2
Sociopolitical Considerations for Developing Future CCLs

INTRODUCTION

In its two previous reports (NRC, 1999a,b), the committee emphasized that the decisions related to the development of a Drinking Water Contaminant Candidate List (CCL), as well as in the prioritization of CCL contaminants for subsequent regulatory or research activities, must be scientifically defensible and transparent. In Chapter 1, the development and implementation status of the first CCL and two related programs are described briefly. This chapter discusses the nature of the perennial task facing the U.S. Environmental Protection Agency (EPA) in developing and implementing a decision-making process for the creation of future CCLs in conjunction with several important social and political issues. More specifically, the interrelated issues of sound science, risk perception, social equity, legal mandates to consider the risks for vulnerable populations, and the proper role of transparency and public participation are discussed. In addition, several potentially helpful conceptual frameworks for risk perception and public participation are described. In so doing, the committee seeks to provide a solid foundation and background support for a systematic, scientifically sound, transparent, and equitable approach to the development of future CCLs, which is fully described later in this report.

LIMITATIONS OF THE FIRST CCL DEVELOPMENT PROCESS

Because of the severe time constraints stipulated by the Safe Drinking Water Act (SDWA) Amendments of 1996 for publication of the first CCL, EPA was forced to develop and utilize a decision-making process for the creation of the 1998 CCL that the committee feels was not suffi-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration 2 Sociopolitical Considerations for Developing Future CCLs INTRODUCTION In its two previous reports (NRC, 1999a,b), the committee emphasized that the decisions related to the development of a Drinking Water Contaminant Candidate List (CCL), as well as in the prioritization of CCL contaminants for subsequent regulatory or research activities, must be scientifically defensible and transparent. In Chapter 1, the development and implementation status of the first CCL and two related programs are described briefly. This chapter discusses the nature of the perennial task facing the U.S. Environmental Protection Agency (EPA) in developing and implementing a decision-making process for the creation of future CCLs in conjunction with several important social and political issues. More specifically, the interrelated issues of sound science, risk perception, social equity, legal mandates to consider the risks for vulnerable populations, and the proper role of transparency and public participation are discussed. In addition, several potentially helpful conceptual frameworks for risk perception and public participation are described. In so doing, the committee seeks to provide a solid foundation and background support for a systematic, scientifically sound, transparent, and equitable approach to the development of future CCLs, which is fully described later in this report. LIMITATIONS OF THE FIRST CCL DEVELOPMENT PROCESS Because of the severe time constraints stipulated by the Safe Drinking Water Act (SDWA) Amendments of 1996 for publication of the first CCL, EPA was forced to develop and utilize a decision-making process for the creation of the 1998 CCL that the committee feels was not suffi-

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration ciently defensible, transparent, or available for public comment. In particular, several major policy decisions were made during the process that lacked sufficient explanation and justification. By identifying and thoroughly discussing these limitations, the committee aims to illustrate how important it is to arrive at a more systematic CCL development mechanism that directly addresses these and other sociopolitical issues. The committee notes that its second report (NRC, 1999b) also describes (albeit summarily) some limitations of the first CCL development process. A major policy decision made during the development of the first CCL was to use completely separate approaches to evaluate potential chemical and microbiological drinking water contaminants. The committee believes that the justification given by EPA (EPA, 1997a) for conducting two independent assessments was not adequate and perpetuates the long-established and often unnecessary regulatory practice of treating chemical and microbial drinking water contaminants separately and differently. As reviewed in the committee’s first report (NRC, 1999a) and in more detail in Chapter 6 of this report, rather than regulating each type of microorganism to a specific concentration as done for chemicals, regulators historically have established a “zero-tolerance” goal for microbiological contaminants. Indicator organisms, particularly fecal coliforms, are then used to show the possible presence of microbial contamination resulting from human waste. While this approach has served well for indicating widespread sewage contamination of surface waters and for controlling diseases such as cholera and typhoid fever, several deficiencies in this approach have come to light in recent decades. For example, some bacteria and many viruses and protozoa show greater resistance to conventional treatment methods than do fecal coliforms. This approach has also led to a deficiency of occurrence databases for microbial contaminants. The committee continues to believe that the time is rapidly approaching when the same risk assessment principles should be applied to the management of microbial contaminants as are applied to chemical contaminants. Indeed, as described in Chapter 1 and elsewhere in this report, the committee recommends a two-step approach for the development of future CCLs that will similarly assess chemical, microbial, and other types of potential drinking water contaminants. Second, all potential chemical drinking water contaminants that were considered initially for inclusion on the CCL were taken directly from existing databases and lists of chemicals produced by various regulatory programs within EPA and by stakeholder groups (see Table 1–1; EPA, 1997a). Although this was useful for developing a CCL in a short time

