v>EPA
United States
Environmental Protection
Agency
Microbial Expert Input and Review for
the Third Contaminant Candidate List
Office of Water (4607M) EPA 815-R-08-0010 February 2008 - Draft www.epa.gov/safewater
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
Table of Contents
1.0 INTRODUCTION 1
2.0 BACKGROUND 1
3.0 PROJECT SUMMARY 2
4.0 MICROBIAL WORKSHOP 3
4.1 Day 1 4
4.2 Day 2 12
Table of Exhibits
Exhibit 1: Meeting Agenda for Tuesday, March 20 - Wednesday, March 21, 2007 3
Exhibit 2: PCCL Rescoring Data 14
Exhibit 3: Rank Ordered Microbe List Based On Expert Review 17
Exhibit 4: Rank Reflecting General Consensus of Experts that Voted 21
Page i of ii
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
List of Acronyms and Abbreviations
CCL Contaminant Candidate List
CCL 1 EPA's first contaminant candidate list
CCL 2 EPA's second Contaminant Candidate List
CCL 3 EPA's third Contaminant Candidate List
EPA United States Environmental Protection Agency
ID50 Infectious Dose 50: The infectious dose of a micoorganism required to
produce infection in 50 percent of the exposed population
NAS National Academies of Sciences
NDWAC National Drinking Water Advisory Council
NRC National Research Council
OGWDW Office of Ground Water and Drinking Water
PCCL Preliminary CCL
PWS Public water system
SDWA Safe Drinking Water Act
US United States of America
WBDO Waterborne Disease Outbreaks
Page ii of ii
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
1.0 Introduction
As part of the process of establishing a microbial Contaminant Candidate List (CCL), the United
States Environmental Protection Agency (EPA) sought expert input on its approach to
identifying and prioritizing contaminants. On March 20 and 21, 2007, an expert panel convened
in Washington, D.C. at EPA Headquarters to provide input and review of the draft third CCL
(CCL 3) microbial prioritization process. A panel of 6 experts was selected based on their
experience in the fields of public health, toxicology, and epidemiology; and familiarity with the
Safe Drinking Water Act regulations and the CCL regulatory process. This document provides a
summary of the proceedings of the two-day workshop, organized and facilitated by Horsley &
Witten, Inc. The workshop agenda is included in section 4.0 of this report.
2.0 Background
The Safe Drinking Water Act (SDWA) includes a process that the EPA must follow to identify
new contaminants that may require regulation. According to the SDWA, EPA must periodically
release a CCL of unregulated contaminants that are known to or anticipated to occur in drinking
water at levels that may pose a risk to public health; and therefore, may require regulation. EPA
typically conducts an extensive research and data collection effort, and solicits comments from
experts and the general public (via the Federal Register), on unregulated contaminants to develop
a CCL. These contaminants are then further evaluated by EPA to determine whether they should
be regulated. When making this determination, the SDWA specifies three criteria to determine
whether a contaminant may require regulation:
• the contaminant may have an adverse effect on the health of persons;
• the contaminant is known to occur or there is a substantial likelihood that the contaminant
will occur in public water systems with a frequency and at levels of public health
concern; and
• in the sole judgment of the Administrator, regulation of such contaminant presents a
meaningful opportunity for health risk reduction for persons served by public water
systems.
The first CCL (CCL 1), established in March of 1998, contained 60 contaminants (50 chemical
and 10 microbial) that were chosen based on expert opinion. EPA then made their regulatory
determinations on the CCL 1 and ultimately decided not to regulate 9 contaminants, based on
their evaluation of "significant risk reduction" as described in the SDWA. The second CCL
(CCL 2), established in February 2005, carried forward the remaining 51 contaminants from
CCL 1 (9 microbiological contaminants and 42 chemical contaminants). During this time, EPA
provided an update on the Agency's work to improve future CCL review processes based, in
part, on recommendations from the National Research Council (NRC) and the National Drinking
Water Advisory Council (NOWAC).
NOW AC and the National Academies of Science (NAS) proposed a broader, more
comprehensive evaluation process than previously utilized by EPA to assist the Agency in
identifying contaminants for the CCL. They recommended that EPA develop and use a process
Page 1 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
for creating future CCLs. As a result, a broad universe of potential drinking water contaminants
were established, assessed, and reduced to a preliminary CCL (PCCL), using simple screening
criteria. The screening criteria indicate public health risk and the likelihood of occurrence in
drinking water. All of the contaminants on the PCCL would then be assessed in more detail
using a classification approach and tools, along with expert judgment, to evaluate the likelihood
that specific contaminants could occur in drinking water at levels and at frequencies that pose a
public health risk. The outcome of the detailed classification approach results in the draft CCL.
EPA began developing CCL 3 in 2006 using the new procedures described above. During this
process, they identified 284 data sources for consideration in the CCL 3 process, including some
contaminants from the CCL 2. Each universe (microbial and chemical) was narrowed down to a
PCCL using simple screening criteria, based on a contaminant's potential to occur in water
systems and to cause adverse human health effects.
The universe for the microbial CCL 3 includes a survey of human pathogens published by Taylor
et. cil., and pathogens nominated through the public nominations process. Screening criteria
were used to indicate the potential for waterborne transmission and identify microorganisms to
move to the PCCL.
All of the contaminants on the microbial PCCL are assessed using attributes (e.g., waterborne
disease outbreaks, occurrence, health effects) to characterize the potential for the microbial
pathogen to occur in PWS, cause waterborne disease outbreaks and adverse health effects. The
outcome of the detailed approach resulted in the draft microbial CCL 3 list.
3.0 Project Summary
The goal of this project was to obtain expert input on the approach EPA is using to establish the
microbial CCL 3. Specifically, the focus of this review was to provide comment on the draft list
of microorganisms, the screening process, and scoring protocols used to establish the lists.
