Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 Evolution and Use of Risk Assessment in the Environmental Protection Agency: Current Practice and Future Prospects Overview EPA risk-assessment concepts, principles, and practices are products of many diverse factors, and each agency program is based on a âunique mixture of statutes, precedents, and stakeholdersâ (EPA 2004a, p. 14). With respect to statutes, Congress established the basic plan through a series of environmental laws, most enacted during the 1970s and most authorizing science-based regulatory action to protect public health and the environment. Another factor is EPAâs case-by-case experience with implementing these laws and the result- ing supplementary principles and practices. Equally important, advisory bodies have drawn on the expertise of scientists and other environmental professionals in universities, private organizations, and other government agencies to recommend corrections and improvements. The net result is that risk assessment in EPA is a continually evolving process that has a stable common core but takes several forms. This chapter traces the origins and evolution of risk assessment in EPA with an emphasis on current processes and procedures as a stepping-off point for the future improvements envisioned in later chapters. This chapter first describes the diverse statutory requirements that have led to a broad array of agency programs with correspondingly varied approaches to risk assessment; it then highlights current concepts and practices, outlines EPAâs multifaceted institutional arrangements for managing the process, and identifies extramural influences. The record shows that EPA continually updates the process with new scientific information and policies, often in response to new laws or advice from advisory bodies as to general principles or individual assessments. Not all external recommendations necessarily warrant agency action, but it is clear that implementation of some recommendations has been in- complete. The chapter closes with process recommendations for implementing some of the substantive recommendations in the chapters that follow. 26
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 27 STATUTORY PLAN AND REGULATORY STRUCTURE The environmental laws enacted by Congress shape EPAâs regulatory structure, which, in turn, influence EPA risk-assessment practices and perspectives. The statutes give EPA authority to regulate many forms of pollution (for example, pesticides, solid wastes, and industrial chemicals) as they affect different aspects of the environment (for example, air quality, water quality, human health, and plant and animal wildlife). The premise central to EPA risk-assessment practices can be found in enabling legislation for its four major pro- gram offices: air and radiation, water, solid waste and emergency response, and prevention, pesticides, and toxic substances. Selected provisions appear below. â¢ The Clean Water Act calls for standards âadequate to protect public health and the environment from any reasonably anticipated adverse effectsâ (CWA Â§ 405 (d)(2)(D)). â¢ The Clean Air Act, when addressing criteria pollutants, directs the agency to de- velop criteria âreflecting the latest scientific knowledgeâ and, on the basis of those criteria, to issue ânational primary ambient air quality standards to . . . protect public health with an adequate margin of safetyâ (CAA Â§Â§ 108,109). â¢ The primary purpose of the Toxic Substances Control Act is âto assure [that tech- nologic] innovation and commerce in such chemical substances and mixtures do not present an unreasonable risk of injury to health or the environmentâ (TSCA Â§ 2 (b)(3)). â¢ Under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), one criterion for registering (licensing) a pesticide is that âit will perform its intended function without unreasonable adverse effects on human health and the environmentâ (FIFRA Â§ 3). â¢ The Superfund National Contingency Plan specifies that âcriteria and priorities [for responding to releases of hazardous substances] shall be based upon relative risk or danger to public health or welfare or the environmentâ (CERCLA Â§ 105 (a)(8)(A)). The term risk assessment does not appear often in the statutes, and it is important to note that these statues were enacted prior to the emergence of risk analysis as an integra- tive discipline in the late 1970s and early 1980s. Rather, EPA risk-assessment principles and practices stem from statutory provisions calling for information on âadverse effectsâ (EPA 2004a, p. 14), ârelative riskâ (p. 82), âunreasonable riskâ (p. 14), and âthe current scien- tific knowledgeâ (p. 104) and for regulatory decisions on protecting human health and the environment. The statutes provide various standards and procedures related to the scientific analyses used to evaluate the risk potential of pollutants subject to the statutes., â ifferent D emphases and terminology lead to different risk-assessment approaches, sometimes for the same pollut- ant, in different agency programs. That can confuse and confound observers. For example, Clean Air Act provisions related to four air-pollution topics use different terms for what is essentially the same statutory finding: â¢ Clean Air Act provisions related to pollutants regulated as national ambient air quality standards are de- signed to âprotect the public health with an adequate margin of safetyâ (CAA Â§ 109, emphasis added). â¢ For welfare (environmental) effects, this provision directs the office to âprotect the public welfare from any known or anticipated adverse effectsâ (CAA Â§ 109, emphasis added). â¢ Standards for âhazardousâ pollutants from stationary sources (for example, factories) are to âprovide an ample margin of safety to protect public health or prevent an adverse environmental effectâ (CAA Â§ 112, emphasis added). â¢ Regarding mobile sources (for example, cars), the statute calls for ensuring that these vehicles do not âcause or contribute to an unreasonable risk to public health, welfare or safetyâ (CAA Â§ 202 (a)(4), emphasis added). â ome statutes call for technology-based standards that require, for example, specific control techniques or S technology-forcing standards that specify emission limits to be achieved within given periods. Such standards are based on costs, engineering feasibility, and related technical considerations. Examples include Clean Air Act Sec- tions 111 (new-source review) and 202 (mobile-source emissions).
28 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT The existence of several medium-oriented statutes explains why EPA has multiple risk- assessment programs. This circumstance often draws criticism as âstovepipingâ that leads to delay and inconsistency in both risk assessment and regulation. In the early 1990s, Congress considered but did not pass legislation to incorporate common risk-assessment terminology, concepts, and requirements into comprehensive risk-assessment legislation. Instead, recent enactments are notable for precise terms that amplify and clarify legislative objectives in individual statutes by specifying elements that assessments subject to particular statutes must include â¢ The 1996 Food Quality Protection Act specifies that âin the case of threshold ef- fects . . . an additional ten-fold margin of safety for the pesticide chemical residues shall be applied for infants and childrenâ (FFDCA Â§ 408 (b)(2)(C)). â¢ 1996 amendments to the Safe Drinking Water Act are similarly explicit about the presentation of risk estimates and uncertainty: âThe Administrator shall, in a document made available to the public in support of a regulation promulgated under this section, specify, to the extent practicable â Each population addressed by any estimate of public health effects â The expected risk or central estimate of risk for the specific populations â Each appropriate upper-bound or lower-bound estimate of riskâ (SDWA Â§ 300g-1 (b)(3)). Provisions like those that apply to individual programs (the examples above appear in pesticide and water legislation, respectively) account for some of the variation in risk- assessment practices and results. However, although the new terms apply directly only to the program governed by the statute, other programs have adopted some of the changes. Despite differences in statutory language, environmental media, and pollutants, several factors common to the major statutes continue to shape EPAâs regulatory structure and func- tion and its perspectives on risk assessment: â¢ The emphasis in each statute on protecting human health and the environment pro- vides the basis of EPAâs purported conservative approach to risk assessment. Examples range from generic âadequate margin of safetyâ language in the Clean Air Act (CAA) amendments of 1971 (Â§ 109) to the required additional safety factor of 10 for protection for infants and children in the 1996 Food Quality Protection Act (FQPA; FFDCA Â§ 408 (b)(2)(C)). As ex- plained recently, âconsistent with its mission, EPA risk assessments tend towards protecting public and environmental health by preferring an approach that does not underestimate risk in the face of uncertainty and variabilityâ (EPA 2004a, p. 11). â¢ Except as noted above (footnote 2) and later in this chapter (page 51), the statutory provisions related to EPAâs main standards for protecting human health and the environment treat scientific analysis as a central element in regulatory decision-making and call for collec- tion and evaluation of scientific information related to the pollutant undergoing regulatory review. Statutes often detail the kinds of information, analyses, and formal documentation required in the rule-making record. â bipartisan coalition of senators sponsored the Thompson-Levin bill (S981), titled âRegulatory Improvement A Bill,â which would have codified the Office of Management and Budget (OMB) role in review of agency regula- tions; some provisions later appeared in the OMB Bulletin (70 Fed. Reg. 2664 ). The Moynihan bill (S123) called for comparative risk assessment.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 29 â¢ Although some sections of statutes focus solely on health-effect considerations, many also identify information and analyses from other fieldsâsuch as economic analysis, technical feasibility, and societal impactsâfor use in making regulatory decisions. âIt is generally recognizedâby the science community, by the regulatory community, and by the courtsâthat it is important to consider other factors along with the science when making decisions about risk managementâ (EPA 2004a, p. 3). The resulting decisionsâwhether or not to regulate and, if so, the nature and form of regulationâseek to protect human health and the environment where appropriate, in part on the basis of scientific analysis and in part on the basis of consideration of information on costs, societal values, legal requirements, and other factors. As the proponent of any new regulation, EPA generally has the burden of proving that the proposed regulation meets statutory standards. That is not a requirement for EPA to prove âcause and effectâ in the customary scientific sense, but rather to demonstrate by way of science-based analysis that the proposed regulation meets statutory criteria related to adverse effects, unreasonable risks, and other statutory thresholds for regulation: Although regulatory agencies do not have the technical burden of proving that a particular companyâs products or activities have caused or will cause a particular personâs disease, they do have the practical burden of assembling a record containing sufficient scientific informa- tion and analysis to survive a reviewing courtâs âhard lookâ review under the âsubstantial evidenceâ or âarbitrary and capriciousâ tests for judicial review of administrative action [McGarity 2004]. The environmental statutes administered by EPA and general administrative law re- quire documentation and review of relevant data and analyses. Some statutory provisions for pesticides facilitate gathering data for risk assessment by enabling the agency to impose data requirements on producers and others (for example, FIFRA Â§ 3); the agencyâs ability to impose data requirements has proved far more limited under the Toxic Substances Control Act (TSCA; GAO 2005) and other statutes. As the primary scientific rationale for many EPA regulations, risk assessment is subject to scientific, political, and public controversy. Building on the statutory foundation, the 1983 Red Book introduced principles, terminology, and practices that have become mainstays of the process. That report, which provided for a common framework for reconciling, to some extent, the differing requirements of the statutes, led to changes in the 1980s and 1990s and continues to shape the process today. The Pivotal Role of the Red Book The 1983 National Research Council Report During the 1970s, the scientific assessment practices of EPA and other federal agencies faced with similar responsibilitiesâthe Occupational Safety and Health Administration, the Food and Drug Administration (FDA), and the Consumer Products Safety Commis- â ection S 109 of the CAA of 1970 is the most often cited example; note, however, that the statute expressly provides for consideration of costs, feasibility, and other factors in state implementation plans (Â§ 110). Such con- siderations influence the time allowed for compliance with the standards. â The situation differs for pesticides. The pesticide statute, FIFRA, requires manufacturers to submit data show- ing a âreasonable certainty of no harmâ before pesticides can be registered and marketed and to maintain the registration.