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration period, EPA excluded from consideration several tens of thousands of chemicals that might pose a threat to safe drinking water but have not yet been identified or included on one of the selected lists (NRC, 1999b). For example, the first CCL development process did not evaluate radionuclides or pharmaceuticals. Moreover, reliance on databases created primarily for purposes other than supporting drinking water research and regulatory activities led to the inclusion of chemicals (e.g., acetone) on the draft 1998 CCL that were ultimately determined not to merit such attention and were subsequently dropped. In addition, reliance on these existing lists and databases implies a policy decision to focus the CCL solely on chemicals already identified by EPA programs, increasing the likelihood that a newly emerging chemical threat would be missed by the process unless it caused a major problem. Although the chemical and microbial contaminants included on the 1998 CCL certainly merit regulatory and research attention, a broader approach to contaminant selection could potentially identify higher-risk contaminants. In short, the committee characterized this overall approach as “looking under the lamp post” for a relatively few types and numbers of drinking water contaminants compared to the universe of potential contaminants (NRC, 1999b; see Chapter 3 for further information). A third major policy decision was the exclusion of contaminants from further consideration if their occurrence in drinking water was not first determined to be demonstrated or anticipated based on available information (EPA, 1997a). By evaluating occurrence before any evaluation of health effects data, drinking water contaminants with potential adverse health effects would be excluded from the 1998 CCL solely because of missing or inadequate data on occurrence. Fourth, EPA decided to defer consideration of 23 chemicals and chemical groups for inclusion on the draft 1998 CCL based solely on the possibility of their being endocrine disruptors (EPA, 1997a). The rationale for this policy decision was that these types of chemicals were then under general review by EPA’s Endocrine Disruptor Screening and Testing Advisory Committee and another NRC committee. Similarly, a list of 35 pesticides that indicated a high risk of leaching into groundwater but met no other criteria for inclusion on the draft 1998 CCL were also deferred pending further evaluation by EPA’s Office of Pesticide Programs for their potential to occur at levels of health concern. Both of these policy decisions imply that research on the health effects of these important chemicals and their occurrence in drinking water could be de-

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration layed at least until the development of future CCLs. Finally, the stipulated size of the CCL reflected an unstated policy decision on the general amount of resources EPA would devote to the regulation of drinking water contaminants. In this regard, a total of 262 chemicals and groups of related chemicals were identified initially for consideration during preparation of the 1998 draft CCL (EPA, 1997a). As noted in the committee’s first report (NRC, 1999a) however, EPA made it clear that the total number of contaminants on the draft and final 1998 CCLs would have to be reduced from 262. Ultimately, the final 1998 CCL included a total of 60 individual chemicals, or groups of related chemicals, and microorganisms (EPA, 1998a). Thus, for planning and discussion purposes, the committee has always considered that future CCLs would similarly be limited to no more than 100 or so total contaminants as noted elsewhere in this report. Indeed, the committee feels that EPA’s difficulty in developing a highly complex research plan for the 1998 CCL (reviewed in Chapter 1) supports this contention. The committee also notes that the entire 1998 CCL development process allowed for only limited public participation. As noted in Chapter 1, EPA relied heavily on the advice of the National Drinking Water Advisory Council (NDWAC) Working Group on Occurrence and Contaminant Selection for the development of the approach used to create the draft 1998 CCL. Although NDWAC’s meetings are open to the public and include experts from various stakeholder groups, it cannot be considered to represent the broad spectrum of public views on this issue. Thus, for the most part, public input into the development of the 1998 CCL was confined to the two-month public comment period for the draft 1998 CCL. The committee, several public commenters, and EPA recognized the need for a more systematic, scientifically sound, and transparent process for selecting contaminants for future CCLs (EPA, 1998a,c; NRC 1999a,b). Specific comments indicative of the lack of transparency in the decision-making process for the draft 1998 CCL included calls for clarification of the process as a whole by EPA so that it could be more fully understood (EPA, 1998c). Requests were also made to EPA for explanations and justifications of the screening criteria used to narrow the field of candidate contaminants. Questions were raised as to how these specific criteria were selected, defined, operationalized, and weighted. Several other comments focused specifically on the health effects criteria (e.g., whether carcinogenic effects received priority over other health effects or whether priority was given to contaminants with more complete toxicological data). Some commenters complained of

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration insufficient explanation and justification for the inclusion or exclusion of specific contaminants. To a large extent, the widespread recognition of these limitations helped lead to the formation of this National Research Council (NRC) committee at the request of EPA to advise it on developing regulatory and research priorities for the 1998 CCL and the creation of subsequent CCLs. Thus, a detailed discussion of these issues provides a necessary and appropriate foundation for much of this report. The committee believes that any proposed CCL classification scheme must directly incorporate the principles of transparency, scientific defensibility, and equity if it is to be successful at the policy level. USE OF SOUND SCIENCE IN FUTURE REGULATORY DECISIONS The CCL is intended to be a central element in EPA’s future regulatory strategy for its drinking water program. Section 1412(b)(3)(A)(i) of the amended SDWA requires EPA to use the “best available, peer-reviewed science and supporting studies conducted in accordance with sound and objective scientific practices…”. The use of peer-reviewed science will not, however, guarantee agreement among all parties that might be affected by the listing of a chemical or microorganism on a CCL since scientists often weigh the different strands of evidence and supporting data differently. Indeed, the committee explicitly noted in its preceding report that expert judgment must play a substantial role in the process of developing a CCL (NRC, 1999b). Furthermore, disagreements on some CCL listings are to be expected and do not necessarily indicate that they are unsound. Rather, the soundness of the judgments will have to be decided in the more-or-less usual way of reasoned and supported argument among the contending and interested parties. This question was dealt with in greater detail in the first report of the committee, and its conclusions are repeated here (NRC, 1999a): [The committee] takes the position that scientific disagreements are the norm and do not signal a deviation from sound science. These disagreements may be based on values other than strictly scientific ones, however, this does not mean that the sides of the debate are not based on sound science. Indeed, it is not unusual for scientists to disagree on the application of sound science to public policy issues. Any scheme that

