitoring violation. Tier 2 needs emphasis on human and animal
pathogen detection and real time monitoring.
14
-------
Charge Question 3. Are the methodology and assumptions used to predict the net impacts
in total coliform positive (TC+) samples, E. co/i-positive (EC+) samples, acute violations,
assessments, and corrective actions between the current TCR (with and without the effects
of the Ground Water Rule), the AIP, and the Alternative Analysis reasonable? If not, what
alternatives might be considered?
The framework and methodology of the Economic Analysis (EA) appear to be consistent
with a properly conducted EA based on the USEPA Guidelines for Economic Analysis
(2000). The USEPA guideline should be referenced so readers understand where the framework
for this EA originates and better understand the reasoning behind the proposed revision to Total
Coliform Rule (rTCR) EA structure. Using this framework, the Agency should guide the reader
through the EA clearly pointing out elements that were completed, where assumptions were
made, and where steps were omitted (along with justification for omitting them).
The Council for Regulatory Environmental Modeling notes in their November, 2003,
guidelines, both an uncertainty analysis and a sensitivity analysis should be conducted on any
model developed and used as a basis for a regulatory decision. The term "uncertainty analysis"
itself can lead to confusion, because an overall Uncertainty Analysis (UA) is composed of two
separate components, sensitivity and uncertainty, one of which uses the same term as the overall
analysis itself. Both components are necessary to complete the overall uncertainty analysis. The
sensitivity analysis illustrates the degree of impact that one variable has on the outcome when all
other values are held constant (possibly using a median or average value), such as the use of the
lower and upper bound estimates used in the proposed rTCR EA.
Unlike the sensitivity analysis, the uncertainty analysis examines the inherent variability
in the data and its subsequent impact on the distribution of the output. Some variability is
inherent and a portion may be associated with the measurements used to characterize the
population. The Agency should identify the sources of uncertainty and determine the level of
confidence that should be placed in the final results.
In reviewing the EA guideline, it is apparent that one deficiency in the rTCR uncertainty
analysis originates from the guideline itself and is not a deficiency introduced by the rTCR EA.
While the EA guideline does admit the approach outlined may be different than what other
disciplines consider a complete uncertainty analysis, the USEPA EA Guideline appears to
discuss conducting an Uncertainty Analysis (UA) only. The Agency should identify that the UA
is comprised of a single part, the sensitivity analysis. Since the Agency's EA guideline is not a
regulation or legislated requirement, its structure and content are not so rigid as to preclude
interpretation and improvement, when and if the application warrants it. In this case, the Agency
should point out that the UA in the EA is only comprised of a sensitivity analysis, which does
not include an actual evaluation of uncertainty associated with the output. The analysis and
discussion in the EA would better inform policy makers by identifying and summarizing all the
variables with known distributions, and those without, to aid risk managers with the
interpretation of the results.
The EA also seems to assume that the variables for which there are no data will not have
a significant impact on the overall uncertainty. Until the EA can better characterize the error
15
-------
associated with the output, one should be cautious about inferring too much from the results.
Statements, such as the one beginning on line 23 pg 5-31 (and others similar to it), should be
made with consideration for the error or uncertainty surrounding the outcome, as the observed
difference is probably not significant. The EA discusses some differences as if they are
significant and other differences as if they are not, without a simultaneous discussion as to how
the difference is judged to be significant in one case, but not in the other.
When evaluating the baseline condition, the admission that data are missing from 15
states, including states with large populations (California, Florida, and Pennsylvania), requires
additional discussion on what impact this data gap might have on the final result. In addition, the
number of states submitting TC and EC data in each tier (1 - 4) is not described. Again, there is
no discussion of how the missing data may have biased the baseline estimate.