30 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT sionâcame under close scrutiny as decisions resulting from those practices took on greater social importance. In 1981, Congress (PL-96528) directed that FDA support a National Research Council study of the âmerits of an institutional separation of the scientific func- tions of developing objective risk assessments from the regulatory process of making public and social policy decisions and the feasibility of unifying risk assessment functions.â The National Research Council organized the Committee on the Institutional Means for Assess- ment of Risks to Public Health in October 1981, and the committeeâs report, the Red Book, was issued on March 1, 1983. In his letter transmitting the report to the commissioner of FDA, the chairman of the National Research Council, Frank Press, stated, The Congress made provision for this study to strengthen the reliability and objectivity of scientific assessment that forms the basis for federal regulatory policies applicable to car- cinogens and other public health hazards. Federal agencies that perform risk assessments are often hard pressed to clearly and convincingly present the scientific basis for their regulatory decision. In the recent past, for example, decisions on saccharin, nitrites in food, formalde- hyde use in home insulation, asbestos, air pollutants and a host of other substances have been called into question. The report recommends no radical changes in the organizational arrangements for perform- ing risk assessments. Rather, the committee finds that the basic problem in risk assessment is the incompleteness of data, a problem not remedied by changing the organizational arrange- ment for performance of the assessments. Instead, the committee has suggested a course of action to improve the process within the practical constraints that exist. As noted in Pressâs letter, the âcourse of actionâ recommended by the committee focused primarily on the process through which complex and uncertain, and often contradictory, scientific information derived from laboratory and other types of research could be made useful for regulatory and public-health decision-making. The committee was also sensitive to the concern, expressed in the congressional language, that scientific assessments should be âobjectiveâ and free of policy (and political) influences. Because all assessments of scientific data are subject to uncertainties and because scientific knowledge is incomplete, it is possible for different analysts to arrive at different interpretations of the same set of data. If the as- sessment involves risks to human health from chemical toxicity or other types of hazards, the differences in interpretation can be large. The committee therefore recognized that risk assessments could be easily manipulated to achieve some predetermined risk-management (policy) outcome. Much of the work of the committee was directed at finding ways to mini- mize that potential problem while avoiding the undesirable step of institutional separation of scientific assessment from decision-making. The 1983 report was not directed at the technical analyses involved in risk assessment. Rather, it offered a coherent and generally applicable framework within which the process of risk assessment could be undertaken. That framework was shown to be necessary to fill the gap between the research setting within which general scientific knowledge and diverse types of information on specific threats to human health are developed and the various types of risk-management activities undertaken by regulatory and public-health agencies to minimize those threats. The committeeâs recommendations gave order to the developing field of risk assessment by defining terms and elucidating the four (now well-known) steps of the risk-assessment process. The committee chose the term risk characterization to describe the fourth and final step of the risk-assessment process, in which there is an integration and synthesis of the information and analysis contained in the first three steps (see Figure 2-1). The committee stated that the term characterization was chosen to convey the idea that both quantitative and qualitative elements of the risk analysis, and of the scientific uncertainties
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 31 RESEARCH RISK ASSESSMENT RISK MANAGEMENT Laboratory and field Hazard Identification Development of observations of (Does the agent cause regulatory options adverse health effects the adverse effect?) and exposures to particular agents Evaluation of public health, economic, social, political Information on Dose-Response Assessment consequences of extrapolation methods (What is the relationship regulatory options for high to low dose between dose and incidence and animal to human in humans?) Risk Characterization (What is the estimated incidence of the adverse effect in a given population?) Field measurements, Exposure Assessment estimated exposures, (What exposures are characterization of currently experienced or populations anticipated under different Agency decisions conditions?) and actions FIGURE 2-1â The National Research Council risk-assessmentârisk-management paradigm. Source: NRC 1983. Figure 2-1.eps in it, should be fully captured for the risk manager. Risks associated with chemical toxicity necessarily involve biologic data and uncertainties, many of which are not readily expressed in quantitative terms. Again, it was beyond the charge of the committee to offer specific technical guidance on the modes of scientific analysis appropriate for each of the steps of risk assessment. The first recommendation of the Red Book is the following (NRC 1983, p. 7): We recommend that regulatory agencies take steps to establish and maintain a clear concep- tual distinction between assessment of risks and consideration of risk management alterna- tives; that is, the scientific findings and policy judgments embodied in risk assessments should be explicitly distinguished from the political, economic, and technical considerations that influence the design and choice of regulatory strategies. Two aspects of that critical recommendation are especially noteworthy. First, the com- mittee emphasized that the distinction between risk assessment and risk management is a conceptual one; that is, it concerns the fact that the content and goals of the two activities are distinguishable on a conceptual level. The Red Book nowhere calls for any other type of âseparationâ of the two activities. Second, the phrase âpolicy judgments embodied in risk assessmentâ (which are said to be different in kind from those involved in risk management) points to one of the most important insights of the committee. In particular, the committee recognized that almost no risk assessment can be completed unless scientific information (data and knowledge) is supplemented with assumptions that have not been documented in relation to the particular risk assessment at hand, although they have probably been supported by substantial evidence
32 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT or theory for the general case. The clearest examples of such assumptions related to risks posed by chemical toxicity concern the shape of dose-response curves in the region of very low doses and the relevance to humans of various toxicity responses observed in high-dose animal experiments; assumptions regarding these and many other aspects of the data used for risk assessment are necessary to provide risk managers useful risk characterizations based on consistent approaches. The Red Book committee recognized that for a given analytic component of any of the steps of a risk assessment for which an assumption is necessary, several scientifically plau- sible assumptions might be available. The committee used the phrase âinference optionsâ to describe the array of possibilities. To bring order and consistency to risk assessments conducted by the federal government and to minimize case-by-case manipulations of risk- assessment outcomes, the committee recommended the development of specific âinference guidelinesâ; these were to contain âan explicit statement of a predetermined choice among alternative inference optionsâ (NRC 1983, p. 4) (see Box 2-1). Thus, agencies should take steps to describe, in explicit guidelines, the technical approaches used to conduct risk as- sessments, and these guidelines should include specification of the assumptions (including, in some cases, models) that would be consistently used to draw inferences in all the analytic components of the risk-assessment process where they are needed. Inference options have come to be called default options, and the inferences selected for risk assessments have come to be called defaults. The development and consistent use of technical guidelines for risk assessment, with the specification of all the necessary defaults, were seen by the Red Book committee as necessary to avoid the institutional separation of scientific assessment from policy development and implementation while minimizing inappropriate and sometimes invisible policy influences on the risk-assessment process. As noted later in this chapter, some critics of the Red Book have raised the concern that the committeeâs commendable effort to avoid âinappropriate influencesâ can readily be taken to mean âno influenceâ from risk managers and other stakeholders. One additional feature of the Red Bookâs recommendations bears on the current commit- teeâs task. Thus, as part of the statement of Recommendation 6, which concerns the criteria for useful risk-assessment guidelines, can be found the following (NRC 1983, p. 165): Flexibility The committee espouses flexible guidelines. Rigid guidelines, which permit no variation, might preclude the consideration of relevant scientific information peculiar to a particular chemical and thus force assessors to use inference options that are not appropriate in a given case. Also, rigid guidelines might mandate the continued use of concepts that become ob- solete with new scientific developments. Large segments of the scientific community would undoubtedly object to such guidelines as incompatible with the use of the best scientific judgment for policy decisions. Flexibility can be introduced by the incorporation of default options. The assessor would be instructed to use a designated (default) option unless specific scientific evidence suggested otherwise. The guidelines would thus permit exceptions to the general case, as long as each exception could be justified scientifically. Such justifications would be reviewed by the sci- â o scientific knowledge is without uncertainty, but it is generally subject to empirical verification; when the N empirical evidence is supportive and no contrary evidence can be found, documentation is said to have been es- tablished, at least tentatively. The assumptions needed to complete risk assessments are generally well supported for the relevant set of past assessments; however, in any specific case it will often be difficult, if not impossible, to verify empirically that a given assumption also holds for the substance at issue.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 33 BOX 2-1 Agencywidea Risk-Assessment Guidelines 1986 Guidelines for Carcinogen Risk Assessment (EPA 1986a) Guidelines for Health Assessment of Suspect Developmental Toxicants (51 Fed. Reg. 34028 ) Guidelines for Mutagenicity Risk Assessment (EPA 1986b) Guidelines for Estimating Exposures (51 Fed. Reg. 34042 ) Guidelines for Health Assessment of Chemical Mixtures (EPA 1986c) 1991 Developmental Toxicity Risk Assessment (revised and updated) (EPA 1991) 1992 Guidelines for Exposure Assessment (EPA 1992a) 1996 Guidelines for Reproductive Toxicity Risk Assessment (EPA 1996a) 1998 Guidelines for Ecological Risk Assessment (EPA 1998a) Guidelines for Neurotoxicity Risk Assessment (EPA 1998b) 2000 Supplementary Guidance for Health Risk Assessment of Chemical Mixtures (EPA 2000a) 2005 Guidelines for Carcinogen Risk Assessment and Supplemental Guidance for Assessing Sus- ceptibility from Early-Life Exposure to Carcinogens (EPA 2005a,b) These guidelines, which are consistent with Red Book recommendations (NRC 1983, p. 7), âstructure the interpretation of scientific and technical information relevant to the assessmentâ and âaddress all elements of risk assessment, but allow flexibility to consider unique scientific evidence in particular instances.â Each guideline is a multiyear project developed by multioffice teams composed of scientists in EPA laboratories, centers, program offices, and regional offices. Draft guidelines are peer-reviewed in open public meetings and published for comment in the Federal Register. In general, each guideline follows the 1983 Red Book paradigm, providing guidance on the use and interpretation of information in each field of analysis, including the role of defaults and assumptions and approaches to uncertain- ties and risk characterization. Some guidelines are accompanied by supplementary reports on special topics, for example, âAssessing Susceptibility from Early-life Exposure to Carcinogensâ (EPA 2005b) and âGuiding Principles for Monte Carlo Analysisâ (EPA 1997a). aEPAâs guideline library includes many other guidance documents and policies, including those specific to individual programs (see, for example, Tables C-1 and D-1 and references). entific review panels and by the public under procedures described above. Guidelines could profitably highlight subjects undergoing relatively rapid scientific development (for example, the use of metabolic data for interspecies comparisons) and any other components in which exceptions to particular default options were likely to arise. They should also attempt to present criteria for evaluating whether an exception is justified. As will be evident throughout this report, it has proved difficult to achieve scientific consensus on judgments regarding the adequacy of scientific evidence to justify, in specific cases, departures from one or more defaults. One of the objectives of the present committeeâs work might be seen as determining whether 25 years of scientific research and of scholarly thinking about the conduct of risk assessments provides new insights into whether there might be better ways of approaching the uncertainties that give rise to the need for defaults.
34 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT Later National Research Council Studies NRC (1993a) advocated the integration of ecological risk assessment into the 1983 Red Book framework. The framework for risk assessment and its four-step analytic process were adopted and promoted in the National Research Councilâs Science and Judgment in Risk As- sessment (NRC 1994) and Understanding Risk: Informing Decisions in a Democratic Society (NRC 1996). Indeed, the framework has been widely adopted in other expert studies of risk assessment (see PCCRARM 1997 and references cited therein) and has been adopted outside the United States (in the European Union and the World Health Organization) (see Figure 2-2). Moreover, as regulatory and public-health institutions have had to bring a greater degree of scientific analysis and consistency to health threats posed by microbial pathogens (Parkin 2007), excessive nutrient intakes (IOM 1997, 1998, 2003; WHO 2006), and other environmental stressors, they have found the Red Book framework both scientifically ap- propriate and useful. One additional theme regarding the risk-assessment process is given great attention by the National Research Council in Understanding Risk (NRC 1996, p. 6): The analytic-deliberative process leading to a risk characterization should include early and explicit attention to problem formulation; representation of the spectrum of interested and INTEGRATED RISK ASSESSMENT STAKEHOLDER PARTICIPATION Problem Formulation With Hazard Identification RISK MANAGEMENT Analysis Characterization Characterization Dose- Exposure of of Response Assessment Exposure Effects Assessment Risk Characterization FIGURE 2-2â The World Health Organizationâs framework for integrated health and ecologic risk as- Figure 2-2.eps sessment. NOTE: Figures 2-1 and 2-2 show different renditions and evolving emphases as to the basic elements of the Red Book paradigm. Source: Suter et al. 2001.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 35 affected parties at this early stage is imperative. The analytic-deliberative process should be mutual and recursive. Analysis and deliberation are complementary and must be integrated throughout the process leading to risk characterization: deliberation frames analysis, analysis informs deliberation, and the process benefits from feedback between the two. That recommendation provides nuance to the Red Bookâs call for âseparationâ of as- sessment and management to facilitate the supreme goal of risk assessment: to provide the scientific basis for public-health and regulatory decisions. As long as âanalysis and delibera- tionâ does not involve efforts by risk managers to shape risk-assessment outcomes to match their policy preferences, but rather involves efforts to ensure that assessments (whatever their outcomes) will be adequate for decision-making, interactive processes involving âthe spectrum of interested and affected partiesâ are seen as imperative. The 1994 National Research Council report Science and Judgment in Risk Assessment evaluated EPAâs risk-assessment practices as they apply to hazardous air pollutants from sources subject to Section 112 of the CAA amendments of 1990. That report did not alter the principles for risk assessment set forth by the Red Book but rather examined EPA guidelines and practices and then recommended ways in which various technical improvements in the conduct of risk assessments and in the presentation of risk characterizations might be ac- complished. Thus, the present committeeâs efforts resemble in many ways those undertaken by the Science and Judgment committee. The issue of default options was given much consideration (see Box 2-2). Indeed, the 1994 National Research Council committee found EPAâs existing technical guidelines for risk assessment to be deficient with respect to their justifications for defaults and with respect to evidentiary standards and scientific criteria to be met for case-specific departures from them. The committee offered a long series of recommendations, each preceded by a discussion of the state of technical understanding, on issues of data needs for risk assessment, uncertainty, variability, aggregation of exposures and risk, and model development. The 1994 committeeâs recommendations extended beyond the technical content of risk assessment and included issues of process, institutional arrangements, and even problems of risk communication. Although there was much focus on air-pollutant risks, particularly the technical issues related to exposure assessment, most of that committeeâs recommendations had broad applicability to risk assessment. In Appendix D to the present report, the committee has selected representative rec- ommendations contained in the three National Research Council reports cited above and attempted to provide a view of how EPA has responded to many of them. It can be seen that EPA has devoted considerable effort to ensuring that its guidelines conform to many National Research Council recommendations, although the record on accepting and imple- menting recommendations is uneven and incomplete (see, for example, Boxes 2-4 and 2-5 and Chapter 6). The present committee has been asked to review current EPA âconcepts and practices,â taking into account the previous National Research Council studies and studies in which new scientific approaches are being evaluated. The present committee is not specifically charged with modifying the fundamental concepts first elucidated in the Red Book unless the scientific understanding on environmental hazards and the research on the conduct of risk assessment that have developed over the past 25 years demand such a modification. Thus, as â Appendix N to the 1994 report contains two views of the issue of defaults, one of committee member Adam Finkel and one of members Roger McClellan and D. Warner North; their papers represent a range of committee perspectives on the appropriate balance of science and policy considerations in a system for departure from default assumptions.
36 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT BOX 2-2 Science Policy and Defaults Science and Judgment (NRC 1994) describes defaults as the âscience policy components of risk assessmentâ (p. 40) and points out that âif the choice of inference options is not governed by guidelines, the written assessment itself should make explicit the assumptions used to interpret data or support conclusions reached in the absence of dataâ (p. 15). The report recognizes âchoiceâ as an aspect of science policy (p. 27): The [1983 Red Book] committee pointed out that selection of a particular approach under such circumstances involves what it called a science-policy choice. Science-policy choices are distinct from the policy choices associated with ultimate decision-making. . . . The science-policy choices that regulatory agencies make in carrying out risk assessments have considerable influence on the results. Those principles are the basis of EPAâs call for âtransparency,â âfull disclosure,â and âscientific con- clusions identified separately from default assumptions and policy callsâ in the Risk Characterization Handbook (EPA 2000b). EPAâs recent Staff Paper (EPA 2004a, p. 12) embraces and expands on the principles: âScience policy positions and choices are by necessity utilized during the risk assessment process.â The Superfund programâs supplemental guidance document Standard Default Exposure Factors was developed in response to requests to make Superfund assessments more transparent and their assumptions more consistent. The guidance states that defaults are used when âthere is a lack of site-specific data or consensus on which parameters to choose, given a range of possibilitiesâ (EPA 2004a, p. 105). as the committee undertook its technical evaluations, it remained sensitive to the question of whether the Red Bookâs framework for risk assessment and its conceptual underpinnings are adequate to meet the challenges of understanding and managing the array of environmental threats to health and the environment that we are expected to face in the foreseeable future. These considerations have also shaped other approaches to thinking about risk assessment including PCCRARM (1997) and a recent publication by Krewski et al. (2007). Current Concepts and Practices EPAâs statement of task for this committee (Appendix B) seeks a âscientific and technical review of EPAâs current risk analysis concepts and practices.â In addition, EPA invites the committee to develop ârecommendations for improvingâ EPAâs risk-analysis approaches, âtaking into consideration past evaluations.â At the outset, the committee approached its task in part by reviewing major National Research Council reports published since 1983. It also examined EPA risk-assessment activities in light of themes and trends in those reports. The discussion that follows highlights EPAâs progress in many spheres and shortfalls and committee uncertainty about the nature and extent of progress. The National Research Council reports and EPA documents arrayed in the timeline diagram in Figure 2-3 and the timeline table in Appendix C are the primary sources for this analysis. The implementation table in Appendix D isolates and highlights National Research Council recommendations on selected risk-assessment topics with relevant EPA responses as documented in a recent EPA Staff Paper (EPA 2004a), guideline documents, and other EPA sources; it also draws on a Government Accountability Office (GAO) study requested by Congress (GAO 2005).