Horsley & Witten, Inc. was contracted by EPA to coordinate the expert review of the CCL 3 for
microbial contaminants. A pool of potential experts recommended by their peers from national
drinking water organizations such as the American Public Health Association, Association of
State Drinking Water Administrators, National Science Foundation, and universities with strong
public health and medical programs were evaluated.
Horsley & Witten selected 6 experts to participate in the microbial review. Experts were
selected based on their experience in the fields of public health, toxicology, and epidemiology;
their familiarity with the SDWA regulations and the CCL regulatory process; as well as their
level of interest. Horsley &Witten organized and facilitated a two-day microbial workshop that
was held in Washington, D.C. at EPA Headquarters (March 20-21), where the experts served on
a panel to discuss their findings regarding the draft CCL 3 microbial process. The workshop
agenda is included in this report under the workshop section.
Experts received an organized packet of information prior to the workshops, which included the
workshop agenda and all CCL 3 associated materials including documentation of the compilation
of the CCL Universe, screening process, scoring process, and contaminant dossiers. Experts
Page 2 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
answered all questions posed by EPA and engaged in productive discussions regarding
contaminants and whether the draft CCL lists developed by EPA were acceptable, based on the
screening and scoring process. A detailed summary of the workshop is included in this
document.
4.0 Microbial Workshop
Exhibit 1: Meeting Agenda for Tuesday, March 20 - Wednesday, March
21,2007
DAYl
Time
8:30 - 8:50 AM
8:50-9:10 AM
9:10-9:40 AM
9:40 -10:00 AM
10:00-10:15
AM
Topic
Introductions
Welcome
Meeting Objectives/Ground Rules, Logistics
Overview of the CCL 3 Process - Historical Perspective
Presentation of Charge
Questions/ Answers
Speaker
Facilitator: Richard Delaney,
Horsley Witten Group, Inc.
Pamela Barr, Standards and Risk
Management Division Director
EPA Office of Ground Water and
Drinking Water (OGWDW)
Richard Delaney
Tom Carpenter,
EPA OGWDW
Crystal Rodgers,
EPA OGWDW
Experts/EPA
Facilitated by Rich Delaney
10:15 -10:30 AM: BREAK
10:30 AM-
12:00
Screening Question 1 : Are the narratives describing the
screening criteria clear, adequate and transparent?
Experts
12:00 - 1:00 PM: LUNCH
1:00- 2:30 PM
Screening Question 2: Do the fact sheets adequately
represent the available and updated information for each
contaminant?
Experts
2:30- 2:45 PM: BREAK
2:45- 3:30 PM
3:30- 4:30 PM
4:30- 5:00 PM
Screening Question 3 : Are the attributes and the
respective scoring protocols reasonable and transparent?
Screening Question 3.1: Do the scoring protocols
adequately address the microbes and available data
necessary to characterize each attribute?
Wrap-up
Experts
Experts
Richard Delaney
DAY 2
Time
8:30 -9:00 AM
9:00-10:30 AM
Topic
Recap of Day 1
Screening Question 3.2: Is the approach for using the
highest score between the waterborne disease and the
occurrence attributes reasonable?
Speaker
Crystal Rodgers,
EPA OGWDW
Experts
Page 3 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
10:30-11:30
AM
Screening Question 3.3: Is the approach
for deriving the overall health effects
from the scores of the general population
and sensitive subpopulation reasonable?
Experts
11:30- 12:30 PM: LUNCH
12:30- 1:45 PM
Screening Question 3. 4: Is the idea of
summing attribute scores to get the total
score for each microorganism reasonable?
Experts
1:45- 2:00 PM: BREAK
2:00- 2:45 PM
2:45 -3: 15PM
3:15 -4:OOPM
4:00- 4:30 PM
General Question: Does the Draft CCL 3
microbe list represent those pathogens
that have the highest potential to occur in
public water systems and cause adverse
human health effects? Are there
pathogens on the Draft CCL 3 list that
should not be listed and, conversely, are
there pathogens that should be listed?
Screening Question 4: Is the ranked
CCL produced by the scoring protocol
process reasonable?
Expert Panel recap of Charge Questions
and recommendations
Wrap-up
Experts
Experts
Experts
Richard Delaney
4.1 Day1
Introduction - Rich Delaney, Executive Vice President, Horsley Witten Group, Inc.
(Facilitator):
Expert Panel:
Mark Borchardt, Ph.D., Research Scientist, Marshfield Clinic Research Foundation
Patrick Murray, Chief Microbiologist, National Institutes of Health
Kellogg Schwab, PhD, Associate Professor, John Hopkins University, Bloomberg School of
Public Health
David Welch, PhD, Clinical Microbiologist, Medical Microbiology Consulting
Jon Mark Hirshon, MD, University of Maryland School of Medicine
Rebecca Hoffman, Wisconsin State Laboratory of Hygiene, University of Wisconsin, Madison
Welcome - Pamela Barr, Director, EPA Standards and Risk Management Division
Ms. Barr discussed the background of the microbial CCL 3. The contaminants on the list are
known or anticipated to occur in drinking water and are most likely to cause public health
concerns. EPA agreed with the NAS recommendation on how the review process should be
transparent. The Agency also agreed with the NDWAC's recommendations on how to
streamline the review process through the completion of the following steps; 1) maintain a
universe of microbes, 2) screen this list to only include waterborne contaminants and, 3) further
Page 4 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
evaluate these waterborne pathogens for their potential to cause adverse health effects. Ms. Barr
explained that the reviewers' goal for this workshop is to review draft list, assumptions, and list
of recommendations to see if they agree with EPA's conclusions regarding the contaminants.