In addition, the Committee recommends that in regard to the effectiveness of the
corrective actions to reduce future TC- and EC-positive samples, the Agency should repeat the
analysis with values for the percentage of communities that will actually implement corrective
actions from 1% for the less effective corrective actions to 50% for the most effective. In
addition, it is not clear that the reduced monitoring strategy offers significant savings in the
annual national costs for the assays that are not offset by an increased risk of waterborne disease
outbreaks. Overall, the Alternative Analysis (AA) appears to provide public health protection in
a more timely fashion than the AIP .
The Drinking Water Committee (DWC) of the Science Advisory Board (SAB), while
providing comments on the methodology and assumptions used to predict the net impacts of the
proposed rTCR under the current TCR, AIP, and Alternative Analysis (AA), cannot provide an
adequate assessment which could serve as a substitute for an Uncertainty Analysis. In this EA,
one can argue the Uncertainty Analysis is not complete, because of the undefined or ill-defined
distributions associated with many of the input variables. Regardless, it appears that the EA
represents the best possible analysis given the paucity of available data.
Suggestion 1. The Safe Drinking Water Information System (SDWIS)/Federal Version
(FED) and 6-year Review Databases
The Total Coliform Rule is the drinking water rule that applies to all public water
systems regardless of size, source or treatment. Utilities across the United States have invested
billions of dollars in monitoring to meet the requirements of this rule. Unfortunately,
management for these data has been challenging. There was only a limited amount of data from
the Safe Drinking Water Information System - Federal Version (SDWIS/FED) and six-year
review databases used in the EA deemed useful for the rTCR analyses. Given this limited data,
there was still a good deal of work, along with a significant amount of thought and effort,
expended on identifying those segments of the database that would be useful. A large amount of
data has been gathered and screened by the USEPA when the States voluntarily submitted
electronic monitoring data reflecting records from 1998-2005.
In estimating the baseline in Chapter 4, there is one troubling assumption - basing the
analysis solely on one year (2005) out of six years of total coliform (TC) and enteric coliforms
16
-------
(EC) occurrence data. The Agency states, "Using only a single year of data was beneficial in
that it [sic] simplified the analysis..." (page 4-6, line 36). This assumption completely ignores
potentially large inter-annual differences in TC and EC occurrence. Such differences could be
driven, for example, by inter-annual variability in precipitation patterns, which the Agency
acknowledges with the statement (line 39): ".. .changing environmental factors may be
important." Even groundwater shows large year to year differences in the occurrence and
quantity of human enteric viruses, as evidenced in studies of municipal wells2. The baseline
occurrence rate could shift up or down depending on which year of data is analyzed. It may be
possible to evaluate the importance of TC and EC inter-annual variability using data from the
states that submitted complete 6 year datasets. It may also be possible to clarify this issue by
expanding on the statement regarding the consistency of the TC-positive rates in the years
outside the year selected for this analysis.
Further, it should be noted that the EA model was developed to produce a reference
baseline and not to be used as a predictive tool. The EA uses the SDWIS/FED database of
violations as the reference for their model. The data for the model came from the six-year data
review data set, which contained all sample results (with the stated limitations). The model
output was compared to the SDWIS/FED data as the means of validating it. The violations in the
SDWIS/FED database should reflect the sampling results in the six-year data review for those
same periods of time, since the sampling results determine PWS compliance. However, this
should not be considered a "validation" of the model as a predictive tool. A model should only
be considered validated as a predictive tool if it can be shown that it accurately forecasts events
that have not yet taken place. Until additional data are available to confirm the model output, the
EA model should be considered for establishing a reference baseline only. The Agency will
need a database that is far more robust than the 1 year of data from the six-year data review
database used in the EA to validate the model and decrease the long list of assumptions used in
the current EA (see comments in next section).
Given the scheduling of the revised total coliform rule (rTCR), the assumptions used to
fill the data gaps identified in the six-year review database should be considered adequate for this
EA. However, they indicate a deficiency in the data collection system that should be rectified
prior to implementation of the proposed rTCR. On page 4-7 one justification includes the line
"...single year data allows for a single database." This statement implies that the databases were
of such different format and structure that the information in the database from the years prior to
2005 could not be combined into a single database. This points to a need to improve data
management and collection practices. To address at least one of the items in Section 3.17 of the
AIP, the Agency will need a database that is far more robust than the six-year data review
database used in the EA.