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration affects the provision of public water is likely to engender legitimate scientific disagreement. The report also recognizes that identifying and agreeing on what is sound science is itself a difficult and error-prone enterprise. It therefore makes no recommendations on what “soundness” entails, letting the accepted mechanisms of peer regard, peer review, and scientists’ habits of critical thinking continue to serve as the ultimate arbiters. NATURE OF THE TASK The daunting task before the EPA and for which the committee is charged to provide advice is how to take a large, unordered set of chemicals, microorganisms, and other potential drinking water contaminants and separate it into two sets:1 one (the CCL) to receive occurrence monitoring or research of some sort preparatory to an eventual decision to regulate or not and another set that will receive no regulatory or research attention aside from that otherwise dictated by the advancing interests of science and commerce. This sorting of contaminants into two discrete sets based upon available occurrence and health effects attributes gleaned from the scientific literature is a classification problem. More specifically, the committee’s objective is to recommend a scientifically sound, transparent, and equitable process that can be used to identify and cull from the universe of potential drinking water contaminants a list (CCL) that contains primarily contaminants for which EPA may justify expending considerable, albeit limited, resources to develop drinking water regulations or to pursue occurrence monitoring or health effects, treatment, or analytical methods research. This is the central idea behind the CCL as required under the SDWA Amendments of 1996. In effect, this means that EPA is probably limited to preparing a CCL that has perhaps up to 100 contaminants and groups of related contaminants on it. This represents a full two-order-of-magnitude reduction from a total universe of potential drinking water contaminants that may number 1   A third set could be considered that includes those drinking water contaminants that are (or can be) determined to present an urgent threat to public health. As noted in Chapter 1, the amended SDWA specifically allows EPA to circumvent the CCL process and issue interim regulations for such contaminants. Thus, they are not included in this discussion and are not a major focus of this report.

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration in the tens of thousands given the number of chemicals currently in commercial use.2 As noted in Chapter 1 (see Figure 1–3), the committee has previously suggested (NRC, 1999b) and continues to suggest a preliminary step that reduces the universe of potential contaminants to a smaller list of perhaps a few thousand (the preliminary CCL or PCCL). Before the problem of reducing this list to the CCL can be discussed; however, an overview of the difficulties of conducting even a coarse screen to get to the PCCL is in order. If one were to assume, as seems plausible, that only a very small proportion of the universe of potential drinking water contaminants are likely to be high-risk contaminants (e.g., of sufficient toxicity and likelihood to become prevalent in drinking water), any screening of this list must be highly accurate with respect to correctly identifying those contaminants that are not a problem. Appendix A of the committee’s first report (NRC, 1999a) included a quantitative analysis of this situation, many elements of which are summarized below. The arithmetic of culling a list containing perhaps tens of thousands of potential drinking water contaminants to a much smaller, but still large, list of “high-risk” contaminants quickly reveals some inherent dilemmas. Assuming one could reduce such a list of contaminants down to a list of a few thousand (the PCCL) with some accuracy (e.g., it would include only chemicals with a genuine potential to occur in drinking water and cause adverse health effects in exposed persons), this judgment must be made with considerable accuracy. If one were to make an error in only one of a hundred contaminants, wrongly believing it has potential for contaminating drinking water or that it has more health significance than it actually does, almost a quarter of the contaminants on a smaller list of around 2,000 (the PCCL) will be “false positives” (one of a hundred contaminants on a list of about 50,000 produces 500 false positives). Thus, a very high level of specificity is necessary to avoid cluttering the PCCL with a high proportion of relatively harmless contaminants. (In this case, a specificity of 99 percent is the proportion of harmless contaminants that are correctly identified as harmless). On the other hand, the requirement for a relatively high level of specificity runs counter to the usual public health emphasis of acting in a health-protective manner. As stated in both of its previous reports (NRC, 2   The total number of contaminants in this universe could be quite large, based on the European experience (see Appendix A of this report) and given that the Toxic Substances Control Act inventory of commercial chemicals alone includes about 72,000 substances (NRC, 1999b).

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration 1999a,b), the committee continues to recommend that EPA “err on the side of public health protection,” opting for high sensitivity in the generation of future CCLs. However, the opposing requirement not to include false positives makes it more likely that certain contaminants that may truly pose a health risk will be passed over as well (i.e., false negatives are left behind while trying to avoid false positives). Another approach would be to proceed “from the other direction” by starting at the final list (the CCL) and trying to populate it with contaminants already known or reasonably suspected to be problematic on other grounds (e.g., chemicals that already have health advisories or are included on some list of environmental concern). Using this strategy, one would be ignoring the huge bulk of contaminants for which little or no data have been accumulated. To a large extent, the draft 1998 CCL was developed using just such an approach. Yet the unwelcome “surprises” (e.g., methyl-t-butyl ether [MTBE]) invariably arise from the large group of contaminants with little or no data. Perhaps an argument could be made that this practice is acceptable since the high-risk characteristics described should receive priority in any regulatory scheme and EPA can circumvent the CCL process to develop interim regulations for any contaminants that are determined to represent an urgent threat to public health (EPA, 1998a). However, it is important to note that one of the goals of the CCL—to avoid such surprises—would be compromised to some extent. The unpleasant arithmetic properties of screening large numbers of contaminants aside, one must also consider the difficulty of conducting any classification task where imperfect and incomplete data must be used to answer a sophisticated question: in this case, Does a chemical, microbiological, or other type of contaminant pose an existing or future threat to drinking water supplies? Because of its general importance, the task of classification has been the subject of a great deal of research in recent years (see, for example, Bowker and Star, 1999). One need only consider commonplace examples to see how difficult and complex a task it can be. For example, applications to college are made on a relatively standard data collection format, including readily quantifiable scales such as SAT scores, grade point averages, and demographic data; have little or no missing data; and have sufficient history and numbers to make quantification and statistical investigations with adequate power possible. Despite these strengths, the level of public confidence in the accuracy of the classification procedure (i.e., admit-don’t admit) is not high, with an understanding that many worthy applicants will be rejected and some relative failures will be ad-