1983 1986 1987 1991 1992 1993 1994 1996 1997 1998 2000 2002 2003 2004 2005 2006 NRCâs EPAâs Unfinished EPAâs EPAâs NRCâs NRCâs EPAâs Presidential EPAâs EPAâs NRCâs EPAâs EPAâs EPAâs EPAâs Risk Guidelines Business Guidelines Framework Pesticides Science Guidelines Commission Guidelines Risk Estimating Framework Risk Guidelines Frame- assess- for report for for in the Diets and for Repro- on Risk for Charac- the Public for Assessment for work for ment in Carcinogen Develop- Ecological of Infants Judg- ductive Assessment Neurotox- terization Health Cumulative Principles Carcinogen Assessing the Risk mental Risk and ment in Toxicity; and Risk icity Risk Handbook Benefits of Risk and Risk Health landscape Federal Assessment Toxicity Assessment Children Risk NRC report Management Assessment Proposed Assessment Practices Assessment Risks of Govern- (expanding Risk NRCâs Assess- Understand- reports; and Air Staff Paper; (expanding Environ- Figure 2-3.eps ment 1976 Assessment Issues in ment ing Risk EPAâs Ecological Pollution Air Toxics 1986 mental report guidelines) (expanding Risk Exposure Risk Regulations Risk guidelines) Exposures 1986 Assessment Factors Assessment Assessment to guidelines) Handbook Reference Children Library FIGURE 2-3â Timeline of major documentary milestones. Documents arrayed here represent major risk assessment reports presented in Tables C-1 and D-1 (see Appendixes C and D). Sources: NRC 1983, 1993a,b, 1994, 1996, 2002; EPA 1986a, 1987a, 1991, 1992b, 1996a, 1997c, 1998a,b, 2000b, 2003d, 2004a,b, 2005a, 2006; PCCRARM 1997. 37
38 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT Environmental Protection Agency Progress in Implementing National Research Council Recommendations In general, as shown in Table D-1, National Research Council committees have rec- ommended improvements related to a broad array of risk-assessment issues and activities. Most recommendations provide technical advice on scientific topics, such as cumulative risk, toxicity assessment, mode of action, and uncertainty analysis; but others address associated matters, such as peer review, guideline development, and principles like the conceptual distinction between risk assessment and risk management. EPA responses to the recommen- dations take several forms, including internal guidance memoranda and formal guidelines, handbooks and manuals, new programs, and standing committees to study identified risk- assessment topics. Table D-1 shows that some recommendations have prompted complementary activities in various agency offices. For instance, the agency has both generic and program-specific guidance related to cumulative risk and aggregate exposure (Table D-1). Agencywide guid- ance issued under the auspices of the Science Policy Council and Risk Assessment Forum includes a 1997 guidance memorandum and supplemental guidelines for chemical mixtures. Individual offices have undertaken separate projects to meet office-specific needs. Examples include, for the Office of Air and Radiation, the Integrated Air Toxics Strategy (64 Fed. Reg. 38705 ), the TRIM model (EPA 2007a), and the Multiple Pathways of Exposure Model (EPA 2004b); for the Office of Pesticide Programs (OPP), the report Guidance on Cumulative Risk Assessment of Pesticide Chemicals That Have a Common Mechanism of Toxicity (EPA 2002a); and for the Office of Research and Development (ORD), the cumula- tive-risk components of the Human Health Research Strategy (EPA 2003a). Table D-1 shows a long-standing emphasis on ârisk characterizationâ in both National Research Council recommendations and EPA guidance memoranda, formal guidelines, and other documents (see Box 2-3). The 1994 National Research Council committee described risk characterization as involving integration of information developed in the hazard-iden- tification, dose-response, and exposure analyses and âa full discussion of uncertainties as- sociated with the estimates of riskâ (NRC 1994, p. 27). The agencyâs risk-characterization guidance, including a handbook (EPA 2000b) devoted to the topic, was consistent with that recommendation in emphasizing âtransparencyâ and âclarityâ in explaining risk-assessment approaches and results, especially specifying strength and weaknesses of data and methods and identifying related uncertainties. Citing 1994 National Research Council recommendations for greater attention to the use of defaults, EPA applies this general risk-characterization guidance to the specific subject of defaults in the proposed (EPA 1996b) and final (EPA 2005a) cancer guidelines (Table D- 1). Those documents articulate the scientific basis of five major defaults used in cancer risk assessment in the absence of scientific data. The Staff Paper (EPA 2004a) explains that the agency âinvokes defaults only after the data are determined to be not usable at that point in the assessmentâ (EPA 2004a, p. 51), emphasizing that this is a âdifferent approach from â ational Research Council recommendations are not by themselves responsible for EPA activities on topics N covered by them. EPAâs Science Advisory Board (SAB) and the International Life Sciences Institute (ILSI) Risk Sci- ence Institute have also provided recommendations on these issues. The burst of activity on cumulative risk and aggregate exposure, for example, reflects a confluence of such factors as new statutory requirements in the 1996 FQPA and advances in the state of the science. â he 1996 proposal cited here and elsewhere (for example, Table D-1) represents an intermediate step in the T evolution of EPA cancer principles from 1986 to 2005; also, although the guidelines were not completed for almost 10 years, the 1996 proposal documented contemporaneous EPA work on the 1994 National Research Council recommendations related to cancer risk assessment.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 39 BOX 2-3 Agency Guidance on Risk Characterization: Attention to Uncertainty A 1992 guidance memorandum reinforces principles enunciated in the 1983 Red Book and in EPAâs 1986 risk-assessment guidelines and was a forerunner of later guidance documents. Highly reliable data are available for many aspects of an assessment. However, scientific uncertainty is a fact of life for the risk assessment process as a whole. . . . Scientists call for fully characterizing risk not to question the validity of the assessment, but to fully inform others about critical information in the assessment. . . . Even though risk characterization details limitations in an assessment, a balanced discussion of reliable conclusions and related uncertainties enhances, rather than detracts, from the overall credibility of each as- sessment [Reprinted in NRC 1994, Appendix B, pp. 352-353]. The Risk Characterization Handbook (EPA 2000b) instructs risk assessors to, among other things, âcarry forward the key information from hazard identification, dose-response, and exposure assess- ment, using a combination of qualitative information, quantitative information, and information about uncertaintiesâ (p. 24) and âdescribe the uncertainties inherent in the risk assessment and the default positions used to address these uncertainties or gaps in the assessmentâ (p. 21). After highlighting the emphasis on âtransparencyâ in EPAâs 1995 risk-characterization policy (EPA 1995), the Staff Paper (EPA 2004a) notes that âone of the major comments on EPA risk assessment practices is that they do not characterize uncertainty and variability transparently enoughâ (p. 33). The statement of task for EPA (2004a) confirms that âthis is an issue EPA is attempting to addressâ (p. 33). (See Box 2-4 for related peer-review commentary on one assessment.) choosing defaults first and then using data to depart from themâ (EPA 2004a, p. 51), as in the past. The committee found this framing of defaults problematic, as discussed at length in Chapter 6. Table D-1 was instructive for the present committeeâs review of EPA risk-assessment concepts and practices called for in the statement of task. For example, GAOâs survey reports broad-based approval in EPA of the program for developing risk-assessment guidelines in line with 1983 Red Book recommendations for inference guidelines (Table D-1). As new methods emerged, the agency revised and updated several of the original 1986 guidelines (on cancer, developmental toxicity, mixtures, and exposure assessment). The addition of new topics to the guideline library, such as neurotoxicity in 1998 and ecologic risk assessment in 1998, suggests that adding other new topics may be a useful way to implement recommendations in the present report. EPAâs response to recommendations from its Scientific Advisory Board (SAB) and the National Research Council for an enlarged peer-review program offers another model for the future. EPAâs 1992 and 1994 peer-review policy memorandums (EPA 1992c, 1994) expanded peer review beyond statutory mandates10 to âmajor scientifically and technically based work products related to Agency decisionsâ (EPA 2000b; Table D-1). The general objective of both the National Research Council recommendations and EPAâs new policy was to add scientific expertise to the overall risk-assessment process. The expanded policy was intended to move assessments not then subject to peer review into the ambit of peer review. The calls for more peer review, like the call for more stakeholder participation, demonstrate concern about both the increasing complexity of risk assessment and the credibility of EPA assessments. However, EPA (2000b) acknowledges the need for upfront planning of the peer 10â ection 109 of the CAA requires peer review of the criteria documents setting forth the scientific analyses S underlying national ambient air quality standards; Section 6 of FIFRA requires peer review of identified pesticide actions. See also 70 Fed. Reg. 2664  (federal peer-review guidelines).
40 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT review to ensure it provides the appropriate insight and direction to the risk assessment. In that regard, the new framework proposed in this report may well require a different kind of peer review in which experience and expertise on decision theory, social sciences, and risk management may be required along with scientific expertise. The enormous variety and scope of EPA risk-assessment responsibilities and activities preclude a detailed and full assessment by the present committee of risk-assessment practices in all parts of the agency. For example, the GAO survey (see GAO , Table D-1) implies extensive use of the guidelines by agency risk assessors but does not provide information on the extent to which individual risk assessments (assessments of particular hazardous air pol- lutants, pesticides, or Superfund sites) follow some of or all of the principles enunciated in the guidelines. Similarly, even with the strong emphasis on identifying uncertainties, explaining defaults, and justifying science-policy choices as critical features of risk characterization in EPA guidance documents (see Table D-1 and Box 2-3), peer reviewers and other commenters recommend greater clarity and transparency in characterizing variability, uncertainty, and risk (GAO 2006; see Box 2-4 for one example). Those concerns raise questions about the extent to which guidance on risk characterization is fully used in practice, whether the guid- ance is adequate, and how to guide characterization during periods when science, practice, and expectations are evolving. The 1994 National Research Council report called for explanation of the scientific basis of default options and identification of âcriteriaâ for departure from defaults. Guidelines for Carcinogen Risk Assessment (EPA 2005a) includes as an appendix an extended discussion BOX 2-4 Commentary on Risk Characterization for the Dioxin Reassessment In a recent report (NRC 2006) on EPAâs dioxin reassessment (EPA 2003b), the peer-review panel complimented some features of EPAâs approach to scientific uncertainties in the assessment and then recommended that the agency âsubstantially revise the risk characterization section of Part III of the Reassessment to include a more comprehensive risk characterization and discussion of the uncertain- ties surrounding key assumptions and variablesâ (NRC 2006, p. 25). For more than 20 years, EPA guidance documents have stressed displaying âall relevant informa- tion pertaining to the decision at handâ (EPA 1984, p. 14), fully informing others about âcritical informa- tion from each stage of a risk assessmentâ (EPA 2000b, p. A-2), and the importance of transparency and âdescribing uncertainties inherent in risk assessment and default positionsâ (p. 21), among other things. See Box 2-3 and Table D-1 (section on risk characterization) for fuller statements and references. In view of this long-standing internal guidance emphasizing complete and transparent characterization in agency risk assessments, the need for âsubstantial improvementâ in EPAâs description of the scientific basis for key elements in this important assessment suggests inattention to principles enunciated in EPA guidance (NRC 2006, p. 9; emphasis in original): The Committee identified three areas that require substantial improvement in describing the scientific basis for EPAâs dioxin risk assessment to support a scientifically robust risk characterization: â¢ Justification of approaches to dose-response modeling for cancer and noncancer end points. â¢ Transparency and clarity in selection of key data sets for analysis. â¢ Transparency, thoroughness, and clarity in quantitative uncertainty analysis. The calls for improved risk characterization in dioxin risk assessment by NRC (2006) illustrate the need for greater clarity and transparency that are often voiced in reviews of EPA risk assessments. Consistent with the statement of task, this report develops information and approaches for addressing these issues.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 41 of the scientific basis of defaults and alternatives but does not provide criteria for invoking defaults (see Chapter 6). Chapter 6 of the present report analyzes EPA implementation of selected recommenda- tions regarding defaults in greater depth than in Table D-1. For example, Table 6-3 charac- terizes some EPA practices as implicit or âmissingâ defaults. As also shown in Table D-1, National Research Council committees have made various recommendations related to un- certainty analysis. However, as noted in Chapter 4, uncertainty analysis and characterization pose difficult technical issues, and in general related best practices have not been established. In the absence of guidelines on the appropriate degrees of detail, rigor, and sophistication needed in an uncertainty analysis for a given risk assessment, it is not surprising that expert advisory committees recommend technical improvements in this regard.11 (See Box 2-5 on importance of implementation of guidelines.) EPA and GAO comments on the Integrated Risk Information System (IRIS) (see Table D-1) may be instructive as to the outlook for the present committeeâs recommendations to the agency. The GAO report details numerous improvements in the IRIS process over the past 10 years. It also indicates that in 2005 EPA completed only eight IRIS reviews, falling âconsiderably shortâ of the recommended (and highly optimistic) goal of 50 each year (GAO 2006). GAO states that agency officials explained the shortfall in terms of such factors as risk-assessment complexity, resource limitations, and peer-review requirements.12 Those fac- tors will also be at play as the agency applies recommendations in this report to the current IRIS backlog and to new risk assessments for individual chemicals or sites.13 Similarly, in reporting that 90% of its 2002 scientific and technical work products were peer-reviewed (Gilman 2003; Table D-1), the agency also tracks how the peer-review comments were ad- dressed (EPA 2000c). In sum, Table D-1 identifies both EPA guidance responding to National Research Council recommendations and an impressive set of practices undertaken to improve agency risk assessments. However, the breadth and scope of EPAâs risk-assessment agenda limit the table to a selected subset of current concepts and practices. Although the record demonstrates the extent to which National Research Council recommendations have been implemented on paper through guidelines and other guidance statements, the committee does not have detailed information on the extent to which the guidelines have been fully and effectively incorporated in practice. As EPA explained to GAO (in relation to IRIS), many factors could lead to partial implementation, including data availability, staff expertise and experience, resource constraints, adequate peer review, and the impact of statutory deadlines and legal frameworks on the risk-assessment process. The Role of Policy Each stage in the risk-assessment process calls for a series of choices, each with the potential to influence, and in some cases determine, the outcome of the risk assessment. As 11â or example, one recent review âfinds that EPA guidance concerning specific use of the Integrated Exposure F Uptake Biokinetic (IEUBK) model and additional use of blood lead studies is incomplete. . . . The Office of Solid Waste and Emergency Response (OSWER) directive fails . . . to give adequate guidance about what to do when [data] and IEUBK model results disagree by a substantial marginâ (NRC 2005a, p. 273). 12â One published paper reports that in 2006 EPA added only two assessments to the IRIS database (Mills 2006). 13âThe impact of these factors on the high-profile IRIS program, which is based in the scientist-rich ORD, raises questions about the capacity of the agency as a whole, where many risk assessors have less experience than those in ORD, to expand its risk-assessment activities in line with recommendations set forth in this report. See Chapter 9.