Ms. Barr reminded the group that the list and the other workbook materials they received are
internal agency deliberative documents and asked the reviewers not to quote, cite or distribute
the information. She explained that EPA is looking for individual expertise, not information
from viewpoint of reviewer's organizations. EPA had received technical information so far, and
the next step is to prepare recommendations and information from the workshop for internal
Agency review. Ms. Barr stated that the draft list will be published in the Federal Register in
February 2008 and the final list will be completed in August of 2009.
Rich Delaney discussed how Horsley &Witten staff will play a listening role and will record
points of agreement and "parking lot" items in order to help reviewers move forward. Meeting
notes will be issued to reviewers for their comments approximately 2 weeks from the date of the
meeting.
History and Process Information - Tom Carpenter, EPA Standards and Risk Management
Division
The history of the process was reviewed describing the steps from the CCL 1 to the CCL 2, and
then to the CCL 3. NRC and NDWAC identified challenges within the CCL process and these
were addressed in this new process for the CCL 3. This process begins with a previously
published survey of microbial pathogens identified in the Taylor et al. report (2001) used as the
CCL universe. Subsequently, this universe of human pathogens is screened to derive a PCCL.
The CCL is then selected from the PCCL microbes using attributes to characterize health effects
and occurrence. The regulatory decisions that EPA will make regarding the contaminants to
include on the CCL 3 will relate to the likelihood for occurrence of the contaminants in a Public
Water System (PWS) at a particular frequency, and what adverse heath concerns they cause.
Balancing Occurrence and Health Effects:
1. The slide presentation shows the progression of the CCL 1 to CCL 3.
2. Literature review 1400 (species).
3. Nominations (what did we miss?).
4. Plausibility that waterborne disease can occur (overall screening criteria - see screening
document in workbook regarding the 12 screening criteria).
5. Results of screening (see table in screening document) - these are the ones EPA wants to
look at now, these are what they feel are the most virulent contaminants.
Page 5 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
Attributes to Scoring the PCCL:
1. Treatment process consideration - Long Term Surface Water treatment rule - drinking
water treatment NOT considered because there are too many variables (e.g.,
chlorination, filtration, political decisions, etc.), also because the goal is public health
protection - basic assumption.
2. Occurrence - direct detection of microbes using cultural, immunochemical, or molecular
detection of pathogens in water.
3. Health effects - including sensitive populations: chronic disease population, pregnant
mothers, elderly, children), and gastrointestinal disease.
4. Waterborne Disease Outbreaks - documentation of occurrence of pathogens in drinking
water by public health officials through adverse health effects in a population and are
direct evidence of exposure.
The goal was to put contaminants on a level playing field. EPA tried to balance Waterborne
Disease Outbreaks (WBDO) and occurrence. The highest score (between WBDO and
occurrence) is selected in order to elevate pathogens that have been detected in US drinking
water or source water above those that have caused WBDOs in other countries but not in the US
or pathogens that have not caused WBDOs in any country but have been epidemiologically
associated with water-related disease.
Overview of Charge and Questions - Crystal Rodgers, EPA Standards and Risk
Management Division
EPA's expectation is that reviewers will provide substantive comments on the draft CCL 3 and
on the scoring and screening processes. EPA hopes that reviewers will work towards answering
the following question: does the CCL 3 represent pathogens that have the most potential to occur
in PWS and cause adverse health effects.
Expert Discussion: Regarding PWS & occurrence - how do distribution systems fit into
occurrence? For example, typically, distribution systems run for miles after treatment, which
adds another layer of complexity.
EPA Response: SDWA regulations define PWS as supplying 25 or more people or having 15 or
more service systems. They do not include private wells or smaller systems. The components of
the PWS do not stop at end of treatment; therefore, the distribution system is included. PWS
also includes premise plumbing, meaning the service line to buildings (water is considered public
until this point). However, the definition of premise plumbing depends on the water system's
jurisdiction.
An important overall question for reviewers to consider is: Are there contaminants that
reviewers think should NOT be on list, taking scoring and screening protocols into consideration,
would you have come up with same list? Please see the 12 screening criteria (section 3 of
screening document). Please note that criteria 1-8 have had numerous discussions, so they are
not likely to be changed.
Page 6 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
There should be 30 fact sheets, each one for the contaminants that made it to PCCL. This is
forum to discuss issues regarding the literature.
This process is partially to determine if reviewers would come up with similar scoring. EPA
called upon 10 microbiologists with different specialties to score the PCCL and choose/rank the
draft CCL 3. EPA has not looked at statistical variation of scores.
The following identifies issues and clarification that experts noted during the presentations:
• There are no substantive changes from the CCL 1 to CCL 2; however, there have been
substantial changes from the CCL 2 to the CCL 3. It seems that EPA has shifted gears
(i.e., gone back on some of their determinations) because treatment is no longer
accounted for. Perhaps some organisms are included that should not be.
• Perhaps there is an issue regarding the measure of reproducibility for scoring (i.e., could
similar scores be reproduced with other experts or is there too much flexibility in scoring
and professional judgment?). The group recommends that another group scoring should
occur to determine scoring variability.
• The reason why 7.0 was selected by EPA as the cut-off point for pathogens that made it
to the draft CCL 3 should be explored. For example, will rescoring make a difference if
more or less "weight" is placed on occurrence? What does the group feel is the rational
break point of scoring? Experts agreed that they should not ignore the lower scored
pathogens even though occurrence is not prevalent.
Discussion of Questions (In the order in which they were approached):
Question #1: Are the narratives describing the screening criteria clear, adequate and
transparent?
The experts began their discussion by going through each screening criterion to determine
whether the criteria were clear and transparent. A number of key questions and points of
clarification were raised by the experts, as follows:
1. Screening Criterion 1: Were all anaerobes excluded from the PCCL through this screening
criterion?
• Expert recommendation: the parenthetical is misleading in document, and therefore, the
reasons why these are excluded should be made clearer. For future iterations of the CCL
there should be allowances for exceptions (e.g., if documentation of a WBDO becomes
available).