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration mitted. Perhaps worse, suspicion exists that students are admitted or rejected in a nonrandom fashion although the underlying mechanisms are often hard to discern. To summarize, the placement of a contaminant first on the PCCL and from there on the CCL involves not one, but several, difficult classification judgments. For example, does the chemical produce a health effect? (“Health effects” is a category.) What are the specific health effects? (Health effects are categorized into different types or diseases that change periodically.) What is the nature of the evidence? (Evidence is placed into several, often overlapping categories such as animal studies, epidemiological studies, and “reliable” studies.) What appears at first to be a “simple” classification exercise quickly reveals itself to be a complex task in which many choices are being made, some explicitly, some as a result of given prior classifications, and some implicitly or without conscious knowledge of the classifiers. RISK PERCEPTION The selection of microorganisms, chemicals, and other types of contaminants for inclusion on a CCL will be based on judgments of their potential health risks, including an evaluation of the severity of their effects, their potency, and the likelihood of their occurrence in drinking water. These risk judgments must be made in a context of considerable scientific uncertainty (e.g., due to data gaps and the use of models to estimate potency and environmental fate), complexity (e.g., variability in the vulnerability of subpopulations and the effects of contaminant mixtures), and controversy (e.g., issues concerning acceptable evidence, relevant data, and the fairness and acceptability of risks). Faced with this context, people rely on their assumptions, values, beliefs, and in general, their worldviews, as well as on the information available to them (which to some extent is also dependent on their worldviews) in order to make judgments about risks. Therefore it should not be surprising that people differ in their perceptions of risk and that these disagreements reflect differences in their worldviews. A worldview can be defined as a deeply held “orienting disposition” toward the world and its social organization that guides a person’s perceptions, interpretations, analyses, and responses in a wide variety of complex situations (Peters and Slovic, 1996). An individual’s worldview develops from his or her life experiences, social interactions, and educa-

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration tion. It includes perceptions of self-identity and local environment (e.g., local hazards, local socioeconomic conditions), political beliefs and moral values, and views and values held by the social groups with which an individual identifies and belongs. The prevailing views of one’s discipline, or the values of the institution by which one is employed, may also be incorporated into one’s worldview and influence how one perceives risks (see, for example, Barke and Jenkins-Smith, 1993; Kraus et al., 1992; Slovic et al., 1995). Another important aspect of an individual’s worldview is the primary reasoning scheme a person uses in complex situations such as those involving the risk perception of an environmental hazard. Research over the past 30 years has emphasized the differences in risk perception between scientists and “lay people” who are not professionally trained in science (e.g., Sowby, 1965; Starr, 1969). According to this research, scientists tend to adopt a quantitative approach to risk and emphasize considerations such as dose and exposure. On the other hand, lay people generally adopt a more qualitative approach, emphasizing the fairness and voluntary nature of the risk, its effects on future generations, and the characteristics of the risk such as whether it is known, uncontrollable, or capable of producing catastrophic effects (Tesh, 1999). At the aggregate level, surveys have found differences in average risk perception scores when scientists as a group were compared with lay people as a group (Barke and Jenkins-Smith, 1993; Kraus et al., 1992; Slovic et al., 1995). However, these surveys also found a wide spectrum of views on risk perception within each group indicating that at the individual level, the risk perceptions of many scientists and lay people may not differ substantially. For example, it is not unusual that when confronted by complex environmental hazards, lay people often become self-taught, scientific experts who work closely with professional scientists to evaluate the technical aspects of the risks (Brown, 1992, 1997; Tesh, 1999). On the other hand, many scientists have attempted to achieve a balance between quantitative and qualitative approaches to the evaluation of environmental risks. The wide spectrum of views on risk perception found at the individual level among scientists as well as among lay people reflects differences in worldviews within these two heterogeneous groups. Three schemes of reasoning can be differentiated theoretically and are discussed below to aid understanding of risk perception (although an individual’s primary reasoning scheme is likely to be some combination of these) as related to the classification of drinking water contaminants for regulatory consideration. In the “utilitarian” reasoning scheme, each