42 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT BOX 2-5â Guideline Implementation and Risk-Assessment Impacts As shown in Box 2-1, EPAâs library of risk assessment guidelines covers a broad array of topics. In 1994, the NRC committee concluded that âthe guidelines were generally consistent with the Red Book recommendations. . . . They include default options which are essentially policy judgments of how to accommodate uncertainties. They include various assumptions that are needed for assessing exposure and riskâ (NRC 1994, p. 5). Despite conformity with the Red Book, approval of peer reviewers (see peer-review history at the front of individual guidelines), and staff appreciation of the guideline documents (GAO 2006), concerns identified in EPAâs Staff Paper (EPA 2004a) and the GAO report (GAO 2005) regarding EPA risk assessments (for example, overconservativism and underconservatism in risk estimates, use or nonuse of defaults, incomplete discussion of uncertainty, and delays in completing assessments) prompt questions about the extent to which the guidelines fulfill their intended function in individual assessments. That is, to what extent are problems associated with EPA risk assessments traceable to guideline content or use? One question is related to the scientific adequacy and general utility of the guidelines themselves as a resource for assessors and managers; that is, do they provide information needed in a usable form? A second question is related to risk assessorsâ use or nonuse of the guidelines in any particular case; that is, do assessors and managers have the technical experience, scientific data, funding, and time to use the guidelines as intended? (See Box 2-4 for an example of incomplete attention to existing guidance.) Factors contributing to ineffective guidelines or guideline use may include â¢ Nonavailability of relevant data, risk-assessment methodology (for example, established de- faults), or both. â¢ Complexity, lack of clarity, or infeasibility in the recommendations by the National Research Council and other bodies that advise the agency. â¢ Complexity, lack of clarity, or infeasibility in the related EPA guidelines. â¢ Optional vs mandatory wording in the guidelines. â¢ Individual or ad hoc policy overriding guideline policy. â¢ Lack of experience on the part of risk assessors. â¢ Management issues, such as lack of experience or oversight on the part of supervisors and decision-makers. In view of EPAâs pattern of developing guidelines to address previous National Research Council recommendations (Table D-1), understanding of factors that influence effective use of the guidelines by assessors and managers could be critical for effective implementation of recommendations in the present report. developed more fully in Chapters 4-7, the data gaps and uncertainties inherent in the process generate the need for defaults and assumptions; in addition, alternative approaches to each assumption introduce the element of choice (NRC 1994, p. 27): Risk assessors might be faced with several scientifically plausible approaches (for example, choosing the most reliable dose-response model for extrapolation beyond the range of ob- servable effects) with no definitive basis for distinguishing among them. The [Red Book] committee pointed out that selection of a particular approach under such circumstances involves what it called a science-policy choice. Science-policy choices are distinct from the policy choices associated with ultimate decision-making. . . . The science-policy choices that regulatory agencies make in carrying out risk assessments have considerable influence on the results.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 43 However, it is critical that science-policy choices underlying risk-assessment guidelines be based on the need for consistency, reproducibility, and fairness. Some choices are normal aspects of scientific endeavors, whether part of a regulatory pro- cess or not. For example, each stage of the risk-assessment process involves an initial survey of the scientific literature and relevant databases to identify and isolate studies pertinent to the pollutant or situation under review. The array includes information from many sources: reports in peer-reviewed journals, reports in the gray literature, personal communications about recent results not yet published, and the like. Some studies have been replicated or oth- erwise substantiated; others may have a questionable provenance. Judgments on those issues parallel judgments made in developing any scientific analysis. Continuing analysis involves reviewing each study for fundamental strengths and weaknesses, for example, quality-assur- ance issues, replicability, consistency with comparable studies, and peer-review status. Other considerations are specific to the regulatory process. They include the relevance of any particular piece of evidence in the decision context (see Chapters 3 and 8), informa- tion submitted by stakeholders and other interested parties, applicability of relevant agency policies and guidelines, and factors that might compromise use of data for standard-setting purposes (for example, the presence of potential conflicts of interest in generating or censor- ing data). It is easy to narrow the options by eliminating nonconforming studies. However, more than one study may meet basic scientific standards, and studies vary with respect to quality attributes. Benchmark dose (BMD) calculations for perchlorate offer an example, as de- scribed in a recent National Research Council peer-review report (NRC 2005b, p. 170): As part of its deliberations on the point of departure, the committee reviewed the BMD analy- ses conducted by EPA (2003c), the California Environmental Protection Agency (CalEPA 2004), and Crump and Goodman (2003) on the data from Greer et al. (2002). Overall these analyses used different models, approaches, parameters, response levels, and input data, so comparison of the results of the analyses is difficult. The task, then, was to identify the âcriticalâ study or studies for use in continuing the risk assessment (see, for example, EPA 2002b, 2004a, 2005a,b), which may involve choosing among or combining varied results from different scientifically adequate studies. When dif- ferent scientists make different judgmentsâthat is, different choicesâamong the alternative studies, related risk-assessment results may differ substantially (Box 2-6). In addition to choosing one set of âhardâ data over another where necessary, risk asses- sors identify uncertainties and unknowns at each stage in the process. In the hazard-identifi- cation stage, questions about the applicability to humans of findings in specific animal studies lead to uncertainty in the animal-to-human extrapolation, an assumption that data in those studies are predictive of adverse effects in humans under particular conditions of exposure. When relevant data are unavailable, other uncertainties lead to questions on other matters, such as the relevance of effects observed in studies on males to females, adults to children, and âhealthyâ workers to the general population. Similar uncertainties are important in all types of risk assessments. The dose-response analysis almost invariably raises questions about the likelihood that effects observed at the generally higher doses used in animal studies (or under conditions of workplace exposures) would be observed at the generally lower doses expected in connection with environmental exposures. As shown for perchlorate, the number of choice points and the options at each point open the door to different reference dose (RfD) values, depending on the combination of choices made:
44 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT Box 2-6 Choices and a Reference Dose Value for Perchlorate In 2002, EPA issued a draft reference dose (RfD) for perchlorate, a contaminant found in public drinking-water supplies for more than 11 million people. After peer-review challenges to the scientific basis of EPAâs proposed RfD, the National Research Council produced an independent analysis at the request of several agencies. â¢ EPA based the RfD on adverse effects in rats; the National Research Council committee chose a key biochemical event seen in healthy humans that would precede adverse effects as the basis of the RfD (NRC 2005b, pp. 14, 166). â¢ EPA used changes in brain morphometry, thyroid histopathology, and serum thyroid-hormone concentrations in rats (oral exposure) as the basis of its point of departure for the RfD calculation; the National Research Council committee recommended using inhibition of iodine uptake by the thyroid in a small group of exposed healthy humans, a nonadverse effect, as the basis of the point of departure (p. 168). â¢ EPA selected a âcompositeâ uncertainty factor of 300 to account for animal-human differ- ences, use of a lowest-observed-adverse-effect level, lack of chronic data, and other database gaps (p. 172). The National Research Council committee used a total uncertainty factor of 10 to account for interindividual variability (p. 178). This was consistent with the use of human data, and assumed that the point of departure was a no-observed-effect level. EPA had proposed an RfD of 0.00003 mg/kg per day; committee recommendations would lead to an RfD of 0.0007 mg/kg per day (p. 178). In 2005, EPA responded to the National Research Council recommendation by issuing a new perchlorate RfD of 0.0007 mg/kg per day (EPA 2005c). The analytic process for this chemical indicates that different scientific bodies can come to different risk conclusions, with a majority of the differences arising from different emphases placed on datasets and on how uncertainty and variability are viewed. Large-scale epidemiology studies can bring these variability and risk issues into sharper focus. For example, a recent large Centers for Disease Control and Prevention study found associations between relatively low perchlorate exposures and reduced thyroid function in sensitive populations of women (Blount et al. 2006). Further followup studies will provide insight as to whether the current RfD is adequate. Further analysis of CDC data suggest an interaction of perchlorate and tobacco smoking (perhaps via thiocyanate) to affect thyroid function (Steinmaus et al. 2007). â¢ Use of BMD or low dose for RfD calculation. â¢ Use of the lowest-observed-adverse-effect level or the no-observed-adverse-effect level. â¢ Use of the ED01, ED05, or ED10 to define the benchmark response.14 â¢ For noncancer end points, an uncertainty factor of 1, 10, 100, 1000, or other. â¢ For carcinogens, a threshold or nonthreshold approach. Exposure assessment can involve an even broader range of uncertainties and related choice points. Some are related to the fate and transport of the pollutant in the environ- ment, others to data on and uncertainties about the metabolism, distribution, and fate of the chemical in the target population. In each case, chemical-specific data are rarely available on all the parameters critical for estimating expected exposures. 14â D , ED , or ED is the dose associated with either a 1%, 5%, or 10% increase in an adverse effect relative E 01 05 10 to the control response (EPA 2008).
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 45 As a result, exposure scenarios are just thatâhypothetical situations based on combi- nations of measured and, where data are unavailable, modeled estimates of the form and amount of a chemical in the environment or human tissue. They often combine data specific to the chemical at issue and, where such data are not available, data on similar chemicals or on the same chemical in different conditions. After examining the database for answers to these questions, EPA risk assessors turn to assumptions and extrapolation to develop information for completing an assessment: â¢ In the absence of chemical-specific data, what data on what other chemicals best represent the chemical under study? â¢ In the absence of reliable measurements of exposure in the environment, which as- sumptions and models can be expected to provide reasonably valid estimates? â¢ In the absence of reliable measurements of tissue exposure in humans, which as- sumptions and models can be expected to provide reasonably valid estimates? â¢ Of several potentially vulnerable populations (for example, infants, children, the elderly, and pregnant women) with comparable exposure potential, which populations are the most sensitive and in need of protection under the standard? â¢ When and how should exposure assessment take account of cumulative or aggregate exposure? Choices at those and other decision points shape predictions of risks to populations of interest and the credibility of the risk assessment itself. Superimposed on those choices among candidate scientific studies, assumptions, mod- els, and the like, policy choices are required as to which scientifically plausible assumptions and models to use in completing the assessment. The process is designed to accommodate discussion of the choices and the reasons for them. The Red Book paradigm and successor reports and EPA guidance documents stress the importance of characterizing risk by advis- ing decision-makers and the public about uncertainties, assumptions, and choices made. A National Research Council report on EPAâs dioxin reassessment illustrates the point (NRC 2006, p. 55): The impact of the choices made in the risk assessment process can be characterized by quan- tifying the impact of plausible alternative assumptions at critical steps. The risk estimates can be most fully characterized by performing probabilistic analyses when possible and by presenting the range of possible risk estimates rather than by reporting the single point estimates. Risk characterization should provide useful information to risk managers to help them understand the variability and uncertainty in the risk estimates. Chapter 6 of the present report provides additional recommendations on developing alternative risk estimates in light of plausible alternatives to defaults. The Red Book points out that ârisk characterization, the estimate of the magnitude of the public health problem, involves no additional scientific knowledge or conceptsâ (NRC 1983, p. 28). Rather, it calls for synthesizing information from the preceding analyses with special attention to identifying uncertainties and their impact on the assessment (see Chapter 4 of this report). The Role of Time Time is a major and rarely acknowledged influence in the nature and quality of environ- mental risk assessment in EPA. Some time factors are immediately obvious. The statutory deadlines for some regulatory decisions necessarily require completed risk assessments to
46 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT meet the deadlines. When EPA fails to meet a standard-setting deadline, as often happens, regulated entities, advocacy groups, and other interested parties exercise their statutory right to bring âdeadlineâ suits, which result in court orders to issue standards by a specified date. The result may bring closure to an assessment that has been languishing or lead to an assessment that meets the deadline but falls short of some scientific standards. Such statutory requirements constitute advance notice of the need for specific risk as- sessments in specified timeframes and can lead to regular schedules for many assessments and related analyses. Examples of such requirements include the 5-year cycle for review and revision of the national ambient air quality standards (Section 109) and the 8-year deadline for maximum achievable control technology (MACT) standards for hazardous air pollutants (Section 112) under the CAA amendments of 1990. In 1996, Congress set new deadlines for pesticide actions under the FQPA, requiring the agency to reassess the risks of all existing pesticide food tolerances (standards) over a ten year period; that same year Congress enacted a new Safe Drinking Water Act requiring the agency to select five new contaminants each year for decisions on maximum contaminant levels (MCLs) for drinking water. Several predictable but highly variable factors can upset the best-laid plans. The most obvious is the unavailability of scientifically reliable and context-relevant data and methods. Other situations can be cited. Some involve new research or monitoring data that identify issues that affect the assessment or information on the imminent appearance of new studies expected to make a substantial difference in the analysis; others involve emergency environ- mental problems or changes in political priorities that result in reassignment of resources and staff to other assessments. Undue political influence in the process can also result in delays (GAO 2008). And initial planning may have been inadequate with regard to what could reasonably be achieved with available data and resources and the corresponding set- ting of unreasonable expectations. In some circumstances, EPA is faced with an abundance of data, especially on high-profile chemicals. Specifically, where chemicals have been studied for many years, multiple studies of comparable quality on a single chemical may yield different results, in some cases large differences in RfDs or risk and in other cases slight but critical differencesâa situation that invites debate and controversy and may take years to resolve. In these circumstances, new studies and new data, while at the same time shedding light on assessments, can complicate reviews (Box 2-7). However, it is important to recognize the value of analyses that synthesize data across a number of different studies and end points, which can result in a more precise and defensible analysis. In addition to recommending attention to previously unavailable new studies, almost ev- ery peer review recommends research that would improve the assessment. Recommendations of both types hold the prospect of reducing uncertainty and contributing to a more reliable risk assessment. Such recommendations also invite delay, require additional resources, and contribute to ambiguity as to whether the assessment is scientifically sufficient. Such delay can have significant impact on communities who are awaiting risk assessment results to make decisions regarding the safety of their neighborhoods where hazards may be present. Iteration is an important feature of an adequate risk-assessment process and should be built into the planning. Addressing late-arising problems uncovered in discussion between assessors and managers will improve the assessment but may also delay its completion. Similarly, stakeholder and peer-review involvement brings many benefits but may extend the process. Changing administrations may also add to the time required.15 15â PAâs recent dioxin reassessment and cancer guidelines are examples. Specifically, the dioxin report or parts E of it were submitted for peer review on several occasions from 1992 to 2003, when a National Research Council
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 47 Box 2-7 Impact of New Studies In 1997, concern about the effects of human exposure to mercury led Congress to request a National Research Council review of EPAâs RfD for methyl mercury (MeHg). At the time, scientists were awaiting results from studies of three populations because the existing RfD was based on a 1987 study of 81 Iraqi children accidentally exposed in utero (NRC 2000a, p. 306). Noting that MeHg exposures in the Iraqi study population were not comparable with low-level chronic exposures expected in North American populations, the National Research Council committee recommended basing the RfD on new studies that were incomplete at the time of the 1997 Mercury Study Report to Congress (EPA 1997b). A National Research Council committee recommended that EPA retain the 0.1-Î¼g/kg per day RfD but replace the study used to set the RfD with new studies: âSince the establishment of the current RfD, results from the prospective studies in the Faroe Islands (Grandjean et al. 1997, 1998, 1999) and the Seychelles (Davidson et al. 1995a,b, 1998), as well as a peer-reviewed re-analysis of the New Zealand study (Crump et al. 1998) have added substantially to the body of knowledge concerning the develop- mental neurotoxic effects of chronic low-level exposure to MeHgâ (NRC 2000a, p. 312). Similarly, National Research Council recommendations on the long-running dioxin assessment expand the scope of the assessment: âEPA is encouraged to review newly available studies on the ef- fects of TCDD on cardiovascular development in its risk assessment for noncancer end pointsâ (NRC 2006, p. 174). Perchlorate (Box 2-6) provides an example of how emerging data may inform risk after an as- sessment has been finalized. The iterative nature of risk assessment and research ensures that new data will enter the process. The salutary effect of new data can also result in additional time for analysis and incorporation of data into the risk assessment. In some ways, problems with timeliness are inherent in a decision-making environment that places a premium on âsound scienceâ or âcredible science.â The nature of the conflict can be understood if it is recalled that the scientific process of seeking the truth, by design and to its credit, has no natural end point. In addition, the training of scientists, by design, and the embedded cultural traditions, such as requiring p values in tests of significance, instill values of prudence, replication, scientific debate, and peer review as prerequisites of a conclu- sion characterized as âsound science.â This issue is discussed in more detail in Chapter 3. Institutional Arrangements For Managing the Process Consideration of EPAâs risk-assessment accomplishments and shortfalls and of the effects of policy and time leads to questions about institutional arrangements for âmanaging the process,â the subtext of the Red Book. EPA has established an enormous array of programs for this purpose. The combination of people and programs reflects close attention to statutory requirements and advisory-body recommendations. That salutary orientation around diverse statutory requirements also leads to criticism of âapparent inconsistencies in risk assessment panel undertook the most recent review. EPAâs cancer guidelines were first published for comment and peer review in 1996; intermediate reviews took place before publication as final guidelines in 2005. Work began on both documents in the late 1980s. The development period included changes in the general approaches to risk assess- ment and specific new data and theories regarding cancer risk assessment and the toxicity of dioxin. In addition, several changes at the White House during this period led, at different times, to EPA decision-makers with different constituencies.