To assess the long-term effectiveness of the proposed rTCR (Section 3.17 of the AIP), the
metrics will need to be established and the data collection systems put into place, before the rule
is promulgated. This will allow data collected, under the current TCR, to be used as a reference
2
MA Borchardt, PD Bertz, SK Spencer, and DA. Battigelli, (2003) Incidence of Enteric Viruses in Groundwater from
Household Wells in Wisconsin. Applied and Environmental Microbiology, p. 1172-1180, Vol. 69, No. 2.
17
-------
baseline against which future regulatory actions and their impacts can be compared. Failure to
do so will result in the need to employ similarly weak assumptions in future regulatory
documents and evaluations.
Suggestion 2. Corrective Actions
Filling data gaps by making assumptions is sometimes necessary, especially when the
data needed are not available or have not been collected. Such is the case for the corrective
actions in Exhibit 7.16. These data have never been collected, so the table was populated using
the professional judgment of the TCRDSAC Technical Workgroup (TWO). However, the
Agency should recognize that current data systems are not set up to collect this information as
the rTCR moves forward. Until the data systems are set up to collect this information, future
EAs and regulation assessments will be limited in their usefulness as quantitative assessments
upon which risk managers can and should base their policy decisions.
The biggest red flag (assumption) is the effectiveness of the corrective actions in
reducing future TC- and EC-positive samples. The Agency should repeat the analysis with more
extreme values for the percentage of communities that will actually implement corrective actions,
e.g., 1% for the less effective corrective actions and 50% for the most effective. It is not clear
how the current values in the analysis -10%, 5%, and 20% - were selected. The extreme values
are justified, in our opinion, given the absence of any previous knowledge on how well these
assessments will be performed or the true long-term effectiveness of the corrective actions.
The assumption that systems following a corrective action will be TC -free for a period of
time should be reconsidered. As the EA implies, there are a number of causes for a TC-positive
result. Each of the causes may or may not be linked to a corrective action. In addition, the
causes should be viewed as independent events. Just because a corrective action has been taken
and a sanitary defect corrected, does not mean the public water systems (PWS) will be free of
TC-positive results for a period of time. Granted, some actions may result in long-term solutions,
such as rebuilding a problematic sampling station or modifying sampling habits. However,
given the reports on the aging infrastructure of our water distribution systems in the U. S. and the
fact that distribution systems are continually being exposed to disruptive external conditions
(plumbing activities, excavation, pipe replacement), assuming a system will be coliform-free for
a period of time following a corrective action is a not a justifiable assumption. Assuming that a
corrective action will result in TC and EC compliance for three to five years assumes that once a
sanitary defect is found and corrected, a similar incursion won't happen again for another three
to five years. As there are a variety of events, separated by time and space, that could produce a
TC or EC violation, there is no guarantee that one of these events would not produce a TC or EC
violation immediately following remedial action in an entirely different part of the system.
Each column or corrective action in Exhibit 7.16 represents an action triggered by an
independent event. These actions are not dependent on each other. Thus, the corrective actions
should not be pooled or linked. The causes listed in Exhibit 7.16 appear to be all inclusive as
each line (for a given size system) seems to add to 100%. By summing to 100% for a given size
18
-------
category, it appears as though the corrective actions listed are the only ones that could mitigate a
sanitary defect identified in a Level 1 or 2 triggered assessment.
The list of corrective actions focuses on actions associated with the distribution system to
correct a TC or EC trigger. The list does not include remedial actions taken on other components
of the multiple barriers (source and treatment) that may contribute to a TC or EC trigger. The
discussion of corrective actions should acknowledge changes or modifications made in the
maintenance or operation of the treatment process train, as the assessments may trigger changes
in or the addition of treatment. While it is assumed that the actions taken on the treatment
barriers would be an appropriate response under the Groundwater Disinfection Rule (GWDR) or
Surface Water Treatment Rule (SWTR), these actions may be triggered by monitoring in the
distribution system under the rTCR and identified by the assessment. The linkage between the
rules and the subsequent impact on the EA needs to be fully and clearly explained so the reader
knows which impacts are being included and which have been attributed to previous rules.