2. Screening Criterion 2: Wording of criterion 2 is confusing. The meanings of "obligate
intracellular or fastidious pathogens" should be clarified.
3. Screening Criteria 3 & 4: The following preamble should be added to both criteria: "These
are examples of those excluded because they did not cause a waterborne disease outbreak",
for clarity sake.
Page 7 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
4. Screening Criterion 5: No discussion, no changes recommended.
5. Screening Criterion 6: Under this criterion, Legionella would be excluded. A preamble
clarifying this criteria is also needed, similar to the one noted above.
6. Screening Criterion 7: It is a little unclear whether EPA is including source water under this
screening criterion. A "fuzzy line" occurs if a contaminant occurs in recreational source
water - according to EPA; it appears that they would not consider this, unless compelling
data suggests inhalation on ingestion. This point should be made clearer in the criterion
description.
7. Screening Criterion 8: No discussion, no changes recommended.
8. Screening Criterion 9): The group discussed this criterion at length, specifically
contaminants that could have met this criterion, but were excluded from the PCCL. For
example, Acanthamoeba was on the CCL 1 but was screened from the CCL 3. In addition, it
is plausible that water is a source for Pseudomonas because their cells grow well in water and
may amplify in distribution systems; however it is not included in the PCCL. Furthermore,
Legionella is included on the PCCL - a contaminant that almost mirrors Pseudomonas in
transfer.
• Expert Recommendation: Additional research on Pseudomonas is needed to agree
with EPA's conclusion. It seems that there is a systematic bias based on occurrence
scoring - weighting positive results, but excluding negative results. Negative results
should also be evaluated for PCCL. The group understands excluding a pathogen if it
was looked for, but not if it was excluded because it was never looked for. This
criterion needs to be reviewed more carefully to account for this issue. In addition, a
preamble should be added to this criterion discussing this uncertainty.
9. Screening Criterion 10: A preamble, similar to what was requested for Screening Criterion 9,
is required.
10. Screening Criterion 11: In general, EPA should better identify which pathogens were
included in PCCL, and why, in the CCL documentation.
11. Screening Criteria 12: Taxonomy clarification is needed.
Question #1 - Experts Summary Answer: No. The exclusion of some pathogens from the
PCCL should be made much clearer. Documentation for drinking water disease should have
a higher weight. In addition, two outbreaks involving 1-2 individuals received the same
weight as 100 outbreaks involving thousands of individuals. Experts concluded that some
assessment of impact of exposure on the general population needs to be factored into the risk
analysis. EPA should clarify who conducted the screening process (e.g., a flowchart could
better illustrate this). Preambles are needed for criteria, as described above. Further
information regarding whether there is a waterborne disease outbreak is needed for a number
of contaminants on the PCCL.
Page 8 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
Question #3: Are the attributes and the respective scoring protocols reasonable and
transparent?
The discussion began with further clarification of the question by EPA. It was stated that EPA
would like to know if there are contaminants that they should not invest research resources in,
but could still be included on the list (i.e., could there be a different weighting system)?
First, an overview of EPA's scoring formula for WBDO or Occurrence + Health Effects, as
follows:
Score Equivalency
WBDO
5
4
3
2
1
Occurrence
None
None
O
2
1
The protocol took the highest score of each category. Occurrence was determined by EPA to
mean any occurrence of the pathogen in the US WBDO was interpreted as multiple outbreaks in
aUSPWS.
Experts began their discussion regarding perceived variances in occurrence and WBDO
classification. For example, there is great variability in US climate and conditions of each PWS.
In addition, it seems that the number of times (frequency - e.g., 2 deaths vs. 20,000 people sick)
a pathogen occurred in a PWS was not taken into account. Experts should apply their technical
expertise in reviewing each fact sheet for PCCL contaminants. Human prevalence disease data
is not included in scoring, just because we don't know if they are waterborne.
An overview of the health effects scoring process was provided by EPA, as follows:
• Surrogates for potency and severity were used ("ID50" model).
• An illness to mortality evaluation was completed by EPA to evaluate sensitive
subpopulations (1-7, respectively).
• Within each subpopulation the normal incidence was selected, not extreme cases. (However,
prevalence data was not always available because not all cases have been reported to the
medical field or by the medical field.)
Expert Discussion: There does not appear to be any weighting of the general population disease
burden, so perhaps the scoring is biased? Rare vs. frequent illness events should be included for
the general population. There also seems to be a glaring omission of some pathogens on the
draft CCL 3; therefore, it is unclear whether there was enough information on some pathogens to
score them properly. There are available studies, but most lump water with food based ingestion.
Page 9 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
Question #3 - Experts Summary Answer: Experts felt that scoring is too subjective and that
the overall score only allows for ranking as long as part of ranking is subjective. They
recommend that the variability and range of scoring process be clarified. Experts have
concerns that the scoring system - mainly health effects - may not be a reasonable
representation of the population burden, based on available knowledge. The lack of
population burden does indeed show up in some studies about rare organisms that cause rare
disease. The general population disease burden (frequency) should be taken into account.
Sub-Question #3.2: Is the approach for using the highest score between the waterborne
disease and the occurrence attributes reasonable?
Experts began their discussion by reviewing the assignment of scores to contaminants on the
PCCL. Experts felt that the water source path definition is too broad, specifically statements
found on Attachment 1 of the PCCL Scoring Protocols document. In addition, "fresh water used
for recreation", should be clarified because another interpretation can lead one to include other
pathogens on PCCL.
Sub-Question #3.2 - Experts Answer: It appears that the approach for using the highest score
between WBDO and occurrence attributes is an acceptable methodology; however, criteria for
associated definitions must be clarified for these attributes to be properly utilized. In addition,
experts conclude that WBDO and occurrence scoring appears like "double counting" and;
therefore, should not be scored separately, rather using the qualifier "either/or".