48 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT practices across EPAâ (EPA 2004a, p. 14), which are traceable to statutory and managerial, as well as scientific, factors and to calls for greater coordination of agency programs. Environmental Protection Agency Risk-Assessment Programs and Activities EPAâs major program offices have scientific responsibilities on the one hand and regu- latory responsibilities on the other. For scientific data development and risk assessment, the agency relies on environmental professionals trained in diverse technical disciplines, such as chemistry, geology, toxicology, epidemiology, statistics, and communication. For risk management and regulatory decision-making, professionals in economics, engineering, law, and other fields work with agency policy-makers to shape regulatory decisions. As indicated in agency guidelines and other documents, assessors and managers have different roles but interact regularly throughout the process (EPA 1984, 2003d, 2004b; Table D-1, sections on âdistinguishing linking risk assessment and risk managementâ and âproblem formulationâ). In addition to different statutes and scientists with expertise in many fields, EPAâs risk- assessment work takes place in a variety of organizational and geographic locations and includes collaborative activities with numerous public and private scientific organizations. The result is a complex set of interactions that strengthen the agencyâs risk-assessment pro- cesses in the main, but the diversity of inputs also introduces drawbacks. Each major program office manages several risk-assessment activities. For example, the Office of Water has programs for conducting health risk assessments under the Safe Drinking Water Act (SDWA) and ecologic risk assessments under the Clean Water Act. The Office of Air and Radiation conducts human health risk assessments for use in setting regulatory stan- dards related to âcriteriaâ pollutants (such as particulate matter [PM] and sulfur dioxide), in a different program âhazardousâ pollutants (such as arsenic and mercury) from stationary sources, and in still another program pollutants from cars and other mobile sources. That office is also responsible for assessments related to stratospheric ozone depletion and acid rain. As evident in EPAâs Science Inventory (EPA 2005d), other agency offices have compa- rably wide-ranging programs for a total set of activities that almost defies description. The diverse risk-assessment tasks impose demands for both breadth and quality in staffing and managing these activities. Several offices have overarching responsibilities to help meet the demands. ORD con- ducts environmental research at more than 10 laboratories and centers around the country. The laboratories are organized around the basic units in the risk-assessment paradigm (for example, effects, exposure assessment, and risk characterization). ORD plans, conducts, and oversees most EPA risk assessments and risk-assessment-related research for the agency as a whole. In addition to its core program of fundamental research, a substantial portion is planned in collaboration with program and regional offices to address data needs for regu- lation. In keeping with congressional and agency guidance priorities, ORD-led multioffice research-planning teams coordinate planning and budgeting in line with data needs identified by program and regional offices. However, it is important to note that because EPA relies heavily on data in the published literature and these are not the studies conducted by EPA, there is no mechanism for developing the data necessary to address emerging issues, and this contributes to a scarcity of data on particular agents. ORD scientists coordinate generic risk-assessment activities, such as guideline develop- ment and the reference-doseâreference-concentration (RfD-RfC) process, including manage-
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 49 ment of the IRIS database. ORD also conducts individual chemical-specific assessments at the behest of program and regional offices and, variably, in collaboration with them.16 Some offices are staffed to meet particular needs. In keeping with its responsibilities to oversee the safety of pesticide products, OPP employs a highly specialized scientific staff to evaluate data related to testing and licensing requirements for new pesticides before they are marketed and to conduct risk assessments to set limits on the use of pesticides as appropriate. Because pesticides are toxic by definition, this office has special statutory authority to man- date testing procedures and require specific scientific data from pesticide manufacturers. The authority to mandate data generation in that way is not generally available to other offices, which depend on ORD, the scientific literature, and outside contractors. One of the paradoxes of the risk-assessment process is that the same scientific uncertainties that hamper and complicate risk assessment stimulate the development of new data and methods. For example, scientific uncertainties and controversy related to standards under development for PM led to special funding for new research to reduce the uncertainties (see NRC 1998, 1999a, 2001a, 2004). EPA regularly incorporates the expertise of external scientists into its risk-assessment activities. The agency has extensive long-term and ad hoc collaborative relationships with numerous risk-assessing entities in the public and private sectors. Public-sector partners include other federal entities, such as the National Toxicology Program, which is admin- istratively housed in NIEHS; Argonne and other Department of Energy national laborato- ries; and the FDA National Center for Toxicological Research. Private-sector collaborators include the Health Effects Institute in Boston, ILSI, and the American Chemistry Council in Washington. EPA scientists also participate in numerous international programs, such as the UN International Programme on Chemical Safety (IPCS), of which the World Health Organization (WHO) is a partner. The IPCS Harmonization Project, which is designed to harmonize approaches to the assessment of risk, has been a particularly influential partner with EPA in advancing the practice of risk assessment. EPAâs ten regional offices have risk-assessment and regulatory activities correspond- ing to those in the major program offices but focused at the local level. They have diverse risk-assessment responsibilities. Scientists interact with EPA program offices in Washington, DC, and ORD risk-assessment centers and laboratories on the one hand and with nongov- ernment organizations and state, local, and tribal entities on the other. In some cases, the regional offices apply risk assessments or toxicity values (for example, RfD, RfC, or potency estimates from IRIS) developed elsewhere to regional problems; in other cases, they develop region-specific assessments. Through those interactions, state, local, and tribal information and perspectives become part of the process. The diverse inputs to risk assessment in EPA are a natural outgrowth of the diverse environmental problems facing the nation and the agency and of the scientific complexities of the risk-assessment process. Several EPA activities, including risk-assessment guidelines and the RfD-RfC process, are designed to counteract the effects of compartmentalization by standardizing and unifying some of the diverse elements. In addition, the Office of the Science Advisor coordinates the work of two standing committees with agencywide, rather than program-specific, risk-assessing responsibilities. The Risk Assessment Forum was char- tered in response to recommendations in the 1983 Red Book. Somewhat later, the agency set up a Risk Management Council composed of senior EPA risk managers with oversight 16â n I addition to the ORD laboratories, program and regional offices manage laboratories, such as that in Ann Arbor for the air program, that in Bay St. Louis for the pesticide program, and the National Enforcement Inves- tigation Center in Colorado.
50 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT responsibilities for forum activities. Later, renamed and rechartered as the Science Policy Council, that group has enlarged its membership and responsibilities to address a variety of science-policy issues. Risk Management: Regulations and Risk Assessment EPA statutes lodge responsibility for regulatory decisions with the EPA administrator and the assistant administrators who head the program offices. All are political appointees who require Senate confirmation and generally change when the White House changes hands. In their roles as risk managers, those officials are responsible for using completed risk assessments with information from other disciplines to shape regulatory decisions. In addition, they and other risk managers provide oversight for the risk-assessment process from inception to conclusion. As indicated above, the 1983 Red Book stressed the importance of a âconceptual dis- tinctionâ (p. 7) between risk assessment and risk management but rejected the concept of âinstitutional separationâ between the processes. EPA adheres to those principles in the sense that, although assessors and managers are colocated and interact regularly, assessors do not set standards and decision-makers do not conduct risk assessments. Owing to the committeeâs statement of task, this chapter has focused on the evolution of risk assessment and related practices. The committee considers that the same degree of concern about uncertainty, variability, and inferences that has been applied to the assessment of risks should also be applied to the assessment of costs, but this was beyond the scope of this report. For example, economists on the administratorâs planning, evaluation, and innovation staff provide information and analyses on costs and benefits for use in making regulatory decisions and for the regulatory impact analyses (RIAs) that accompany major regulatory actions. (The benefits are computed from the results of risk assessments.) In addi- tion, many program and regional offices have units responsible for analysis of the economic benefits of proposed decisions and regulatory actions. ORDâs National Risk Management Research Laboratory in Cincinnati conducts engineering research for use in developing and evaluating the technical feasibility of pollution-control methods used in formulating regula- tory options. In accord with statutory directives, EPA program and regional offices interact with state and local offices on implementation and compliance issues, such as schedules, costs, feasibility, impacts, and enforcement. Regarding regulation development, as indicated earlier, the Red Book emphasis on the âconceptual distinctionâ between risk assessment and risk management reflects the statutory dichotomy between information used in assessing risk and other kinds of informationââthe public health, economic, social, political consequences of regulatory optionsâ (Figure 2- 1)âused with risk-assessment results to determine âagency decisions and actions.â For example, in evaluating whether a pesticide poses an âunreasonable riskâ to health or the environment, the pesticide law (FIFRA) calls for consideration of the economic, social, and environmental costs of using the pesticide. EPA âinterprets this broad statutory language to mean that any significant benefits to public health through disease control or prevention, or through vector control, need to be considered in the suspension, cancellation, or denial of an application for registration or a determination of ineligibility for deregistration of a public health use of any pesticide that offers such benefitsâ (EPA 2007b). In the same vein, the 1996 amendments to the SDWA explicitly direct EPA to evaluate incremental benefits, costs, and risks associated with compliance with alternativesâa more specific delineation of nonscience considerations than in the original enactment. Differences between the information base for risk assessment, which has science at its
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 51 core, and that for the regulatory decision, which takes account of costs and other nonrisk factors, mean that regulatory decisions are not necessarily congruent with risk assessment. That is, concern about, for example, economic consequences or societal impacts may out- weigh public-health or environmental concerns in such a way as to make a regulatory deci- sion more or less protective than if the decision were based solely on the risk assessment. An additional asymmetry is that the uncertainties associated with cost and benefits are rarely considered although these uncertainties are often explicitly acknowledged in the risk assess- ment. The distinction between the SDWAâs maximum contaminant level goal (MCLG) and the maximum contaminant level (MCL) illustrates the point: the MCLG for a carcinogenic contaminant may be zero, but costs and feasibility concerns may lead the agency to set a regulatory standard, the MCL, to allow a higher level of contamination (see Box 2-8). Some statutes authorize a combination of risk assessment and âtechnology-basedâ processes in setting regulatory standards. Such standards as the SDWAâs MCL illustrate the special case of âtechnology-basedâ standards for which a decision does not depend only on risk assessment. Rosenthal et al. (1992) explain that the SDWA calls for MCLGs, âwhich are concentrations at which no adverse human health effects are believed to occur.â A health- based MCLG is not an enforceable limit. For enforcement purposes, the statute directs EPA to establish a MCL as close to the MCLG as âfeasible with the use of the best technology, treatment techniques, and other means which the Administrator finds after examination for efficiency under field conditions . . . are available (taking costs into consideration)â (42 USC Â§ 300g-1). Other examples appear in the CAA. The 1990 amendments introduced a two-part schemeâpart technology-based, part risk assessmentâfor 189 toxic pollutants regulated under Section 112 of the CAA. The first step directs EPA to identify major emitters of the BOX 2-8 Arsenic in Drinking Water: Uncertainties and Standard-Setting On January 22, 2001, EPA issued a pending standard of 10 Î¼g/L as the maximum contaminant level of arsenic in drinking water. Although the scientific analysis underlying the proposal and the pro- posal itself had been peer-reviewed by both the EPA SAB (1995) and the National Research Council (1999b) and had gone through the public comment process, EPA on March 23, 2001, issued a notice delaying the effective date of the standard to address questions about the science supporting the rule and about the expected implementation costs for affected communities. The National Research Council peer-review committee identified uncertainties and data gaps of several kinds (NRC 2001b): More research is needed on the possible association between arsenic exposure and cancers other than skin, bladder, and lung, as well as noncancer effects. . . . In addition, more information is needed on the variability in metabolism of arsenic among individuals, and the effect of that variability on an arsenic risk assessment. Laboratory and clinical research is also needed to define the mechanisms by which arsenic induces cancer to clarify the risks at lower doses [p. 10]. Nonetheless, the committee made it clear that data gaps and uncertainties do not disqualify the risk assessment for decision-making. There is a sound database on the carcinogenic effects of arsenic in humans that is adequate for the purposes of risk assessment. The subcommittee concludes that arsenic-induced internal (lung and bladder) cancers should continue to be the principal focus of arsenic risk assessment for regulatory decision making, as dis- cussed and recommended in the 1999 NRC report [p. 10]. A final 10-Î¼g/L standard was issued in 2002; EPA and Congress continue to study costs and technical issues associated with implementing the standard (Tiemann 2005).