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration person is assumed to be a self-interested “utility calculator,” determining the optimum balance of personal satisfaction from among various options and the information available (Anderson, 1993). At the societal level, the goal is the efficient and optimum distribution of costs and benefits so as to maximize benefits for the majority of society’s members. In this scheme, risks would be ranked on a common metric such as expected number of deaths per year, quality adjusted life-years (QALYs), or disability adjusted life-years (DALYs), and this metric would provide the basis for comparisons and cost-benefit trade-offs among risks. This reasoning scheme tends to juxtapose widely different forms of death (e.g., immediate death, death after a long and debilitating illness) as well as voluntary and involuntary risks (Bennett, 1999). It is also generally indifferent to issues of equity or social justice (Morrow and Bryant, 1995). In the context of risk perception, this reasoning scheme tends to assume that some level of risk is necessary for “growth” and therefore attempts to answer in a quantitative, cost-benefit fashion the question: How safe is safe enough? Policies associated with (but not necessarily logically entailed by) this reasoning scheme tend to emphasize managing risks over risk avoidance and avoiding false-positive errors (e.g., ameliorating a risk that later proves harmless) over false-negative errors (e.g., failing to detect an important health hazard). A second reasoning scheme focuses on the interpretation, elaboration, and assertion of rights (or entitlements to rights) and has been called “liberal rationalism” (Anderson, 1993). Reciprocity (e.g., mutual tolerance, respect, trust, goodwill) is the key value, and social justice is the goal of this approach. In the context of environmental risk, this approach would be concerned with the fair distribution of hazards among social groups and would attempt to answer the following question: How fair is safe enough (Rayner and Cantor, 1987)? In this scheme of reasoning, the “average person” does not exist; instead risks are perceived as varying by social group. Policies addressing societal demands for environmental justice, the protection of vulnerable subpopulations, and the right to tap water that is safe to drink are based on this reasoning scheme. A third reasoning scheme, called “critical reason,” is concerned with “big-picture” issues such as a search for the common good, the essential purposes of policies and activities, and the characteristics of the ideal society (Anderson, 1993). It involves thinking self-consciously about big-picture goals and purposes and integrating the grounds for society’s basic values, beliefs, practices, priorities, and institutions with one’s own values, beliefs, interests, and activities. In this, ideals such as the goods

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration ately identified and protected (Parkin et al., 2000). As noted above, susceptibility can be defined either on the individual or the population scale. Thus, an updated working consensus of “susceptible” will be needed to identify and protect vulnerable subpopulations in accordance with the amended SDWA and in the creation of future CCLs. Without this clarification, different stakeholders are likely to approach the term with different conceptual frameworks regarding their perceptions of risk. As a consequence, they may have conflicting but unrecognized differences about what persons and groups should be considered eligible for inclusion in such subpopulations. For example, are susceptible people those who are at elevated risk because of exposure or because of an inherent, nonmodifiable trait? Are risk and susceptibility conceived as related primarily to a contaminant’s inherent characteristics (e.g., chemical structure or virulence of a waterborne pathogen), the host’s immune status, or some characteristic of the subpopulation itself? Until such issues are resolved, any one definition of susceptible is likely to have important public policy consequences because it may not necessarily address all people who need to be protected from the adverse health events associated with all contaminants. However, the committee emphasizes that none of these questions or the important issue of what “meaningful portion of the general population” actually means under the amended SDWA can be answered based solely on scientific findings. Rather, the answers will depend in part on societal values and on viable, democratic means of resolution. In this regard, the resulting deliberative process may require several iterations before a working consensus emerges (Franz and Jin, 1995; Malone, 1994). Strong and widespread social support will be important in implementing effective programs to ensure safe drinking water for vulnerable subpopulations. TRANSPARENCY AND PUBLIC PARTICIPATION Increasing public concern and activism on environmental issues have resulted in demands that federal, state, and local regulatory and enforcement policies be made more transparent and incorporate public participation. For example, one of the key demands of environmental justice advocates is “the right to participate as equal partners at every level of decision-making including needs assessment, planning, implementation, enforcement and evaluation.” (Gibbs and CCHW, 1995) To respond adequately to such legitimate demands, the decision-making process for

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration developing future CCLs should include an oversight mechanism to allow public participation in all aspects of its design, development, and implementation. The process would also have to be transparent (i.e., “…easily understood, where information about the policy is available, where accountability is clear, and where citizens know what role they play in the implementation of the policy”; Finkelstein, 2000) Transparency Fundamental to the notion of transparency are the principles of equity, fairness, and democracy. The transparency of a decision-making process may help to ensure that resulting decisions are not perceived as having been made capriciously or “behind closed doors.” Yet how can the transparency of a policy or decision-making process be evaluated when there is no consensus standard to measure it? To this end, the committee recommends a recently proposed information communication standard for risk assessment that could have broader applicability as a general standard for transparency in decision-making. This standard can be summarized as providing sufficient information on the decision-making process such that citizens are allowed “…the opportunity to place themselves in a similar position as the [decision-maker]…to make as informed a choice…as if they themselves had gone through the [decision-making] process…” (Hattis and Anderson, 1999). In other words, one of EPA’s major goals in developing future CCLs should be to explain the process sufficiently so that with the information supplied an informed citizen could arrive at their own reasonable and informed judgments. To meet such a standard would require that transparency be incorporated into the design and development of the decision-making process (and any models used in the process) in addition to being an integral component in communicating the details of the decision-making process to the public. The issue and importance of transparency were raised at a November 1999 stakeholder meeting to discuss the implementation status of the 1998 CCL. Notably, EPA pledged that it would produce as part of the forthcoming CCL regulatory determination process a detailed support document that would describe comprehensively the rationale for all decisions to regulate (or not) CCL contaminants and provide a review of all the data used to support such decisions (Michael Osinski, EPA, personal communication, 2000).