Sub-Question #3.3: Is the approach for deriving the overall health effects from the scores
of the general population and sensitive subpopulation reasonable?
The panel began their discussion regarding the general assignment of health effects scores. All
experts agreed that the scoring is confusing. They did not fully understand how the rating was
determined. Experts noted that certain viruses are not being tested, and are therefore excluded
from the PCCL and draft CCL 3. Many pathogens on the PCCL list are those w/ available
information.
The group discussed whether some organisms could be lumped into a representative category.
(e.g., how total coliform represents a group of human-based bacteriologic organisms). Experts
discussed how the health effects ratings seem inconsistent (e.g., Human enterovirus received a
rating of 6 as well as Escherichia coli, which is regulated and controlled by treatment,).
Experts concluded that these types of variations are not visible in the document, but they should
be. There needs to be a definition spelled out regarding rating variables, e.g., most severe versus
most common.
Experts continued their discussion regarding the definition of "meaningful opportunity". There
seems to be quite a bit of flexibility on EPA's part in determining what this means. Does this
lend to consistent scoring? The discussion then turned back toward the variability in the human
effects scoring. Experts concluded that documented outbreak is so variable; therefore, it could
be better to weigh score based on occurrence (i.e., yes/no), severity, and waterborne disease.
The experts concluded that there is a "systematic bias" to the WBDO scoring protocol. Experts
Page 10 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
recommended that EPA recognize and discuss the "systematic bias" to allow the public to better
evaluate the protocol.
Sub-Question #3.3 - Experts Summary Answer: There is a general lack of transparency and
consistency in the health effects scoring. In particular, there is a lack of transparency and
consistency for the general population & subpopulations. Experts identified the following
specific issues:
1) The methodology applied to the general population and subpopulations appears to be
inconsistent, particularly regarding burden of disease.
2) There needs to be a more definitive balance between health effects & WBDO.
Solutions can be offered with further review.
Sub-Question #3.4: Is the idea of summing attribute scores to get total score for each
microorganism reasonable?
Sub-Question #3.4 - Experts Summary Answer: Experts agreed that the attribute scoring
resulted in a ranked list that was reasonable, but not necessarily appropriate.
Day 1 Summary Discussion:
The reviewers concluded Day 1 with a summary of their recommendations regarding scoring,
ranking, and issues that need to be further discussed, as follows:
• Prioritization/ranking appears to be a good approach, but it needs to be more clearly
defined.
• A systematic bias towards occurrence and WBDO scoring is prevalent. Experts should
review this in more detail to determine a best answer to the general question on Day 2.
• The population burden or disease frequency is missing from the scoring system.
• Qualification of the regulation of potential pathogens versus unknown or newly identified
pathogens should be addressed. Experts share an overall concern that EPA is not
proactive about new pathogens because the information is not as available. Including
only pathogens that have been fully researched will dilute the effort.
• Methods and limitations section should be included in PCCL Scoring and Screening
documentation.
Expert Summary of Questions 1, 3.1-4: Experts expressed that the CCL process included an
impressive research effort on EPA 'spart. However, there are some gaps in the data that
experts would like further clarification on from EPA. In particular, experts would like a
response from EPA regarding the likelihood or frequency of pathogen occurrence to make
would make scoring more accurate.
Page 11 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
Rich Delaney: Tomorrow will begin with scoring clarification and then dive into fact sheets.
Experts have requested the following supplemental information from EPA:
1. Goal of CCL (summary).
2. CCL 1, 2 & 3 status information.
3. Overview of nominated CCL 3 contaminants.
4. Logic for scoring process (i.e. issues related to scoring inconsistency).
5. Text and journals regarding health effects data for reference.
4.2 Day 2
Day 2 of the workshop began with a follow-up on the expert's questions above, and an overview
of supplemental information provided to reviewers today by Eric Bergman. EPA began with a
detailed clarification of the CCL 3 goals, as requested:
• Purpose of the CCL: EPA must identify non-regulated contaminants that are known or
anticipated to occur in public drinking water, cause adverse health effects, and may
require regulation.
• Research priorities are secondary.
• It is acknowledged that there is a line between how much information is available about
contaminants, and being able to list them. Contaminants should have data sufficient to
make a regulatory determination.
EPA continued to discuss the issue reviewers raised regarding the quantification of the number
of pathogens to be included on the CCL 3. EPA clarified that they haven't quantified the number
of CCL 3 contaminants that can be researched/regulated; therefore, reviewers should refocus on
whether EPA has the right contaminants on the list versus the broad number of contaminants
included. However, the disadvantage of a large CCL 3 list is that it would be difficult for EPA to
effectively research and make regulatory recommendations on each contaminant due to resource
constraints.
Regulatory determination process includes the following steps:
1. Demonstrate adverse health effects;
2. Identify likelihood for the contaminant to occur in PWS at a level and frequency to cause
health effects and;
3. Determine whether there is a meaningful opportunity for changing this (e.g. treatment or
changing exposure).
The challenge is having scientific certainly regarding risk assessment to compare health effects
and occurrence, as well as WBDO data. This will drive the regulatory determination. EPA
acknowledges reviewers concerns regarding pathogens that are on the PCCL that already have
Page 12 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
regulation associated with them. However, EPA will evaluate whether existing treatment/limits
should be changed or if new treatment/regulations should be developed for these contaminants.
EPA provided the following information, per expert's request:
1. Handout describing the status of the CCL 1, 2 & 3.
2. Handout describing the nomination process.
3. Regarding scoring variability; scores were brought forward from the internal EPA
review, as shown on the 3 example scoring worksheets that EPA provided the group.
Individual experts scored each on their own and then a group came together to discuss
scores. Citations of data were required, especially for any alteration in original scores
after group discussions. All reviewers received the same data sheets.
Discussion on New Materials Presented by EPA During Day 2:
Expert Question: The generaMycobacterium are grouped. What is EPA's rationale for scoring
this way?