52 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT pollutants among diverse source categories and requires that these sources use MACT within specified time limits. The second step takes risk into consideration: 8 years after promulgation of MACT standards to limit emissions of the 189 (later reduced to 187)17 pollutants; EPA was required to evaluate the residual risk to the population and promulgate more stringent standards if necessary âto provide an ample margin of safety to protect public healthâ (1990 Amendments to the Clean Air Act, Title III, Â§ 301 (d)(9)). The law specifies that for known, probable, or possible human carcinogens, the administrator is to promulgate revised standards if the MACT standards do not reduce the risk incurred by âthe individual most exposed to emissionsâ from the source of pollution to less than one in a million. With the focus on the âindividual most exposed,â EPA models exposure with fine spatial resolution to characterize the maximum level of exposure associated with a toxic air pollutant. Chapter 4 reviews the current state of the science on variability in susceptibility to cancer, and Chapter 5 provides recommendations to EPA for considering this variability in risk assessments. The CAA takes a different approach in setting national ambient air quality standards for criteria air pollutants (ozone, PM, carbon monoxide, sulfur dioxide, nitrogen oxides, and lead). Those standards are based solely on health criteria18 without consideration of the cost and feasibility of compliance, which are reserved for later evaluation in developing state implementation plans. In this decision context, risk assessment plays a role in setting the NAAQS and in the RIAs generally used to evaluate control strategies for criteria air pollutants. Strategic Planning, Priority-Setting, and Data Development Scientifically informed strategic planning is critical. Reliable and relevant scientific data are major determinants of the quality of any risk assessment. As a result, the availability of such data strongly influences the agencyâs ability to improve its assessments in line with new methods, statutory directives, or advisory-body recommendations. In turn, the scientific quality and timeliness of reliable data depend in part on factors common to scientific work in general, such as the availability of methods and data needed to complete the assessment of any particular chemical. Near-term examples include emerging data and methods to understand modes of action that contribute to clarifying and reducing uncertainty in risk assessments. Another example is related to current studies of the use of new genomics and nanotechnology data and methods for environmental risk assessment. In addition, and separate from state-of-the-science questions, data availability depends on congressional and White House subject-matter interests that determine budget priorities for annual and long-range data development. Examples include a 12-year congressional earmark for PM research and chemical-specific allocations or directives related to arsenic. At a different level, agencywide strategic planning, priority-setting, and budgeting processes determine how risk-assessment resources are allocated among EPA programs (for example, air vs water vs IRIS), entities (external grants vs EPA laboratories), practices (basic research vs routine monitoring), and prospective risk assessments (for example, dioxin vs arsenic vs a particular Superfund site). Decisions on those issues are part of the annual planning and budgeting process, which involves scientists and managers with risk-assessment responsibilities in ORD laboratories 17â The original list of hazardous air pollutants (HAPs) contained 189 compounds; however, caprolactam (see 61 Fed. Reg. 30816 ) and methyl ethyl ketone (see 70 Fed. Reg. 75047 ) were later delisted, reducing the number of HAPs to 187. 18â he statute also calls for âsecondary,â or welfare standards to protect the environment and property. T
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 53 and program and regional offices. The resulting budget and subject-matter priorities are crucial in the availability or nonavailability of relevant data for risk-assessment purposes and thus in the quality of agency risk assessments. Although changes in budget allocations and priorities have resulted in more funding in such fields as computational toxicology and nanotechnology and less funding for postdoctoral research fellowships and intramural and extramural research, the fact remains that, in real dollar terms, EPAâs research and develop- ment funding is nearly unchanged since at least 1990, and has been steadily declining since fiscal year 2004 (Coull 2007). The resulting budget and subject-matter priorities also influ- ence the availability and workload of scientists who have the risk-assessment experience needed to study issues raised in the statement of task.19 Extramural Influences and Participants Executive Orders: Risk-Assessment Policy As indicated above, congressional legislation determines the broad outlines of risk-as- sessment principles and practices. The White House influences the process through executive orders addressing diverse risk-assessment topics and activities. Executive orders directing EPA (and other agencies) to expand the scope of their risk-assessment programs to cover cu- mulative risks20 and childrenâs risks,21 in combination with related congressional legislation, led to new emphases as to data collection and approaches to risk analysis.22 Furthermore, such provisions as Section 3-301(a) in Executive Order 12898 on environmental justice are highly specific as to the kind of data required: Environmental health research, whenever practicable and appropriate, shall include diverse segments of the population in epidemiological and clinical studies, including segments at high risk from environmental hazards, low income populations, and workers who may be exposed to substantial environmental hazards. Historically, Office of Management and Budget (OMB) oversight of EPA regulatory activities has focused on planning and budget, congressional directives and priorities, cost- benefit issues, and related administrative and accountability matters. In recent years, OMB has greatly expanded its involvement in risk-assessment practices to include governmentwide information-quality guidelines (67 Fed. Reg. 8452 ), an âInformation Quality Bul- letin for Peer Reviewâ (70 Fed. Reg. 2664 ), a âProposed Risk Assessment Bulletinâ (OMB 2006), and a memorandum on âUpdated Principles for Risk Analysisâ (OMB/OSTP 2007). The present committee did not assess the impact of OMB oversight on EPA risk assessment.23 19â uch S advisory bodies as the National Research Council, the National Science Foundation, EPAâs SAB, and EPAâs Board of Scientific Counselors regularly review and comment on EPAâs research priorities, both annual and for long-term strategic planning. See, for example, NRC 1998, 1999a, 2000b, 2001a, 2004; and www.EPA.gov/SAB. 20â rom Executive Order 12898 (February 11, 1994): âEnvironmental health analysis, whenever practicable and F appropriate, shall identify multiple and cumulative exposures.â 21â rom Executive Order 13045 (April 21, 1997): Federal agencies âshall make it a high priority to identify and F assess environmental health risks and safety risks that may disproportionately affect children.â 22â ndeed, these executive orders led to the creation of the EPA Office of Environmental Justice and, later, the I Office of Childrenâs Health Protection. 23â MB and several government agencies asked the National Research Council to review the âProposed Risk O Assessment Bulletin.â In its report (NRC 2007), the review committee lauds the goal of increasing the quality and objectivity of risk assessment in the federal government, but âconcludes that the OMB bulletin is fundamentally flawed and recommends that it be withdrawnâ (p. 6).
54 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT In sum, many factorsâstatutory requirements, the diverse array of environmental prob- lems and agency programs, executive orders, OMB directives, and the vagaries of the risk- assessment processâgive rise to risk-assessment practices and individual assessments that differ in form, information content, and analytic quality. Such diversity demands informed and experienced attention to managing the process. Executive Orders: Regulatory Policy Several executive orders illuminate the role of the White House in risk management and regulatory decision-making. Described as a âcornerstone of White House administra- tive policyâ (OMB Watch 2002), Executive Order 12866 (October 4, 1993)24 calls for each agency head to designate a regulatory-policy officer and outlines requirements related to risk assessment, cost-benefit analysis, performance-based regulatory standards, and other aspects of regulation development. A recent amendment, Executive Order 13422 (Jan. 18, 2007), requires the regulatory-policy officer to be a presidential appointee. The present committee did not assess the impact of those and other executive orders on EPA risk assessment. Public Participation EPA relies on information from the public in developing both general principles and risk assessments of individual chemicals. By law, EPA, like other federal agencies, is required to publish proposed regulations (including any underlying scientific analysis) in the Federal Register, invite public comments, and consider the comments in its final decision. EPA often follows that process for guidance documents that apply only internally (for example, risk- assessment guidelines) and for preliminary analyses used in rule-making. In addition, sepa- rately from the peer-review activities discussed above, the agency often convenes scientific experts to discuss strategic planning and research priorities and to introduce and develop background documents. Notice is given in the Federal Register, and the public is invited to observe and comment during the session. Public meetings, workshops, and the notice and comment process are avenues for stakeholders to present risk-assessment-relevant information and opinion. One example is the Pesticide Program Dialogue Group, a forum established in 1995 for a diverse group of stakeholders to provide feedback on issues from nonanimal testing to endangered spe- cies to risk assessment. The group includes pesticide manufacturers, public-interest and advocacy groups, and trade associations. It is one of several groups on pesticide issues, with corresponding groups in other agency offices, such as those which involve air-program consultation with state and local air-pollution programs and waste-office consultation with responsible parties and community groups regarding Superfund sites. EPA regional offices work closely with the Indian tribes on selected issues. Thus, EPA expressly solicits informa- tion from interested and knowledgeable parties, whether scientists or nonscientists. EPAâs statement of task anticipates near-term and long-term improvements in risk as- sessment as a result of the present report. New approaches can be expected to require ad- 24â Executive Order 12866 replaces and extends Executive Orders 12291 and 12498, issued during the Reagan administration. It directs federal regulatory agencies, including EPA, to âassess both the costs and the benefits of the intended regulation and, recognizing that some costs and benefits are difficult to quantify, propose or adopt a regulation only upon a reasoned determination that the benefits of the intended regulation justify its costsâ [Sec. (b)(1)]. The order requires EPA to conduct a formal RIA for proposed regulations expected to impose economic costs in excess of $100 million per year.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 55 BOX 2-9 Risk-Assessment Planning: Multiple Participants The committee that produced Understanding Risk (NRC 1996) identified several criteria for judg- ing success at the end of the process: getting the science right, getting the right science, getting the participation right, getting the right participation, and developing an accurate, balanced, and informative synthesis. As discussed below (Chapter 3), achieving those objectives depends in part on informed âplanning and scopingâ activities involving risk assessors, risk managers, and interested and affected parties. The emphasis on the ârightâ participants as well as the ârightâ science is important (McGarity 2004): There is little evidence that the scientific information that the agencies are currently using and disseminating is unreliable. Virtually all of the challenges that have been filed so far under the [2004 Information Quality Act] have involved disputes over interpretations, inferences, models and similar policy issues, and not the âsoundnessâ of the underlying data. justments of agency processes for allocating funds, scheduling research, expanding training, and other activities. New methods may also require enhanced peer review and expanded public participation to ensure that affected and interested parties in and outside the regu- lated community have an opportunity to contribute to new approaches and are prepared for change (see Box 2-9). Peer Review, Quality Control, and Advisory Committees Quality-control and peer-review procedures are particularly important when new ap- proaches are introduced into the risk-assessment process. EPA uses several mechanisms to ensure the quality and relevance of laboratory and field data. In addition to general methods and guidelines, including uniform guidance applicable to all federal agencies, the major pro- grams have program-specific methods related to, for example, air emissions, microbiologic contaminants, and underground storage tanks (EPA 2007c). Similarly, EPAâs peer-review program gives attention to new approaches and individual risk assessments. For example, a subcommittee of EPAâs SAB monitored the development of EPAâs first guidelines for ecologic risk assessment. Of course, assessments of individual chemicals based on new methods are subject to statutory requirements for peer review, such as the CAA requirement for review of the scientific basis of national ambient air quality standards and the FIFRA requirement for EPAâs Scientific Advisory Panel (SAP) review of the scientific basis of some pesticide decisions. Other statutes require SAB review of a wide variety of analyses (see Box 2-10).25 Independent advisory committees that provide information and advice on special topics may contribute to new approaches. In addition to advisory committees required by statute, 25â n response to recommendations from the EPA SAB and others (EPA 1992d), EPA peer-review policies issued I in 1992 call for external review of scientific assessments not subject to statutory requirements. The processes were reinforced and augmented (and in some ways redefined) by OMBâs 2002 governmentwide directive on peer re- view applicable to all federal agencies (67 Fed. Reg. 8452 ). EPA risk assessments and underlying scientific analyses are also peer-reviewed when laboratory scientists, as well as those in program and regional offices, publish work developed for risk-assessment use in scholarly journals. That work includes individual laboratory or field studies on toxicology, epidemiology, and monitoring and subunits of risk assessment, such as hazard identification and exposure analysis.
56 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT BOX 2-10 After Peer Review Peer review is not an end in itself. Ideally, peer review identifies deficiencies, suggests modifica- tions, and otherwise leads the agency to improve a risk assessment to conform more fully with scientific standards and to guide decision-making and support regulatory standards. Two situations invite inquiry and attention because, while enhancing the assessment, they also cause delays and add costs to the risk-assessment process. â¢ Peer-review âspiralsâ involve repeated reviews that return assessments to the agency for further revision because the agency has not responded adequately to science-based recommendations in earlier reviews or because of science-policy debates or inadequacies in the peer-review process itself (GAO 2001). Recent examples include the reviews of dioxin and the cancer risk-assessment guidelines (see 68 Fed. Reg. 39086 ; EPA 2005a; NRC 2006). â¢ Some assessments fail to reach closure or completion within a typical period after peer review. An example of such an unfinished assessment is that of dichloromethane (methylene chloride), which was peer-reviewed by the SAB in 1987; the health assessments remain in draft form (EPA 1987b,c), and the SAB comments have never been incorporated (EPA 2003e). The EPA assessment (EPA 1987b,c) at the time was regarded as a good example of the use of pharmacokinetic modeling. Specifically, the SAB review stated (EPA SAB 1988, p. 1) that âthe Subcommittee concludes that the Addendum [EPA 1987c] was one of the best documents it has reviewed in terms of its clarity, coverage of the data and analysis of scientific issues. This document clearly demonstrates the potential utility of pharmacokinetic data in risk assessment. EPA should continue to use this approach in future risk assessments, when- ever scientifically possible.â A confluence of factors may explain extended timeframes and unfinished assessments, including scientific complexity and controversy, a continually evolving database, and stakeholder and advocacy- group demands. Contributing factors in the case of dichloromethane were the absence of strong regulatory pressure for the assessment; the increasing importance of other chemicals, including tri- chloroethylene and tetrachloroethylene; and the replacement of dichloromethane with substitutes (L. Rhomberg, Gradient Corporation, Cambridge, MA, personal commun., May 31, 2007). EPA is scheduled to update the IRIS value for dichloromethane in the middle of 2009 (Risk Policy Report 2007; 40 CFR Part 63 ). such as the SAB and SAP, EPA has chartered committees to provide advice on selected issues pertinent to risk assessment, such as research planning and priorities (the Board of Scientific Counselors), endocrine-disrupting chemicals (the National Committee on Endocrine Disrupt- ing Chemicals and Toxic Substances), and childrenâs health (the Childrenâs Health Protection Advisory Committee) (www.EPA.gov). International Organizations EPA consults and collaborates with programs associated with the risk-assessment arms of numerous international organizations. EPA scientists sit on numerous international com- mittees including the IPCS, the International Agency for Research on Cancer (IARC)/WHO, the International Commission on Radiological Protection, and the Intergovernmental Forum on Chemical Safety; participate in the writing of scholarly papers; and conduct risk-assess- ment training in conjunction with these international organizations. As with state and local regulatory bodies, EPA and these organizations share scientific data, exchange information on developments in risk assessment, and work to harmonize risk-assessment concepts and
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 57 guidelines. Those interactions provide opportunities for EPA scientists to be alert to advances made in the organizations that will contribute to new approaches under way in EPA. In sum, several mechanisms are available to inform and upgrade EPA risk-assessment processes. Beyond the basic procedures outlined above, complementary planning and over- sight activities make it clear that the risk-assessment enterprise involves more than its basic scientific elements. Numerous overarching factorsâtangible and intangible, scientific and nonscientificâshape the process and influence the quality of agency assessments. Conclusions and Recommendations Congressional mandates give EPA a diverse set of risk-assessment and regulatory re- sponsibilities. The process is informed by many factors, including congressional legislation, generic guidance, and advice from scientific advisory bodies, peer-review recommendations specific to individual risk assessments and guidelines, information from stakeholders and other interested parties, and the principle of comity with other government entities (state, local, and international) on risk-assessment issues. The result is a complex set of risk-assess- ment activities that have drawn high praise in many cases and sustained criticism in others. The process recommendations below identify institutional and management issues that require sustained attention by agency leadership. Except for the longer timeframe expected for new guidelines (see final recommendation), the committee contemplates implementation in the immediate and near future. Conclusions â¢ Some deficiencies in current EPA risk-assessment practices can be attributed in part to the unavailability of relevant data and methods. Those limitations head the list of EPA concerns about implementing future recommendations for improvement (Appendix E). Implementing several of the recommendations in the present report will require additional data and methods related to each of the three analytic fields in the Red Book paradigm. In addition, new kinds of data or methods will be required to enable EPA to undertake analyses that are given new emphasis or recommended for the first time here. â¢ Although EPA has a 20-year history of issuing guidelines and other reports designed to implement recommendations for improvement offered by the National Research Council and other advisory bodies, moving from policy to practice has in some cases been incomplete or only partially effective (as to provisions put into practice) and in others uneven (as to use for all assessments in all parts of the agency, where applicable). â¢ Effective use of new methods and attention to new policies require instruction and training for both experienced risk assessors and newcomers. And putting new policies and methods into practiceâthat is, moving beyond policy documentsârequires understanding and appreciation on the part of agency managers and decision-makers. â¢ Historically, guideline development in EPA has taken from as little as 3 years to more than 15 years (for example, the cancer guidelines were issued in 2005 after a 15-year development period). Improvements in risk assessment will involve issuing new guidelines, revising existing guidelines or issuing supplemental guidance, and implementing existing guidelines more effectively.