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration EPA has previously attempted to address the issue of transparency in its Guidance for Risk Characterization (EPA, 1995). In an accompanying memo, the former EPA administrator noted that stakeholders in environmental issues desire sufficient information to allow independent assessments and judgments about the significance of environmental risks and the reasonableness of EPA’s corresponding risk reduction actions. The administrator stated that “we must adopt as values transparency in our decision-making process and clarity in communication with each other and the public regarding environmental risk and the uncertainties associated with our assessments of environmental risk.” This entails a full and open characterization of risks. The EPA administrator added that by doing so, EPA would have to disclose the key scientific analyses and policy choices, uncertainties, and core assumptions that underlie its decisions. Furthermore, the guidelines state that achieving transparency in risk characterization would require a frank and open discussion of the uncertainty associated with an assessment as a whole and its components along with the impacts of key factors or variables on the overall decision-making process (EPA, 1995). They acknowledge that information from different sources carries different kinds of uncertainty and that understanding these differences is important when uncertainties are combined. The guidelines reassure risk assessors and managers that a frank and comprehensive discussion of uncertainties and their impacts on the decision-making process would not necessarily reduce the public’s perceived validity of the process, but instead would likely enhance public trust and serve as a useful indicator of the confidence decision-makers had in the process itself and in resulting decisions. Although not mentioned in the guidelines, an added benefit of such a discussion of uncertainties is that “it helps the [decision-maker] function more honestly in a context that may often exert pressures for more unambiguous answers than can readily be produced” (Hattis and Anderson, 1999). A recent evaluation of the use of models in an environmental assessment effort called the “ULYSSES Project” in Europe has similarly reported that a full discussion of uncertainties could promote public trust and enhance the credibility of the assessment (Dahinden et al., 1999). However, the authors warned that a full discussion of uncertainties could also produce public doubts and skepticism about the results of the assessment. In this regard, Yearley (1999) and Lopez and Gonzalez (1996) found that how the public receives a presentation of uncertainties is most influenced by its existing “lay knowledge” of a particular issue, its history of experience and participation with a related regulatory agency, and

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration the resulting accumulation (or deterioration) of trust and credibility concerning the agency’s agenda, policies, and decisions. An important question for EPA in developing future CCLs is, How much information should be provided to the public in order to effectively characterize the uncertainties in the decision-making process? The 1996 NRC report Understanding Risk: Informing Decisions in a Democratic Society warns that “simple characterizations are likely to give an erroneous impression of the extent of uncertainty, but more careful and elaborate characterizations may be incomprehensible to nonspecialists…”. Thus, the goal would be to characterize the uncertainties in such a way that citizens would understand the level of uncertainty in the process and “…appreciate where scientists agree and where they disagree.” In this regard, the committee again maintains that such disagreements are normal and expected and do not necessarily signal a deviation from the application of sound science. As discussed later in greater detail (Chapters 3 to 5), this report recommends and outlines a type of triage approach to move forward from a universe of potential drinking water contaminants to arrive ultimately at a much shorter list (the CCL) that will largely form the basis of EPA’s future drinking water program. A major aspect of this triage approach will involve the development and use of a “prototype” classification tool that may have a similar degree of complexity and uncertainty as models that have previously been utilized for the purposes of environmental risk assessment (e.g., those employed in the ULYSSES Project). Therefore, the committee believes that in the implementation of this recommended approach, EPA will have to deal with the same issues of transparency and the full discussion of uncertainties that have emerged from the use of these environmental risk assessment models. The proposed classification tool is similar in concept to those already used in medicine for clinical diagnoses of a wide variety of illnesses from appendicitis to myocardial infarction (Baxt, 1995). Classification models have also been used successfully in marketing and financial contexts to sort people into various consumer niches or credit risk groups; in security contexts for fingerprint, speech, and face recognition; and in weather forecasting. Although the recommended classification model approach is innovative and indicates a willingness to adopt techniques successfully employed elsewhere, the committee cautions that it may run the risk of being viewed by the public in a “conspiratorial” fashion. For example, the public may wonder whether such a modeling approach is merely “a vehicle to prove what we think we already know” or whether it

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration represents “an honest attempt to find answers that are not predetermined” (Oreskes, 1998). In addition, the public could perceive the process as subject to manipulation to achieve or support results wanted in the first place. Finally, the process could be perceived as simply a means to remove accountability from decision-makers by placing it on an “objective” modeling process (Yearley, 1999). As noted previously, to gain widespread credibility and acceptance, the CCL decision-making process must be transparent. It is important to note, however, that transparency is not necessarily synonymous with simplicity. A process that appears on the surface to be simple and easy to understand may in fact be riddled with hidden assumptions. In contrast, it is possible for a CCL decision-making process that involves complex classification modeling to be made relatively transparent. This will require that transparency be integral to its design, and EPA must be ready and able to support each step in the process as if the process itself were on trial. First, the use of a classification tool needs justification. The successful experience of using such tools in other contexts and applications would help make the case for its use in sorting various potential drinking water contaminants. In addition, referring to the use of the tool in contexts that are familiar in the day-to-day experience of the public (e.g., marketing and credit profiling) could enhance public understanding of the process (Dahinden et al., 1999). Second, the methodology for designing and calibrating the decision-making process must be defended. This important issue and related considerations are discussed in depth in Chapter 5. To this end, the committee recommends that EPA make clear that the design of the process was in part based on a “retrospective” approach whereby EPA starts with a scenario it wants to end up with (e.g., a scenario that builds on past correct EPA decisions to regulate or not to regulate certain chemicals) and then calibrates the model based on this scenario. The results of citizen focus groups conducted in Europe (for the ULYSSES Project) indicated that this retrospective type of approach to model development would be acceptable to the public (Dahinden et al., 1999). However, the choice of a scenario upon which to calibrate the model may not be a trivial, non-controversial one and therefore may have to be vigorously defended. In addition, disputes may occur (e.g., among decision-makers and between decision-makers and the public) regarding the issue of what minimum levels of sensitivity and specificity the model must have in order to be judged “adequate.” Some justification must also be made for the selection of the key parameters of the model (i.e., measures or attributes of