EPA: Mycobacterium avium was selected as a representative organism because it has significant
occurrence and transmission; therefore, treatment of this pathogen would control all species -
rapid or slow growing.
Expert Response: This grouping is acceptable if rapid growers do not cause more significant
disease than the representative organism, under EPA's scoring scheme. However, the experts
expressed caution about this potential grouping issue. Experts collectively suggested that a
better description of the scoring process is necessary. Also, scoring variability should be
presented in the CCL 3 documentation, possibly graphically. In the future, an independent group
of clinicians is recommended to best score the pathogens on health effects to potentially
eliminate some scoring variability.
Expert Question: Although good scoring examples were presented, there are other
manifestations of disease associated with these pathogens. Furthermore, most common
manifestations would score differently than other possible or extreme manifestations. How does
this get factored in?
EPA Response: This statement is true. However, the scoring process included a mechanism to
avoid variability. Specifically, EPA reviewers were instructed to refer to the data facts sheets
and score on the common disease presentation with each subpopulation. Variations occurred
when staff brought in additional, data or background information.
Expert Summary Comments: EPA has scored subjectively. Severity, potency, and frequency
were "rolled" into health effects scoring because there was limited data on these criteria.
However, the information is available. Perhaps for future CCLs there should be separate scoring
process for other manifestations of disease, other than the most common manifestations. In
Page 13 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
addition, there should be a different process for chronic diseases. The following limitations to
the scoring and screening approach should be well documented in the CCL 3 literature:
1. Bias towards pathogens that cause acute WBDO. For example, Helicobacter pylori takes
a long time to grow; and therefore, does not cause an immediate WBDO. It may take
weeks to months to recognize disease caused by Helicobacter pylori, so it is unlikely that
a WBDO would ever be recognized.
2. Difficulties in evaluating the quality of information on which scoring is based (e.g., what
journal should be cited).
3. Subjectivity in converting scientific data (non quantitative) to a numerical score.
4. Difficulties in quantifying uncertainty in scoring, particularly reproducibility (i.e., limited
resources to do statistical analysis and missing information).
Experts clarified that they would try to reproduce the scoring using the data given to them and
scoring process provided by EPA. Therefore, other (or extreme) manifestations of disease other
than primary will be ignored, except when there are pathogens with severe outcomes that have a
higher percentage of occurrence - these should be ranked higher. The following table reflects
the experts' illustration of how the health effects score was normalized to the occurrence score.
It shows minor shifts in the health effects scoring (i.e., 5 vs. 4 for a subpopulation) would have
minimal impact in the overall scoring process. However, some scoring changes would have an
effect on pathogens clustered around a cutoff developed by EPA. Experts also used this scoring
table to assist them in their rescoring of the PCCL (described in later sections).
Exhibit 2: PCCL Rescoring Data
Health Effects Score
14
13
12
11
10
9
8
7
6
5
4
3
2
Normalized Adjustment Ratio
(Maximum WBDO or Occurrence over
the Maximum Health Effects Score)
5/14
5/14
5/14
5/14
5/14
5/14
5/14
5/14
5/14
5/14
5/14
5/14
5/14
Adjusted Score
5.0
4.6
4.3
3.9
3.6
3.2
2.9
2.5
2.1
1.8
1.4
1.1
0.7
Page 14 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
Question #2: Do the fact sheets adequately represent the available and updated
information for each contaminant?
Expert General Issues/Comments:
• Mycobacterium avium: Most infections are minor aside from elderly and those with chronic
disease, so reviewers questioned why Mycobacterium avium received a score of 5 in the
general population. Experts noted that HIV patients could have more complications in
overall population, but the scoring does not account for this disease burden. Experts
concluded that the scores will be re-evaluated.
• Salmonella enterica: Experts stated that this is a marker organism in terms of testing, so it
represents all Salmonella organisms. Experts recommended that EPA change the child and
elderly health effects scores from 3 to 4 due to disease-related complications in these
subpopulations. Experts noted that there is no distinction in occurrence (waterborne) or
frequency because of limited data. Experts concluded that their scoring recommendations be
verified.
• Hepatitis E virus: Experts noted that it is often difficult to determine transferability from
one species to another (e.g., from swine to human) of this pathogen. However, it is important
to look for future transferability data and consider this data when evaluating the CCL.
Outside the US, Hepatitis E has high serial presence and significant mortality in acute cases.
There is epidemiological evidence of exposure and high mortality rate; however, it is difficult
to determine what is common versus extreme exposure issues. Experts discussed the lack of
US epidemiological data and whether it is appropriate to default to extreme or common
disease manifestation in this case. The unknown factor is of persons exposed - how many
become infected? Experts concluded that they are unable to properly evaluate the health
effects data without information regarding disease frequency.
• Cyclospora cayetanensis: Experts recommended that EPA change the fact sheet to reflect
one US outbreak detection in water.
• Naegleriafowleri: Experts noted that there has been only 1 outbreak in drinking water;
others were in recreational surface water. This information should be corrected on the fact
sheet. Experts also commented that the health effects data are unclear. They suggested that
EPA consider developing a health advisory, if the Agency decides not to regulate this
pathogen.
• Exophialajeanselmei: Experts noted that common manifestation of illness is minimal - sub
clinical (soft tissue infection); however, the pathogen can cause severe disease (uncommon).
Experts raised the question whether it is truly waterborne. Cadmus explained that fungi were
found in PWS distribution (probably from construction - soil contamination). Experts
recommended that EPA re-evaluate the screening of all fungi for future CCLs, and possibly
eliminating fungi, since they are not believed to be waterborne (soil based). However, they
concluded that they do recommend changing current CCL 3 criterion # 9.
