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.
6 Framework for the Development of Policy Options to Reduce Exposure to Dioxins and Dioxin-like Compounds A central element of the committee's charge was to identify, evaluate, and recommend policy options for reducing human exposure through diet to dioxin and dioxin-like compounds (referred to collectively as DLCs). In considering DLC exposure reduction options, the committee was charged with taking into account both possible reductions in human health risk associated with reducing DLC intake and the possible nutritional consequences of dietary changes that might result from the selected measures. In response to this charge, the commit- tee has identified a number of possible exposure reduction (or risk-management) options, which are described in Chapter 7; in Chapter 8, the committee makes risk-management recommendations. One of the committee's primary conclusions, however, is that the data re- quired to evaluate and recommend a comprehensive set of risk-management interventions for DLCs are lacking. There remains substantial uncertainty about the magnitude of the risks posed by low levels of dietary exposure to DLCs and corresponding uncertainty about the magnitude of the risk reductions that are achievable by reducing such exposures. This uncertainty will be a substantial limiting factor for the government as it considers risk-management strategies, especially to the extent it considers the use of its traditional food safety regulatory tools, as discussed below. The study's sponsors, however, asked the committee not to evaluate or repeat the extensive risk assessment work that has been done on DLCs (see AEA Technology, 1999; ATSDR, 1998; EPA, 2000; Fiedler et al., 2000; IARC, 1997; and Scientific Committee on Food, 2000, 2001~. In keeping with its charge, the committee has proceeded with its identification of risk-man- agement options on the premise that reductions in DLC exposure are desirable for 150
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 151 health reasons in light of: (1) the extensive animal data demonstrating the toxicity of DLCs at very low exposure levels, (2) the cumulative effect even low levels of exposure can have on an individual's total DLC body burden, and (3) the exist- ence of vulnerable subpopulations, such as fetuses and breastfeeding infants and individuals (e.g., subsistence fishers) whose traditional diets, for cultural or eco- nomic reasons, may result in higher levels of DLC exposure. Even with the assumption that reducing dietary exposure to DLCs is desir- able, insufficient data are available to assess and support at least some of the potential risk-management options for DLCs. For example, the data on DLCs in the food supply, including the incidence, levels, and geographical distribution of DLC contamination in various animal production systems and in raw, processed, and cooked foods, are extremely limited, as discussed in Chapters 4 and 5. Such data would be required to most effectively target and prioritize risk-management interventions and support the establishment of regulatory limits on DLCs in food, if the government considered that an appropriate option for reducing DLC expo- sure. The committee also found very little data that would help to assess the feasibility and calculate the costs and other impacts of the wide range of risk- management options that are potentially available to reduce DLC exposure through the food supply. This paucity of data is particularly evident in the animal production arena, where the ecology and epidemiology of these compounds un- der normal production conditions are not available. Further examples of data limitations that affect the consideration of risk-management options are provided later in this chapter and in Chapter 7. Because of the large gaps in the data required to devise an optimal, compre- hensive risk-management strategy for DLCs and to assess specific options, the committee was unable to reach conclusions on the likely effectiveness of a num- ber of the possible risk-management options it identified, or on a relative ranking of these options. It thus focused its efforts on devising a framework for the development and evaluation of options for reducing DLC exposure. The pro- posed framework is described in this chapter. The purpose of the framework is to provide a systematic basis to identify risk-management options and to collect and analyze the data required to evaluate the options. As discussed below, the framework divides the universe of options into three categories: (1) options to reduce DLCs in animal forage and feeds, (2) options to reduce DLCs in the human food supply, and (3) options to change food-consumption patterns to reduce DLC exposure. The framework is suitable for evaluating a wide array of risk-management options, including regulatory interventions, voluntary and collaborative interventions, and nonregulatory in- centives. In identifying possible options, the committee assumed that the govern- ment would be operating within current statutory authorities; the key food safety statutory provisions and other legal tools applicable to DLCs are thus summa- rized in this chapter.