The rTCR allows for reduced monitoring, which opens up a policy and technical issue
regarding the adequacy of sampling. 1) The public water systems given more options to qualify
for reduced monitoring are the small systems. Unfortunately, these are the very same systems
most likely to have water sanitary quality problems. This is evident in Exhibit 4.10 where the
public water systems (Community Water Systems + Non-transient, non-community water
systems + Transient Non-Community Water Systems) serving a population size less than 1,000
are responsible for 90% (7,822/8,734) of the TCR violations! The DWC believes that these
small water systems simply do not have the capital investment to provide the necessary
monitoring and corrective actions to reduce this high level of violations. A possible solution is
to encourage smaller water systems to combine with larger systems which may reduce the
number of TCR violations. 2) The TC and EC assays are the only routine, widely-available
distribution system monitoring tool for assessing and ensuring the microbiological quality of
drinking water delivered to the public. The TC and EC monitoring has its limitations, but why
does that translate to requiring less rather than more monitoring? 3) The TC and EC assays are
inexpensive. It is difficult to believe that a reduced monitoring strategy offers significant savings
in the annual national costs for the assays that are not offset by an increased risk (and costs) of
waterborne disease outbreaks.
4) Experience with viral pathogens in groundwater and distribution systems shows that
occurrence is amazingly variable in space and time (Borchardt et al). Thus it may be there is no
such thing as steady-state fecal contamination. Granted, TC and EC are not 100% correlated
with virus occurrence, but the bacterial indicators do provide some information, particularly if
there is a positive test result. The only way to counter the variability and not miss an event is to
perform more frequent sampling, not less. A quarterly or annual sampling plan is highly
unlikely to detect a TC or EC event lasting only one week. Therefore, once a public water
system is placed in this reduced monitoring regimen, it is highly unlikely, unless the sanitary
defect is egregious, to be triggered into more frequent sampling.
Delaying rule implementation to improve the data sets on which this EA is based would
not be protective of public health. From the discussion, it is clear that the Agency needs to
improve the databases containing monitoring and compliance information. Without
improvements, future EAs will be hampered by the issues discussed herein. As the TCRDS FAC
19
-------
recognized, it is important to characterize the impacts of promulgated rules by identifying,
establishing, and collecting the necessary metrics to do so. The long list of assumptions needed
to cover data gaps in the EA reinforces the need for better data collection.
It is understood the analytical and labor costs listed represent national averages for
routine sampling. However, a coliform-positive test will trigger additional sampling,
assessments, or other actions that may not be accommodated within normal work schedules. As
noted in the previous paragraph, other responsibilities may force TCR activities, especially non-
routine activities, into overtime. Hence, some cost factor to account for these activities outside
of the normal labor rates should be considered. After all, one would hope that repeat sampling
and assessment activities are not routine, but that they are conducted by personnel trained and
knowledgeable in these areas. (If assessments are conduced using outside contracts, the labor
costs may actually be higher than the utility's pay scale, considering the level of expertise that
might be required.). In addition, the projected labor costs include fringe benefits, but do not
include overhead (an additional multiplier to cover administrative activities, utilities, office space,
etc.).
Not including the state costs in the ratepayer cost table could skew the underlying cost
distribution. Those states (CA and AL) that operate their regulatory programs on a cost recovery
basis will have their fees passed along directly to the ratepayer (in 2005, the labor cost for an
hour of a state engineer's time was $105). This raises a concern, because when the state costs
and the utility costs (both of which are higher in CA than the national average) are combined, the
cost to the CA ratepayer will be greater. If one considers the distribution of ratepayers, this will
skew the ratepayer costs among the states, pushing the states with higher costs even further out
(widening one end of the distribution).