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration potential health effects and occurrence), as well as a discussion of the magnitude of the effect each chosen parameter has on the model’s results and the quality of data for each parameter (e.g., uncertainties, data gaps, timeliness of the data). The choice of parameters must not be based simply on the fact that quantitative data are available for them, in other words: “There is a human tendency to count what is easy to count, and then, as a kind of ‘doublethink,’ to mistake what is counted for what counts” (Hattis and Anderson, 1999). Third, if decision-making for including or excluding certain contaminants on future CCLs will ultimately depend on a combination of EPA judgment and the results of a classification tool (as recommended in this report), then this relationship must be fully articulated along with the background assumptions underlying agency judgments. Furthermore, EPA will have to justify how the consistency and explicitness of its decisions can be maintained in these situations. Other related issues, such as to what extent political (including budgetary) contingencies will affect EPA’s decisions in conjunction with scientific and efficiency considerations (e.g., regarding the size of future CCLs), will also have to be fully aired. Fourth, it can be expected that some decisions to include or exclude specific contaminants from the CCL will be controversial. For example, EPA received and responded on a contaminant-by-contaminant basis to public comments that opposed the inclusion of several contaminants and groups of related contaminants (e.g., aluminum, organotins) on the draft 1998 CCL (EPA, 1998a,c). Thus, to help ensure transparency and legitimacy in the development of future CCLs, any key criteria, data, or assumptions that ultimately made the difference between inclusion or exclusion in such cases must be clearly identified and carefully justified. Although a classification model approach is recommended for the creation of future CCLs, the committee notes it may also be worthwhile to use the tool to aid thinking on the issue of contaminant selection for regulation and research. For example, if the use of the tool produces unexpected results, it will force decision-makers to review the assumptions, parameter choices, and uncertainties incorporated in the model as well as how the tool works (Ravetz, 1999). Conducting a review of this kind might advance knowledge on the science of decision-making as well as the science involved in the regulation of drinking water contaminants. In addition, if such a review were discussed fully in a public forum, the transparency of the model would also be enhanced (i.e., it would look less like a “black box” mysteriously churning out results).

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration Two additional ways to enhance the transparency of the classification model would be (1) to remove or add parameters to determine how the selection of contaminants changes (a discussion of such results would also serve to improve the transparency of the process) and (2) to make the software and databases used to design, develop, and implement the model available to the public so that citizens can attempt to go through or recreate the process themselves. However, the committee realizes that the latter may be impractical. To conclude this discussion, Hattis and Anderson (1999) provide an appropriate quote summarizing the importance of transparency in decision-making: …We quantifiers will only succeed in being helpful to democratic decision-making if our work is designed and presented in such a way that it helps affirm, rather than supplant, the decision-making autonomy of our audience. Public Participation “Public participation encompasses a group of procedures designed to consult, involve, and inform the public to allow those affected by a decision to have an input into that decision” (Rowe and Frewer, 2000). Furthermore, Renn and colleagues (1993) noted that the central tenet of public participation is that the public is, in principle, capable of making wise and prudent decisions. The 1996 NRC report Understanding Risk: Informing Decisions in a Democratic Society adopts this central tenet and is critical of agencies that rely on a “decide-announce-defend” strategy that involves the public only after the deliberation process is over (NRC, 1996). However, the report uses the term “deliberation” rather than “participation” to emphasize the need for substantive public involvement throughout the decision-making process (Chess, 2000; NRC 1996). The report discusses at length the importance of “getting the right participation” (i.e., sufficiently broad participation that includes the range of interested and affected parties) and “getting the participation right” (e.g., incorporating public values, viewpoints, and preferences into the process). Public participation procedures vary from one-way flows of information (e.g., surveys, focus groups, public comment), where the aim is to elicit public opinions, to collaborative forms of decision-making such as negotiated rule making (Fiorino, 1990; Laird, 1993), consensus confer-

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration ences (Einsiedel and Eastlick, 2000), citizen juries (Lenaghan et al., 1996), and citizen panels (Renn et al., 1993), where the aim is to elicit decisions and judgments that will form the basis for actual policies (Beierle, 1998). A brief description of several public participation procedures and evaluative criteria to assess their strengths and weaknesses has been proposed by Rowe and Frewer (2000). Not surprisingly, there is no procedure that is preferred in all situations (NRC, 1996). However, an effective public participation procedure might entail a hybrid or combination of procedures such as the three-step procedure proposed by Renn et al. (1993). This approach attempts to integrate (1) technical expertise, (2) values and concerns of stakeholders, and (3) common sense and personal experience of the general public to balance the legitimate demand for public participation with the needs for a scientifically sound policy and agency accountability. In a given situation, the most effective procedure will depend, among other factors, on the aims and rationale for participation. Regarding the development of future CCLs, at least three rationales can be used to justify increased public participation: political, normative, and epistemic (Perhac, 1998). Each rationale leads to a different definition of who should participate on behalf of the “public,” and this in turn affects the choice of participation procedure. The political rationale is that public participation enhances the political viability, legitimacy, and transparency of the process as well as the credibility of the regulators. It is the recognition that even the most scientifically sound process will be difficult to implement if it is perceived by the public as unfair or biased. Using this rationale, the “public” would be defined as those people whose acceptance is crucial for the viability and legitimacy of the process, such as representatives of specific stakeholder groups. For example, negotiated rule making is a public participation procedure that corresponds to this definition of public. The normative rationale appeals to democratic principles and holds that the public has a right of involvement since it is the “owner” of publicly funded regulatory policies and the most appropriate source of the value judgments that are necessary in any decision-making process. This rationale assumes that all are affected by regulatory policies and therefore all have the right to participate, not just representatives of stakeholder groups. Under this rationale, a public participation procedure should ensure that all citizens’ values and preferences are fairly represented—regardless of whether these citizens are organized—and that the output of the procedure (e.g., decisions or recommendations) has a