Page 15 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
Question #2 Expert Summary Answer: Experts decided that they can not properly answer this
question because there was not enough time for the group to thoroughly evaluate data
references. However, it appears that the draft fact sheets may not adequately represent the
available and updated information for each contaminant; a number of data omissions were
pointed out.
Sub-question #3.1: Do the scoring protocols adequately address the microbes and available
data necessary to characterize each attribute?
The experts commented that they needed to thoroughly review the fact sheets and evaluate each
pathogen using the scoring protocols provided by EPA to accurately answer this question. They
also detected errors in the translation of the information on the fact sheet to an actual score and
recommended that EPA conduct a quality assurance check to ensure that translation of factual
data to numeric values.
Based on their judgment, the experts re-scored the draft PCCL. The draft PCCL was ranked by
total score, in descending order (see table on page 20). Experts disagreed on the order in which
the pathogens exist on this list. For example, some commented that magnitude of disease is still
not properly reflected in this draft PCCL. Others commented that some pathogens were still not
scored high enough and others are ranked too high on the PCCL.
Experts also discussed whether they should accept imperfections in the scoring methodology or
should they recommend specific changes. Experts raised the question whether it would be
possible to develop a weighting rule for health effects based on population burden. EPA staff
noted that frequency of disease is an evaluation that is reviewed, and further evaluated frequency
during the regulatory determination.
Sub-question # 3.1 - Expert Summary Answer: Experts agreed that the current revised
scoring does not reflect the organisms that they are most concerned about. The group also
agreed that frequency and population disease burden is missing in the scoring criteria. This
information would help to properly reflect the public aspect of the public health risks. It is
important to factor in the public burden of disease and frequency to make it reasonable in
terms of true public health risk. Experts concluded that the current scoring system was not
acceptable, particularly since the overall scores were the factors that determined which
organisms would be selected for intervention. They felt that the scoring should be revised;
however they did not receive an answer from EPA whether this would be possible. They
recommended that the public burden of disease needs to be included in the scoring method.
Question #4: Is the ranked CCL produced by the scoring protocol process reasonable?
Question #4 - Expert Summary Answer: Inconsistencies have been pointed out in above
discussions and should be included in the CCL documentation - if this occurs, they would
agree that the scoring protocol process is reasonable. Experts would recommend their revised
PCCL list is if it is quality assured/quality controlled by outside experts (clinicians).
Page 16 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
Expert Microbial PCCL List:
Experts evaluated and rescored each microbe on the EPA PCCL. This table reflects the rank
ordered list based on expert review. The comments note why scores were changed.
Exhibit 3: Rank Ordered Microbe List Based On Expert Review
Pathogen
Naegleria
fowleri
Legionella
pneumophila
Escherichia
coli
(pathogenic)
Shigella
sonnet
Giardia
duodenalis
Human
enterovirus
Salmonella
enterica
WBDO
4
5
5
5
5
5
5
Occurrence
3
3
3
3
3
2
3
Health Effects
General
7
4
3
3
3
2
3
Child
7
3
6
6
5
6
4
Elderly
7
6
6
4
3
4
4
Pregnant
Women
7
4
o
J
o
J
3
2
3
Chronic
Disease
7
6
o
J
o
J
3
2
3
Normalized
Health
Effects Sum
5.0
3.6
3.2
3.2
2.9
2.9
2.5
Total
Score
9.0
8.6
8.2
8.2
7.9
7.9
7.5
Comments
No score
changes, yet only
1 WBDO -
drinking water
(others rec.) -
data must be
changed on fact
sheets.
Significant
possibility of
pneumonia in
chronic & elderly
- raised scores.
Pathogenic
#0157. Marker
org.
Hemmoralogic
colitis - only
major disease
assoc.
Same
complications as
Salmonella in
children.
Did not re-score
or evaluate b/c
already regulated.
A collection of
viruses that vary
in severity &
occurrence.
Question of
severity vs.
commonality -
changed scores to
reflect.
Marker org. - rep.
all Salmonella
orgs. Raised
child score -
Page 17 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
Pathogen
Campylobact
erjejuni
Cryptosporid
ium parvum
Hepatitis A
virus
Vibrio
cholerae
Rotavirus
Arcobacter
butzleri
Entamoeba
histolytica
Human
adenovirus
Human
caliciviras
WBDO
5
5
5
5
4
5
5
5
5
Occurrence
3
3
3
1
2
3
1
2
3
Health Effects
General
3
3
3
3
3
3
3
2
2
Child
6
3
3
4
6
3
3
4
4
Elderly
4
3
4
4
o
J
o
J
3
4
4
Pregnant
Women
o
J
3
3
4
o
J
o
J
3
2
2
Chronic
Disease
o
J
4
3
4
o
J
o
J
3
2
o
J
Normalized
Health
Effects Sum
3.2
2.5
2.5
2.5
3.2
2.1
2.1
2.1
2.1
Total
Score
8.2
7.5
7.5
7.5
7.2
7.1
7.1
7.1
7.1
Comments
complications.
Gastroenteritis is
self-limiting
(except in
children).
Doc. Occurrences
were not
waterborne.
Hospitalization
not always nee. -
lowered scores.
Only 1
documented case
(lower occ score),
50-80 child
deaths per year
(R. Glass) - raise
scores to reflect.
Outbreaks in
Idaho & Ohio.
Taxonomy
grouping
questionable.
Lesser score b/c
complications are
uncommon, no
hospitalization.
More adult cases
than children, do
not always
require
hospitalization -
lowered child
score.
Documented
deaths &
hospitalization &
resistant to
chlorine - raised
chronic score.