The Drinking Water Committee believes the model output with respect to the relative
impacts of the AIP and AA is satisfactory, given the limitations of the corrective actions input
data. Since the frequency distributions for TC and EC monitoring results for the AIP and AA are
the same, the only principle difference between the AIP and AA appears to be the initial
monitoring requirements (AIP allows for a transitional period, while AA does not). Since the
AA requires more initial monitoring, not allowing for a transition period, the impacts of the
rTCR under the AA will be observed sooner. However, the overall frequency of TC- or EC-
positives remains essentially the same in both cases. From the standpoint of public health
protection, the AA would be preferred because the endpoint improvements are achieved
relatively sooner than the timeline in the AIP. In addition, getting such a program running at the
small utility and state level would pose initial logistical and administrative challenges. The net
cost for the AA is slightly higher than the AIP, but given the absence of underlying distributions,
the significance of the net difference is not known. In fact, given all the assumptions made, the
difference in the net cost between the AA and AIP is likely not significant, which an uncertainty
analysis would probably verify.
Suggestion 3. Alternatives and Modifications
The following is a summary of some of the changes that could be considered.
• Consider revisions to the Federal data collection system to ensure the data collected are
adequate to establish a baseline from which to measure the future rTCR impact.
20
-------
• Rerun the EA using more extreme values for the proportion of water systems that
implement corrective actions, e. g., 1% and 50%.
• Revise the assumption that corrective actions will lead to extended periods of TC- and
EC-negative results.
• Revise cost estimates to compensate for the overtime needed for repeat sampling and
analyses in response to TC- or EC positives samples or the additional costs for
contracting out an assessment.
• List, identify, and separate those actions and costs associated with TC- or EC- positives
that would be mitigated by other rules, such as the surface water treatment rule or
groundwater disinfection rule. This will more clearly identify those actions and impacts
associated with the different rules and show how the rules are linked.
The Agency should continue its long-term research efforts to develop tests to identify
specific disease-causing organisms, particularly pathogenic strains of E. coli and other
pathogenic bacteria, to enable public water systems to eliminate these disease-causing organisms
from their water systems. This would require a long-term research project(s) which should
commence now with high priority. It is important to take advantage of the latest techniques in
molecular biology, such as Polymerase Chain Reaction (PCR), which have exquisite sensitivity
and specificity, to advance the problem of identification of pathogenic bacteria in the water
systems. One could visualize first moving to pathogenic strains of E. coli by culture methods
and biochemical methods, and later by PCR methodology, which is very rapid and very specific.
The PCR detection of pathogens fits within the rubric of other recommendations made by the
DWC to the EPA concerning the monitoring for, and the rapid detection of, pathogens. This is
yet one more example of how moving along to this form of monitoring would be helpful and
protective of public health. The DWC recognizes that this form of monitoring acts to address
multiple threats to public health, and has the potential to provide timely and specific information.
This should be done carefully and with extensive validation of molecular biology methodology
against classical culture and microbiology methodology to determine whether adoption of these
methodologies would actually advance water sanitation. In addition, the Agency still has to
balance maintaining a broad bacterial screen vs. moving to screens for specific pathogenic
organisms.
Those small and large water systems that fail to comply with corrective action
requirements should perhaps be tasked with more frequent monitoring and reporting
requirements to encourage them to rapidly become compliant with the proposed rTCR for long
periods of time, at which point the requirements could be relaxed.
Overall, the Drinking Water Committee (DWC) advises USEPA to move forward
deliberately to ensure any changes made in generating the rTCR actually result in a significant
reduction in the frequency and severity of Total Coliform observations in our drinking water
systems. DWC recommends substantial caution in developing the rTCR. To date, most U. S.
water systems are maintained well, hence, DWC recommends being very careful in adopting
new changes to the TCR, unless there is a very high probability they will improve water
distribution system sanitation, i.e., The DWC recommends caution in making changes to the
TCR, and incorporating only those changes that have a high probability of making water systems
more sanitary as this is likely a means to lower frequencies of water-borne illnesses.