58 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT Recommendations â¢ The committee seconds the Government Accountability Office recommendation that the administrator of the Environmental Protection Agency direct agency offices to âmore proactively identify the data most relevant to the current risk assessment needs, including the specific studies required and how those studies should be designed, and communicate those needs to the research communityâ (GAO 2006, p. 69). The committee recommends that the Environmental Protection Agency consider recommendations in the present report as part of that process. â¢ Putting recommendations from this report into practice will require additional staff in fields that are now lightly staffed (for example, epidemiology and quantitative uncertainty analysis) and new staff in fields that are generally understaffed relative to this reportâs em- phasis on the social-science components of environmental decision-making (for example, psychology, sociology, economics, and decision theory). â¢ Agency leaders should give high priority to establishing and maintaining risk-assess- ment and decision-making training programs for scientists, managers responsible for risk- assessment activities, and other participants in the process. This reinforces the Government Accountability Office recommendation that the Administrator of the Environmental Protec- tion Agency âensure that risk assessors and risk managers have the skills needed to produce quality risk assessments by developing and implementing in-depth trainingâ (GAO 2006, p. 69). A regular schedule of refresher courses is critical for such a program. This recommen- dation calls for training to ensure that all relevant managers and decision-makers are fully informed on risk-assessment principles and principles related to the other disciplines (such as economics and engineering) that, with risk assessment, influence regulatory decisions. â¢ To reduce the effects of the compartmentalization resulting from the Environmental Protection Agencyâs organization around diverse statutory mandates, the administrator can buttress the scientific talent brought to bear on improvement activities by revitalizing and expanding interoffice and interagency collaboration through existing structures (for example, the Risk Assessment Forum, the Science Policy Council, and the National Science and Tech- nology Council Committee on Environment and Natural Resources) and by joining scientists from other agencies (for example, the National Institute of Environmental Health Sciences and the Food and Drug Administration) in these activities. This reinforces the Government Accountability Office recommendation that the administrator of the Environmental Protec- tion Agency âdevelop a strategy to ensure that offices engage in early planning to identify and seek the expertise needed, both within the EPA workforce and from external subject matter expertsâ (GAO 2006, p. 69). â¢ The administrator of the Environmental Protection Agency should give special at- tention to expanding the scientific and decision-making core in the regional offices to ensure that they have the capacity to use improved risk-assessment methods and to meet their ob- ligations for interaction with stakeholders, local agencies, and tribes. â¢ The Environmental Protection Agency should establish a tiered schedule for guide- line implementation: (1) immediate and uniform use and oversight as to existing guidelines and risk-assessment policies (for example, 1-2 years), except where inapplicable; a shorter- term schedule for revision or updating of existing guidelines where appropriate (for example, 2-6 years); and a longer-term but definite schedule for development and issuance of new guidelines (for example, 6-15 years).
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 59 REFERENCES Blount, B.C., J.L. Pirkle, J.D. Osterloh, L. Valentin-Blasini, K.L. Caldwell. 2006. Urinary perchlorate and thyroid hormone levels in adolescent and adult men and women living in the United States. Environ. Health Perspect. 114(12):1865-1871. CalEPA (California Environmental Protection Agency). 2004. Public Health Goal for Chemicals in Drinking Water, Perchlorate. Office of Environmental Health Hazard Assessment, California Environmental Protection Agency [online]. Available: http://www.oehha.ca.gov/water/phg/pdf/finalperchlorate31204.pdf [accessed August 25, 2004]. Coull, B.C. 2007. Testimony of Dr. Bruce C. Coull, President of the U.S. Council of Environmental Deans and Directors, National Council for Science and the Environment, Before the Subcommittee on Energy and Envi- ronment, Committee on Science and Technology, U.S. House of Representatives. March 14, 2007. Crump, K.S., T. KjellstrÃ¶m, A.M. Shipp, A. Silvers, and A. Stewart. 1998. Influence of prenatal mercury exposure upon scholastic and psychological test performance: Benchmark analysis of a New Zealand cohort. Risk Anal. 18(6):701-713. Crump, K., and G. Goodman. 2003. Benchmark Analysis for the Perchlorate Inhibition of Thyroidal Radioiodine Uptake Utilizing a model for the Observed Dependence of Uptake and Inhibition on Iodine Excretion. Pre- pared for J. Gibbs, Kerr-McGee Corporation. January 24, 2003. (Presentation at the Fifth Meeting on Assess the Health Implications of Perchlorate Ingestion, July 29-30, 2004, Washington, DC.) Davidson, P.W., G.J. Myers, C. Cox, C. Shamlaye, O. Choisy, J. Sloane-Reeves, E. Cernchiari, D.O. Marsh, M. Berlin, M. Tanner, and T.W. Clarkson. 1995a. Neurodevelopmental test selection, administration, and perfor- mance in the main Seychelles child development study. Neurotoxicology 16(4):665-676. Davidson, P.W., G.J. Myers, C. Cox, C.F. Shamlaye, D.O. Marsh, M.A. Tanner, M. Berlin, J. Sloane-Reeves, E. Cernichiari, O. Choisy, A. Choi, and T.W. Clarkson. 1995b. Longitudinal neurodevelopmental study of Sey- chellois children following in utero exposure to methylmercury from maternal fish ingestion: Outcomes at 19 and 29 months. Neurotoxicology 16(4):677-688. Davidson, P.W., G.J. Myers, C. Cox, C. Axtell, C. Shamlaye, J. Sloane-Reeves, E. Cernichiari, L. Needham, A. Choi, Y. Wang, M. Berlin, and T.W. Clarkson. 1998. Effects of prenatal and postnatal methylmercury exposure from fish consumption on neurodevelopment: Outcomes at 66 months of age in the Seychelles child development study. JAMA 280(8):701-707. EPA (U.S. Environmental Protection Agency). 1984. Risk Assessment and Management: Framework for Decision Making. EPA 600/9-85-002. Office of the Administrator, U.S. Environmental Protection Agency, Washington, DC. December 1984. EPA (U.S. Environmental Protection Agency). 1986a. Guidelines for Carcinogen Risk Assessment. EPA/630/R- 00/004. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. September 1986 [online]. Available: http://www.epa.gov/ncea/raf/car2sab/guidelines_1986.pdf [accessed Jan. 7, 2008]. EPA (U.S. Environmental Protection Agency). 1986b. Guidelines for Mutagenicity Risk Assessment. EPA/630/ R-98/003. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. September 1986 [online]. Available: http://www.epa.gov/osa/mmoaframework/pdfs/MUTAGEN2.PDF [accessed Jan. 7, 2008]. EPA (U.S. Environmental Protection Agency). 1986c. Guidelines for Health Risk Assessment of Chemical Mix- tures. EPA/630/R-98/002. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. September 1986 [online]. Available: http://www.epa.gov/ncea/raf/pdfs/chem_mix/chemmix_1986.pdf [accessed Jan. 7, 2008]. EPA (U.S. Environmental Protection Agency). 1987a. Unfinished Business: A Comparative Assessment of Envi- ronmental Problems. Office of Policy Analysis, Office of Policy Planning and Evaluation, U.S. Environmental Protection Agency, Washington, DC. EPA (U.S. Environmental Protection Agency). 1987b. Technical Analysis of New Methods and Data Regard- ing Dichloromethane: Pharmacokinetics, Mechanism of Action and Epidemiology. External Review Draft. EPA/600/8-87/029A. PB-87-228557/XAB. Office of Health and Environmental Assessment, U.S. Environmen- tal Protection Agency, Washington, DC. June 1, 1987. EPA (U.S. Environmental Protection Agency). 1987c. Update to the Health Assessment Document and Addendum for Dichloromethane (Methylene Chloride): Pharmacokinetics, Mechanism of Action and Epidemiology. Re- view Draft. EPA/600/8-87/030A. PB-87-228565/XAB. Office of Health and Environmental Assessment, U.S. Environmental Protection Agency, Washington, DC. July 1, 1987. EPA (U.S. Environmental Protection Agency). 1991. Guidelines for Developmental Toxicity Risk Assessment. EPA/600/FR-91/001. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. De- cember 1991 [online]. Available: http://www.epa.gov/NCEA/raf/pdfs/devtox.pdf [accessed Jan. 10, 2008].
60 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT EPA (U.S. Environmental Protection Agency). 1992a. Guidelines for Exposure Assessment. EPA600Z-92/001. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. EPA (U.S. Environmental Protection Agency). 1992b. Framework for Ecological Risk Assessment. EPA/63-R- 92/001. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. EPA (U.S. Environmental Protection Agency). 1992c. Guidance on Risk Characterization for Risk Managers and Risk Assessors. Memorandum to Assistant Administrators, and Regional Administrators, from F. Henry Habicht, Deputy Administrator, Office of the Administrator, Washington, DC. February 26, 1992 [online]. Available: http://www.epa.gov/oswer/riskassessment/pdf/habicht.pdf [accessed Oct. 10, 2007]. EPA (U.S. Environmental Protection Agency). 1992d. Safeguarding the Future: Credible Science, Credible Deci- sions. The Report of an Expert Panel on the Role of Science at EPA. EPA/600/9-91/050. U.S. Environmental Protection Agency, Washington, DC. March 1992. EPA (U.S. Environmental Protection Agency). 1994. Peer Review and Peer Involvement at the U.S. Environmental Protection Agency. Memorandum to Assistant Administrators, General Counsel, Inspector General, Associate Administrators, Regional Administrators, and Staff Office Directors, from Carol M. Browner, Administrator, U.S. Environmental Protection Agency. June 7, 1994 [online]. Available: http://www.epa.gov/osa/spc/pdfs/ perevmem.pdf [accessed Oct. 16, 2007]. EPA (U.S. Environmental Protection Agency). 1995. Guidance for Risk Characterization. Science Policy Council, U.S. Environmental Protection Agency, Washington, DC. February 1995 [online]. Available: http://www. epa.gov/osa/spc/pdfs/rcguide.pdf [accessed Jan. 7, 2008]. EPA (U.S. Environmental Protection Agency). 1996a. Guidelines for Reproductive Toxicity Risk Assessment. EPA/630/R-96/009. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. Octo- ber 1996 [online]. Available: http://www.epa.gov/ncea/raf/pdfs/repro51.pdf [accessed Jan. 10, 2008]. EPA (U.S. Environmental Protection Agency). 1996b. Proposed Guidelines for Carcinogen Risk Assessment. EPA/600/P-92/003C. Office of Research and Development, U.S. Environmental Protection Agency, Wash- ington, DC. April 1996 [online]. Available: http://www.epa.gov/NCEA/raf/pdfs/propcra_1996.pdf [accessed Jan. 7, 2007]. EPA (U.S. Environmental Protection Agency). 1997a. Guiding Principles for Monte Carlo Analysis. EPA/630/R- 97/001. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. March 1997 [online]. Available: http://www.epa.gov/ncea/raf/montecar.pdf [accessed Jan. 7, 2008]. EPA (Environmental Protection Agency). 1997b. Mercury Study Report to Congress, Volume 1. Executive Sum- mary. EPA-452/R-97-003. Office of Air Planning and Standards and Office of Research and Development, U.S. Environmental protection Agency, Washington, DC. December 1997 [online]. Available: http://www.epa. gov/ttn/oarpg/t3/reports/volume1.pdf [accessed Jan. 16, 2008]. EPA (U.S. Environmental Protection Agency). 1997c. Exposure Factors Handbook. National Center for Environ- mental Assessment, Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC. August 1997 [online]. Available: http://www.epa.gov/ncea/efh/report.html [accessed Aug. 5, 2008]. EPA (U.S. Environmental Protection Agency). 1998a. Guidelines for Ecological Risk Assessment. EPA/630/R- 95/002F. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. April 1998 [online]. Available: http://oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=36512 [accessed Feb. 9, 2007]. EPA (U.S. Environmental Protection Agency). 1998b. Guidelines for Neurotoxicity Risk Assessment. EPA/630/ R-95/001F. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. April 1998 [online]. Available: http://www.epa.gov/ncea/raf/pdfs/neurotox.pdf [accessed Jan. 10, 2008]. EPA (U.S. Environmental Protection Agency). 2000a. Supplementary Guidance for Conducting Health Risk As- sessment of Chemical Mixtures. EPA/630/R-00/002. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. August 2000 [online]. Available: http://www.epa.gov/ncea/raf/pdfs/chem_mix/ chem_mix_08_2001.pdf [accessed Jan. 7, 2008]. EPA (U.S. Environmental Protection Agency). 2000b. Risk Characterization Handbook. EPA 100-B-00-002. Office of Science Policy, Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC. December 2000 [online]. Available: http://www.epa.gov/OSA/spc/pdfs/rchandbk.pdf [accessed Jan. 7, 2008]. EPA (U.S. Environmental Protection Agency). 2000c. Peer Review Handbook, 2nd Ed. EPA 100-B-00-001. Science Policy Council, U.S. Environmental Protection Agency, Washington, DC. December 2000 [online]. Available: http://www.epa.gov/osa/spc/pdfs/prhandbk.pdf [accessed May 16, 2008]. EPA (U.S. Environmental Protection Agency). 2002a. Guidance on Cumulative Risk Assessment of Pesticide Chemicals That Have a Common Mechanism of Toxicity. Office of Pesticide Programs, U.S. Environmental Protection Agency, Washington, DC. January 14, 2002 [online]. Available: http://www.epa.gov/pesticides/trac/ science/cumulative_guidance.pdf [accessed Jan. 7, 2008].