Page 18 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
Pathogen
Plesiomonas
Cyclospora
cayetanensis
Mycobacteri
um avium
Hepatitis E
virus
Fusarium
solani
Toxoplasma
gondii
Exophiala
jeanselmei
WBDO
5
4
4
2
1
2
1
Occurrence
3
3
3
2
3
1
3
Health Effects
General
2
3
3
51
4
2
3
Child
3
4
3
3
4
2
3
Elderly
3
4
4
3
4
2
3
Pregnant
Women
2
o
J
3
7
4
7
3
Chronic
Disease
3
o
J
3
3
4
2
3
Normalized
Health
Effects Sum
1.8
2.5
2.5
4.3
2.9
3.2
2.1
Total
Score
6.8
6.5
6.5
6.3
5.9
5.2
5.1
Comments
Not on orig.
PCCL,butEPA
added b/c of new
data.
US occurrence -
changed score.
Most infections
are minor (except
elderly & those
w/ chronic
disease).
Can't eval.
General health
effects w/out
manifestation
data (extreme or
common) or
frequent
(exposure to
disease), except
in pregnant
women.
Exposure is very
frequent but
common
manifestation
asymptomatic.
Causes fetal
illness so raised
pregnancy score.
Asymptomatic
disease in general
pop, so score
lowered.
Common
manifestation of
illness is
minimal, sub-
clinical (soft
tissue infestation
mainly). Severe
disease possible
but unlikely -
scores reflect.
Page 19 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
Pathogen
Helicobacter
pylori
Yersinia
enterocolitic
a
Human
astrovirus
Aspergillus
fumigatus
Aeromonas
hydrophila
Human
coronavirus
Microsporidi
a
Blastocystis
hominis
Isospora
belli
WBDO
1
1
2
1
1
1
1
1
1
Occurrence
3
2
2
3
3
1
2
1
1
Health Effects
General
3
3
2
3
2
2
2
2
2
Child
3
4
2
3
3
2
2
2
2
Elderly
3
o
6
2
3
3
2
2
2
2
Pregnant
Women
3
o
3
2
3
2
2
2
2
2
Chronic
Disease
3
o
3
2
3
2
2
2
2
2
Normalized
Health
Effects Sum
2.1
2.5
1.4
2.1
1.8
1.4
1.4
1.4
1.4
Total
Score
5.1
4.5
3.4
5.1
4.8
2.4
3.4
2.4
2.4
Comments
Lesser score - too
weighted re:
malignancy
concern (less than
10% to
gastroenteritis).
Common
manifestation
literature avail.
EPA cited
outbreaks in
water so occ.
Will remain as 2.
Appendicitis
possibility in
children only -
adjusted scores
accordingly.
No data of
hospitalization.
No changes in
scores.
Outbreaks in
Bangladesh &
Nigeria.
Only 1 case,
overseas - sewage
inhalation (not
drinking water).
Detection is not
up to par, adverse
health effects less
likely in sub-pops
- lowered scores.
No score
changes.
1. Subsequent to the workshop, reviewers provided varying comments regarding the general scoring for Hepatitis E
Virus. One expert requested that the score of 5 is lowered to a 2 due to the minimal probability of subclinical
infection. Another expert requested that the general score be changed to a 3 rather than a 2 because the incidence
data for subclinical infections is not well-defined. The score was not lowered due to the lack of group consensus.
Page 20 of 22
-------
EPA-OGWDW
Microbial Expert Input and Review
for CCL 3
EPA815-R-08-010
February 2008 - DRAFT
General Charge Question: Does the Draft CCL 3 microbe list represent those pathogens
that have the highest potential to occur in public water systems and cause adverse human
health effects? Are there pathogens on the Draft CCL 3 list that should not be listed and,
conversely, are there pathogens that should be listed?
The experts went through each of the contaminants on their newly developed PCCL list and
voted on what should and should not be included on the draft CCL. It is important to note that
pathogens carried forward from the expert's PCCL list are not necessarily based on score or
rank. The panel felt that any "cutoff value would be arbitrary. Selection of a population of
organisms for control would have to consider additional factors than simply an overall score.
This is illustrated in this exercise.
Experts voted on each on the pathogens based on their potential to cause adverse human health
risks and their prevalence in water. In addition, not all experts were prepared to render a
definitive vote on all pathogens. The ranking below reflects the general consensus only of those
who voted.
Exhibit 4: Rank Reflecting General Consensus of Experts that Voted
Expert CCL
Pathogen
Legionella pneumophila
Escherichia coll
Shigella sonnei
Human enterovirus
Campylobacter jejuni
Salmonella enterica
Hepatitis A virus
Rotavirus
Human adenovirus
Human caliciviruses
Mycobacterium avium
Total
Score
8.6
8.2
8.2
7.9
7.5
7.9
7.5
7.2
7.1
7.1
6.5
Number of Expert Panel*
For Inclusion
6
5
4
6
4
4
6
3
6
6
4
Against Inclusion
0
0
0
0
0
0
0
3
0
0
2
Abstained
0
1
2
0
2
2
0
0
0
0
0
* All other organisms received one or no votes for inclusion.
However, this exercise was done only to gauge the experts' reaction to some changes to the CCL
based on their new scoring. What was ultimately decided is that overall, EPA's approach was
reasonable. Experts agreed with EPA's ranking, but they identified some concerns. For
example, although Naegleriafowleri received a high score, it may not be appropriate to include
it on the CCL due to its rarity in the environment. Experts also discussed the issue of organisms
that are controlled with treatment such as Salmonella enterica, Shigella sonnei, Yersinia
enterocolitica and Vibrio choleme, and their standing on the CCL; however, there was no
consensus on this issue. In addition, experts discussed the issue of emerging contaminants such
as Helicobacter pylori that require additional research, yet may be discovered to have great
Page 21 of 22
-------
EPA-OGWDW Microbial Expert Input and Review EPA 815-R-08-010
for CCL 3 February 2008 - DRAFT
health effects consequences; and therefore, should be elevated on the CCL. Furthermore, it was
discussed that the current protocol may not account for chronic diseases and may focus on acute
disease manifestation and data, such as the case with Helicobacterpylori.
Page 22 of 22
------- |