21
-------
The rTCR should be written to make it clear and easy to comply with for the convenience
of the PWS. In addition, the DWC strongly recommends placing all information related to
aspects of the rTCR on one website with links to supporting materials. The DWC believes that
development of a similar book embodying all the rules contained within the rTCR would be very
valuable to the water quality community.
Charge Question 4. Are reductions in E. coli and TC occurrence and acute violations
appropriate endpoints for informing benefits? Do they appropriately capture the added
value of the proposed revisions? If not, what other analyses or endpoints might be
considered?
The Committee struggled with its response to Charge Question 4 in trying to determine
whether there are measurable health-related benefits that could be attributable to the revised
Total Coliform Rule (rTCR). If the goal of the Rule is to protect public health, the Committee
feels that measuring reductions in total coliforms (TC) and E. coli (EC) occurrences are not
effective sole endpoints for informing benefits because of the difficulties (discussed below) in
linking these indicators to human health outcomes. It is expected that there will be a decrease in
the number of acute violations with the assumption that assessment, followed by corrective
action will decrease the occurrence. The Committee believes there is value in the TT-model.
However, there are a number of other indicators that need to be considered, as TC is not an
adequate measure of health risk. Even though E. coli is viewed as a more appropriate measure of
risk of enteric illness, they do not capture the health risks from Legionella, for example Other
measures, including structural and hydraulic integrity, have been recently considered in a report
by the National Research Council and may provide valuable supplemental information on health
risks of distributed water.
Suggestion 1. TC as an appropriate endpoint
With respect to the use of TC, the Committee notes the following:
• TC can be considered an indicator of treatment efficacy at the point of treatment. To that
end, it may be judicious to encourage utilities to sample for TC at plant effluent.
However, there is no evidence (of which we are aware) to suggest that the detection of
TC in the distribution system indicates risk to human health. The Committee recognizes
that there are many reasons why TC could be detected in the distribution system, e.g.
release from biofilms, intrusion, regrowth, improper sampling, nitrification, and cross-
connections. Therefore, the significance of TC detection is difficult to interpret and
depends in part on context (such as temperature, season and climate).
• For many systems, the number of samples that will be collected under the revised TCR is
inadequate to measure statistically significant reductions in TC and EC occurrence.
• There exist acknowledged problems with false negative and false positive results that
further complicate the interpretation of results.
• These indicators are not used by most other industrialized countries around the world as a
measure of drinking water quality. We believe that they are used only by the US and
Canada (and only under some circumstances).
22
-------
Despite these limitations, the Committee recognizes that:
• The use of TC represents a tool that is already in place, relatively inexpensive, and
familiar to users. Specifically, there is already existing expertise and infrastructure with
which to conduct these analyses in most water systems.
• The presence of TC is indicative of gross contamination, a breach in treatment or
distribution system failure.
• The fact that there exist differences in occurrence between disinfected and non-
disinfected water systems suggests that, at some level, there is a correlation with water
quality improvement.
With respect to E. coli, the Committee felt that it represents a more credible indicator of
public health risk. The utility of E. coli testing, however, is limited by the rarity of its detection.
It is useful as a confirmatory, follow-up test, and the Committee felt it appropriate to retain it as
an MCL. The Committee notes that the Agency's modeling does not predict much effect on E.
coli occurrence, potentially limiting its usefulness as an indicator of improvement.