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 61 EPA (U.S. Environmental Protection Agency). 2002b. Guidelines for Ensuring and Maximizing the Quality, Utility and Integrity of Information Disseminated by the Environmental Protection Agency. EPA/260R-02-008. Office of Environmental Information, U.S. Environmental Protection Agency, Washington, DC. October 2002 [on- line]. Available: http://www.epa.gov/QUALITY/informationguidelines/documents/EPA_InfoQualityGuidelines. pdf [accessed Feb. 9, 2007]. EPA (U.S. Environmental Protection Agency). 2003a. Human Health Research Strategy. EPA/600/R-02/050. Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://www.epa.gov/nheerl/humanhealth/HHRS_final_web.pdf [accessed July 31, 2008]. EPA (U.S. Environmental Protection Agency). 2003b. Exposure and Human Health Reassessment of 2,3,7,8- Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. NAS Review Draft. National Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC. December 2003 [online]. Available: http://www.epa.gov/NCEA/pdfs/dioxin/nas-review/ [accessed Jan. 9, 2008]. EPA (U.S. Environmental Protection Agency). 2003c. Disposition of Comments and Recommendations for Re- visions to âPerchlorate Environmental Contamination: Toxicological Review and Risk Characterization, External Review Draft (January16, 2002).â National Center for Environmental Assessment, Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://cfpub2.epa.gov/ ncea/cfm/recordisplay.cfm?deid=72117 [accessed Jan 4, 2008]. EPA (U.S. Environmental Protection Agency). 2003d. Framework for Cumulative Risk Assessment. EPA/600/ P-02/001F. National Center for Environmental Assessment, Risk Assessment Forum, U.S. Environmen- tal Protection Agency, Washington, DC [online]. Available: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm? deid=54944 [accessed Jan. 4, 2008]. EPA (U.S. Environmental Protection Agency). 2003e. Dichloromethane. (CASRN 75-09-2). Integrated Risk Infor- mation System, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://www.epa. gov/NCEA/iris/subst/0070.htm [accessed Jan. 10, 2008]. EPA (U.S. Environmental Protection Agency). 2004a. Risk Assessment Principles and Practices: Staff Paper. EPA/100/B-04/001. Office of the Science Advisor, U.S. Environmental Protection Agency, Washington, DC. March 2004 [online]. Available: http://www.epa.gov/osa/pdfs/ratf-final.pdf [accessed Jan. 9, 2008]. EPA (U.S. Environmental Protection Agency). 2004b. Risk Assessment and Modeling-Air Toxics Risk Assess- ment Library, Vol.1. Technical Resources Manual, Part III. Human Health Risk Assessment: Multipathway. EPA-453-K-04-001A. Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA (U.S. Environmental Protection Agency). 2005a. Guidelines for Carcinogen Risk Assessment. EPA/630/P- 03/001F. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, DC. March 2005 [online]. Available: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=116283 [accessed Feb. 7, 2007]. EPA (U.S. Environmental Protection Agency). 2005b. Supplemental Guidance for Assessing Susceptibility for Early-Life Exposures to Carcinogens. EPA/630/R-03/003F. Risk Assessment Forum, U.S. Environmental Pro- tection Agency, Washington, DC. March 2005 [online]. Available: http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=160003 [accessed Jan. 4, 2008]. EPA (U.S. Environmental Protection Agency). 2005c. Perchlorate and Perchlorate Salts: Reference Dose for Chronic Oral Exposure. Integrated Risk Information System, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://www.epa.gov/iris/subst/1007.htm [accessed Jan. 10, 2008]. EPA (U.S. Environmental Protection Agency). 2005d. Science Inventory (SI) Database. U.S. Environmental Protec- ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ tion Agency, Washington, DC [online]. Available: http://cfpub.epa.gov/si/ [accessed Jan. 11, 2008]. EPA (U.S. Environmental Protection Agency). 2006. Framework for Assessing Health Risks of Environmental Exposures to Children (External Review Draft). EPA/600/R-05/093A. National Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC. March 2006 [online]. Available: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=150263 [accessed Oct. 11, 2007]. EPA (U.S. Environmental Protection Agency). 2007a. Total Risk Integrated Methodology (TRIM) â General Information. Office of Air and Radiation, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://www.epa.gov/ttn/fera/trim_gen.html [accessed Feb. 22, 2008]. EPA (U.S. Environmental Protection Agency). 2007b. Explanation of Statutory Framework for Risk Benefit Bal- ancing for Public Health Pesticides. Public Health Issues, Office of Pesticides, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://earth1.epa.gov/pesticides/health/risk-benefit.htm [accessed Jan. 11, 2008]. EPA (U.S. Environmental Protection Agency). 2007c. Environmental Test Methods and Guidelines. Information Sources, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://www.epa.gov/ Standards.html [accessed Jan. 11, 2007].
62 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT EPA (U.S. Environmental Protection Agency). 2008. IRIS Glossary. Integrated Risk Information System National Center for Environmental Assessment, U.S. Environmental Protection Agency, Washington, DC [online]. Available: http://www.epa.gov/ncea/iris/help_gloss.htm#e [accessed Feb. 28, 2008]. EPA SAB (U.S. Environmental Protection Agency Science Advisory Board). 1988. Assess Health Effects Associated with Dichloromethane (Methylene Chloride). Final report. SAB-EHC-88-013. PB-89-108997/XAB. U.S. En- vironmental Protection Agency, Science Advisory Board, Washington, DC. May 9, 1988 [online]. Available: http://yosemite.epa.gov/sab/sabproduct.nsf/0708B8BE9D86939685257328005CF44F/$File/DICHLOROME THANE+++++++EHC-88-013_88013_5-22-1995_264.pdf [accessed Jan. 11, 2008]. EPA SAB (U.S. Environmental Protection Agency Science Advisory Board). 1995. SAB Review of Issues Related to the Regulation of Arsenic in Drinking Water. EPA-SAB-DWC-95-015. U.S. Environmental Protection Agency Science Advisory Board, Washington, DC. July 1995 [online]. Available: http://yosemite.epa.gov/sab/ sabproduct.nsf/D3435DF2A691B8328525719B006A500E/$File/dwc95015.pdf [accessed July 31, 2008]. GAO (Government Accountability Office). 2001. EPAâs Science Advisory Board Panels. Improved Policies and Pro- cedures Needed to Ensure Independence and Balance. Report to the Ranking Minority Member, Committee on Government Reform, House of Representatives. GAO-01-536. Washington, DC: Government Accountability Office. June 2001 [online]. Available: http://www.gao.gov/new.items/d01536.pdf [accessed July 31, 2008]. GAO (U.S. Government Accountability Office). 2005. Chemical Regulation: Options Exist to Improve EPAâs Abil- ity to Assess Health Risks and Manage Its Chemical Review Program. GAO-05-458. Washington, DC: U.S. Government Accountability Office [online]. Available: http://www.gao.gov/new.items/d05458.pdf [accessed Jan. 10, 2008]. GAO (U.S. Government Accountability Office). 2006. Human Health Risk Assessment: EPA Has Taken Steps to Strengthen Its Process, but Improvements Needed in Planning, Data Development, and Training. GAO- 06-595. Washington, DC: U.S. Government Accountability Office [online]. Available: http://www.gao.gov/ new.items/d06595.pdf [accessed Jan. 10, 2008]. GAO (U.S. General Accountability Office). 2008. Chemical Assessments: Low Productivity and New Interagency Review Process Limit the Usefulness and Credibility of EPAâs Integrated Risk Information System. GAO-08- 440. Washington, DC: U.S. General Accountability Office. March 2008 [online]. Available: http://www.gao. gov/new.items/d08440.pdf. [accessed June 11, 2008]. Gilman, P. 2003. Statement of Paul Gilman, Assistant Administrator for Research and Development and EPA Science Advisor, U.S. Environmental Protection Agency, before the Committee on Transportation and Infra- structure, Subcommittee on Water Resources and the Environment, U.S. House of Representatives, March 5, 2003 [online]. Available: http://www.epa.gov/ocir/hearings/testimony/108_2003_2004/2003_0305_pg.pdf [accessed February 9, 2007]. Greer, M.A., G. Goodman, R.C. Pleus, and S.E. Greer. 2002. Health effects assessment for environmental perchlo- rate contamination: The dose response for inhibition of thyroidal radioiodine uptake in humans. Environ. Health Perspect. 110(9):927-937. Grandjean, P., P. Weihe, R.F. White, F. Debes, S. Araki, K. Yokoyama, K. Murata, N. SÃ¸rensen, R. Dahl, and P.J. JÃ¸rgensen. 1997. Cognitive deficit in 7-year-old children with prenatal exposure to methylmercury. Neuro- toxicol. Teratol. 19(6):417-428. Grandjean, P., P. Weihe, R.F. White, and F. Debes. 1998. Cognitive performance of children prenatally exposed to âsafeâ levels of methylmercury. Environ. Res. 77(2):165-172. Grandjean, P., E. Budtz-JÃ¸rgensen, R.F. White, P.J. JÃ¸rgensen, P. Weihe, F. Debes, and N. Keiding. 1999. Meth- ylmercury exposure biomarkers as indicators of neurotoxicity in children aged 7 years. Am. J. Epidemiol. 150(3):301-305. IOM (Institute for Medicine). 1997. Dietary Reference Intakes for Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride. Washington, DC: National Academy Press. IOM (Institute for Medicine). 1998. Dietary Reference Intakes: A Risk Assessment Model for Establishing Upper Intake Levels for Nutrients. Washington, DC: National Academy Press. IOM (Institute for Medicine). 2003. Dietary Reference Intakes: Applications in Dietary Planning. Washington, DC: The National Academies Press.Â Krewski, D., V. Hogan, M.C. Turner, P.L. Zeman, I. McDowell, N. Edwards, and J. Losos. 2007. An integrated framework for risk management and population health. Hum. Ecol. Risk Asses. 13(6):1288-1312. McGarity, T.O. 2004. Our science is sound science and their science is junk science: Science-based strategies for avoiding accountability and responsibility for risk-producing products and activities. Kans. Law. Rev. 52(4):897-937. Mills, A. 2006. IRIS from the inside. Risk Anal. 26(6):1409-1410. NRC (National Research Council). 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press. NRC (National Research Council). 1993a. Issues in Risk Assessment. Washington, DC: National Academy Press.
EVOLUTION AND USE OF RISK ASSESSMENT IN EPA 63 NRC (National Research Council). 1993b. Pesticides in the Diets of Infants and Children. Washington, DC: Na- tional Academy Press. NRC (National Research Council). 1994. Science and Judgment in Risk Assessment. Washington, DC: National Academy Press. NRC (National Research Council). 1996. Understanding Risk: Informing Decisions in a Democratic Society. Wash- ington, DC: National Academy Press. NRC (National Research Council). 1998. Research Priorities for Airborne Particulate Matter: I. Immediate Priori- ties and a Long-Range Research Portfolio. Washington, DC: National Academy Press. NRC (National Research Council). 1999a. Research Priorities for Airborne Particulate Matter: II. Evaluating Research Progress and Updating Portfolio. Washington, DC: National Academy Press. NRC (National Research Council). 1999b. Arsenic in Drinking Water. Washington DC: National Academy Press. NRC (National Research Council). 2000a. Toxicological Effects of Methyl Mercury. Washington, DC: National Academy Press. NRC (National Research Council). 2000b. Strengthening Science at the U.S. Environmental Protection Agency. Washington, DC: National Academy Press. NRC (National Research Council). 2001a. Research Priorities for Airborne Particulate Matter: III. Early Research Progress. Washington, DC: National Academy Press. NRC (National Research Council). 2001b. Arsenic in Drinking Water: 2001 Update. Washington, DC: National Academy Press. NRC (National Research Council). 2002. Estimating the Public Health Benefits of Proposed Air Pollution Regula- tions. Washington, DC: National Academy Press. NRC (National Research Council). 2004. Research Priorities for Airborne Particulate Matter: IV. Continuing Research Progress. Washington, DC: The National Academies Press. NRC (National Research Council). 2005a. Superfund and Mining Megasites/Lessons from Coeur DâAlene River Basin. Washington, DC: The National Academies Press. NRC (National Research Council). 2005b. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. NRC (National Research Council). 2006. Health Risks from Dioxin and Related Compounds/Evaluation of the EPA Reassessment. Washington, DC: The National Academies Press. NRC (National Research Council). 2007. Scientific Review of the Proposed Risk Assessment Bulletin from the Office of Management and Budget. Washington, DC: The National Academies Press. OMB (Office of Management and Budget). 2006. Proposed Risk Assessment Bulletin. Office of Management and Budget, Washington, DC. January 9, 2006 [online]. Available: http://www.whitehouse.gov/omb/inforeg/ proposed_risk_assessment_bulletin_010906.pdf [accessed Jan. 4, 2008]. OMB/OSTP (Office of Management and Budget/Office of Science and Technology Policy). 2007. Updated Prin- ciples for Risk Analysis. Memorandum for the Heads of Executive Departments and Agencies, from Susan E. Dudley, Administrator, Office of Information and Regulatory Affairs, Office of Management and Budget, and Sharon L. Hays, Associate Director and Deputy Director for Science, Office of Science and Technology Policy, Washington, DC. September 19, 2007 [online]. Available: http://www.whitehouse.gov/omb/memoranda/ fy2007/m07-24.pdf [accessed Jan. 4, 2008]. OMB Watch. 2002. Executive Order 12866. OMB Watch. February 10, 2002 [online]. Available: http://www. ombwatch.org/article/articleview/180/1/67 [accessed Jan. 11, 2008]. Parkin. 2007. Foundations and Frameworks for Microbial Risk Assessments. Presentation at the 4th Meeting on Improving Risk Analysis Approaches Used By the U.S. EPA, April 17, 2007, Washington DC. PCCRARM (Presidential/Congressional Commission on Risk Assessment and Risk Management). 1997. Frame- work for Environmental Health Risk Management - Final Report, Vol. 1. [online]. Available: http://www. riskworld.com/nreports/1997/risk-rpt/pdf/EPAJAN.PDF [accessed Jan. 4, 2008]. Risk Policy Report. 2007. EPA Says Future Studies May Force Tighter Toxics Limits for Solvents. Inside EPAâs Risk Policy Report. 14(19):14. May 8, 2007. Rosenthal, A., G.M. Gray, and J.D. Graham. 1992. Legislating acceptable cancer risk from exposure to toxic chemicals. ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ Ecol. Law Q. 19(2):269-362. Steinmaus, C., M.D. Miller, R. Howd. 2007. Impact of smoking and thiocyanate on perchlorate and thyroid hormone associations in the 2001-2002 national health and nutrition examination survey. Environ. Health Perspect. 115:1333-1338. Suter, G., T. Vermeire, W. Munns, and J. Sekizawa 2001. Framework for the Integration of Health and Ecological Risk Assessment. Chapter 2 in Integrated Risk Assessment Report. Prepared for the WHO/UNEP/ILO Inter- national Programme on Chemical Safety. WHO/IPCS/IRA/01/12. The International Programme on Chemical Safety, World Health Organization. December 2001 [online]. Available: http://www.who.int/ipcs/publications/ en/ch_2.pdf [accessed July 30, 2008].
64 SCIENCE AND DECISIONS: ADVANCING RISK ASSESSMENT Tiemann, M. 2005. Arsenic in Drinking Water: Regulatory Developments and Issues. CRS Report for Congress 05-RS-20672a. Washington, DC: Congressional Research Service. October 20, 2005 [online]. Available: http://assets.opencrs.com/rpts/RS20672_20051020.pdf [accessed Jan. 7, 2008]. WHO (World Health Organization). 2006. A Model for Establishing Upper Levels of Intake for Nutrients and Related Substances. Report of a Joint FAO/WHO Technical Workshop on Risk Assessment, May 2-6, 2005, Geneva, Switzerland. World Health Organization, Food and Agriculture Organization of the United Nations. January 13, 2006 [online]. Available: http://www.who.int/ipcs/highlights/full_report.pdf [accessed July 30, 2008].