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration genuine impact on policy. Public participation procedures that follow from this rationale would include referenda, public opinion surveys, focus groups, consensus conferences, and citizen juries. Lastly, the epistemic rationale recognizes that the public possesses important factual knowledge (e.g., local knowledge) as well as special insight on societal values and therefore public participation will result in better decisions. Under this rationale, the “public” might be defined as those people who possess special insight on values or unique factual knowledge that is relevant to a given decision-making process. A procedure that follows from this definition of the public would be citizen advisory committees. General recommendations to facilitate public participation in environmental programs are provided in The Model Plan for Public Participation, developed for the EPA by the Public Participation and Accountability Subcommittee of the National Environmental Justice Advisory Council (EPA, 1996c). Hampton (1999) provides recommendations to ensure that public participation procedures satisfy the criteria of equity, fairness, and justice. Among these are the following: The public should be involved in defining the process of participation. Public involvement should start early in the process (e.g., at the time of agenda setting or when value judgments become important to the process). Participants should have access to appropriate resources such as the information necessary to participate fully in the process, access to scientists, technical assistance, and sufficient time to prepare for the deliberations. Prior agreement should be reached with the participants as to how the output of the procedure (e.g., recommendations, decisions) will be used and how it will affect agency policy decisions. The NRC Committee on Risk Characterization concluded in its 1996 report, Understanding Risk: Informing Decisions in a Democratic Society, that it was not possible to predict which public participation procedure would work most effectively in any given situation, and this committee concurs. Each procedure has its advantages and limitations, but a successful outcome will usually depend less on the inherent aspects of a procedure and more on other factors such as the history of an issue, the level of conflict, the level of public trust in the agency, the agency’s

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration intentions and expectations for the selected procedure, and how well the agency implements the procedure (Chess, 2000; NRC, 1996). SUMMARY: CONCLUSIONS AND RECOMMENDATIONS The committee recognizes that the development of a PCCL from the universe of potential drinking water contaminants, as well as a contaminant’s movement from a PCCL to the corresponding CCL, is a complex task requiring numerous difficult classification judgments in a context where data are often uncertain or missing. The evaluation of contaminant attributes, such as severity of health effects, potency, and occurrence in drinking water (see Chapter 4), will often entail making assumptions because of data gaps and uncertainties. Moreover, evaluation of the severity of health effect and making comparisons of severity among different health effects (e.g., cancers versus impotence) will depend on explicit and implicit value judgments as well as on the choice of reasoning scheme (e.g., utilitarian, quantitative). Because of this complexity, the committee believes that to be scientifically sound as well as publicly acceptable, the process for developing future CCLs must depart considerably from the process used to develop the first (1998) CCL. The committee recommends that the process for selecting contaminants for future CCLs be systematic, scientifically sound, and transparent. The development and implementation of the process should involve sufficiently broad public participation. The ultimate goal of the contaminant selection process is the protection of public health through the provision of safe drinking water to all consumers. To meet this goal, the selection process must place high priority on the protection of vulnerable subpopulations. More specifically, the committee makes the following recommendations: The definition of vulnerable subpopulations not only should comply with the amended language of the SDWA, but also should be sufficiently broad to protect public health; in particular, EPA should consider including (in addition to those subgroups mentioned as examples in the amended SDWA) all women of childbearing age, fetuses, the immuno-compromised, people with an acquired or inherited genetic disposition that makes them more vulnerable to drinking water contami-

OCR for page 45
Classifying Drinking Water Contaminants for Regulatory Consideration nants, people who are exceptionally sensitive to an array of chemical contaminants, people with specific medical conditions that make them more susceptible, people with poor nutrition, and people experiencing socioeconomic hardships and racial or ethnic discrimination. Transparency should be incorporated into the design and development of the classification and decision-making process for future CCLs in addition to being an integral component in communicating the details of the process to the public. Otherwise, the public may perceive the process as subject to manipulation to achieve or support desired results. Therefore, sufficient information should be provided such that citizens can place themselves in a position similar to decision-makers and arrive at their own reasonable and informed judgments. This may require making available to the public the software and databases used in the process. The central tenet that the public is, in principle, capable of making wise and prudent decisions should be recognized and reflected in the choice of public participation procedures used to help create future CCLs. A “decide-announce-defend” strategy that involves the public only after the deliberation process is over is not acceptable. Substantive public involvement should occur throughout the design and implementation of the process. EPA should strive to “get the right participation” (i.e., sufficiently broad participation that includes the range of interested and affected parties) as well as to “get the participation right” (e.g., incorporating public values, viewpoints, and preferences into the process).