Suggestion 2. Reductions in acute violations as appropriate endpoints for informing
benefits
It is expected that there will be a decrease in the number of acute violations and
associated Public Notifications following implementation of the rTCR. This expectation is
fueled by the assumption that many Level 1 and Level 2 assessments will be done, that
corrective actions will be taken, and that EC-positive occurrences will decrease. Because
information on the relationship between EC-positive occurrence rates and illness rates is not
available, we have to assume that a reduction in acute violations will lead to reduction in
waterborne illness, which is reasonable. This seems like a reasonable assumption but it is
currently not supported by data. Further, in the years since the implementation of the original
TCR, there has been a persistent level of acute violations among small water systems that has not
changed substantially. Generally, these small water systems lack sufficient resources to bring
their systems into compliance. It is assumed that the number of systems in this acute violation
category will be reduced by the new emphasis in the revised rule on assessments and repairing
defects. This is an appropriate and measurable endpoint.
Suggestion 3. Added value of the proposed revisions
The DWC agrees that there is value in moving from an MCL-model to a TT-model to
better address the nature of the issue. It is more logical to treat these microorganisms as
indicators of the possible presence of pathogens and to require corrective action than to employ a
fixed number (e.g. 5% of all samples) as a "bright line" indicator of a public health problem.
There is value in having a regulation that is more rational and closely aligned with the nature of
the contamination problems. This is discussed briefly at the end of Chapter 6. Increasing the
awareness and familiarity of the operators of small water systems with their specific issues and
23
-------
focusing efforts on correcting deficiencies rather than meeting strict numerical targets is more
likely to decrease overall risk to the communities served.
Suggestion 4. Other analyses or endpoints that might be considered
Given the limitations of TC and EC as endpoints (as measures of health benefit), the
Committee questioned whether there may be value in adding additional endpoints, even though
these additional endpoints may also have limitations. The question raised was whether several
endpoints, in combination, may be more effective than the use of single indicators. It was not
our intention to suggest these additional endpoints as replacements for TC and EC. Several
possible additional indicators were discussed.
First, the Committee considered measures of improvement in public health, such as
surveillance for waterborne disease outbreaks and/or measures of endemic gastrointestinal
illnesses in communities. The Committee recognized that there are issues of sensitivity,
timeliness and cost associated with any surveillance system. However, there may be
circumstances when enhanced surveillance systems (such as monitoring nurse hotline calls,
monitoring sales of anti-diarrheal medication, monitoring hospital emergency department visits
for gastrointestinal illness) could be useful in areas where there is concern about water quality
and/or vulnerable populations. For example, New York City used enhanced surveillance systems
for gastrointestinal disease as part of their strategy to protect public health while avoiding water
filtration (see Watershed Management for Potable Water Supply: Assessing the New York City
Strategy, National Research Council, 2000).
In addition to considering health endpoints, the Committee also considered indicators of
health risk and asked the following questions and looked to the NAS study for guidance, a) What
factors are "known" to be associated with health risk? b) What factors can be measured by small
systems?; by large systems? A recent study by the National Research Council entitled Drinking
Water Distribution Systems: Assessing and Reducing Risks (2006) suggested the following
measures as useful indicators of risk associated with drinking water distribution systems:
1. Measures of hydraulic integrity following the use of the indicators:
• Decreased frequency of pressure drops
• Areas of the distribution system with extreme water age
2. Measures of structural integrity and distribution system management
• Chlorine residual
• Heterotrophic Plate Count (HPC) bacteria
• Assessment of biofilms using snaking cameras
• Number of utilities with routine prophylactic flushing program
• Number of utilities with active leak detection program
• Number of utilities with active cross-connection detection program
• Number of utilities with active backflow prevention program (e.g. Increased number of
utilities that adopt better management practices)
24
-------
Some combination of these measures may serve as helpful endpoints when trying to assess the
impact of the revised TCR.
Finally, the Committee also discussed concerns about health risks that are not adequately
captured by these endpoints, in particular, risks due to Legionella that is associated with a
significant number of waterborne disease outbreaks each year (Surveillance for Waterborne
Disease and Outbreaks Associated with Drinking Water and Water not Intended for Drinking -
United States, 2005-2006, MMWR, September 12, 2008 / Vol. 57 / No.SS-9). Thus, some of these
pathogen-specific measures may need to be monitored if they are not captured by the fecal
endpoints.
25
------- |