52 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY Finally, the framework addresses the committee's charge to consider the possible nutritional consequences of dietary changes that might occur as a result of measures taken to reduce exposure to DLCs. The committee recommends for this purpose an approach called "risk-relationship analysis," which takes a com- prehensive approach to assessing how measures taken to prevent or mitigate a particular risk can affect other risks to individuals or populations, whether by causing a countervailing increase in risk or by causing ancillary risk reductions. This chapter concludes with a discussion of risk-relationship analysis and how it applies to the evaluation of risk-management options for DLCs. CATEGORIZING POSSIBLE RISK-MANAGEMENT OPTIONS In developing an inventory of possible risk-management options to reduce exposure to DLCs, the committee concluded that opportunities to reduce dietary exposure fall into three major categories, which correspond to the three distinct stages of food production, processing, and consumption within the farm-retail chain. The committee's framework organizes the possible options into these three categories, each of which has its own distinct attributes and its own features that affect the feasibility and effectiveness of possible interventions. Animal Production Systems As discussed in Chapter 4, the forage and feeds consumed by food-produc- ing animals, including fish, is one of the most critical pathways for the entry of DLCs into human food and for human exposure to these compounds. It is also the most promising pathway that provides an opportunity to interdict and reduce DLC contamination of the human food supply. One major source of DLC contamination is air deposition of DLCs onto plants that are used for forage and feeds. When animals consume plants, their by- products, sediment-contaminated waters, or soils directly, DLCs are ingested and stored in the animals' fat. When humans consume the animals' fat as part of their diets, they are exposed to DLCs. The importance of animal feeds as sources of DLCs in human diets is en- hanced by the practice at abattoirs of rendering fats and animal proteins from fallen (recently dead) animals and offal not acceptable for human consumption, and using the salvaged fat in animal feeds, which is then distributed nationally in the animal feed system. This practice has the effect of recycling and potentially concentrating the DLCs that are present in animal fat because of environmental exposures, and of increasing DLC exposure to humans through the food supply. Interrupting this cycle of DLC contamination of the food supply via the animal feed pathway is, in the committee's judgment, a high-priority risk-management option.
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 153 The advantage of forage and feeds as a focal point for measures to reduce DLC exposure is that steps taken at this stage can prevent DLCs from entering human food. Such steps include limiting access to forage in areas known to be high in DLC contamination and limiting the use of recycled animal fat in animal feeds or setting direct limits on the levels of DLCs permitted in animal feeds. Measures of this kind impose costs (e.g., loss of use of land for grazing in highly contaminated areas, increased production costs) and may have other consequences that the committee believes should be considered under the proposed framework. However, any reductions in DLC contamination that are achieved at the forage and feed stage have the important benefit of directly reducing the reservoir of DLCs in food to which humans are potentially exposed. Human Food The second major category of opportunities to reduce DLC exposure identi- fied by the committee involves reducing DLC levels in the foods people con- sume. Once DLCs are present in human food or in the fat of food-producing animals, risk-management options are available, but they may be more limited in their potential to reduce DLC exposure than interventions at the forage and feed stage. In some cases, such as when DLCs are present on the exterior surfaces of fruits and vegetables through direct airborne deposition onto leaves or the pres- ence of DLC-contaminated soil that clings to peels, washing leafy surfaces or removing the outer peels from produce can reduce DLC exposure. DLCs can also be reduced in human food by trimming excess fat from cuts of meat or otherwise reducing the amount of animal fat in food. At present, processing methods to directly remove DLCs from animal foods do not exist. When the removal or reduction of DLCs through washing or other process- ing interventions is not possible, reduction of DLC exposure at the human-food stage can be achieved by limiting the levels of DLCs permitted in the foods. This would require removing foods from the food supply if they contain DLCs above some permissible level. Because human foods or food components generally have greater economic value than animal feeds or feed components, the cost issues that would have to be considered under the food safety laws or in the committee' s analytical framework could be greater. However, this would depend on a wide range of cost factors associated with intervention strategies at different points in the farm-retail supply chain. Food-Consumption Patterns Finally, the committee identified changes in individual food-consumption patterns as a vehicle for reducing DLC exposure. This involves reducing the consumption of foods or food components known or expected to be high in DLCs
54 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY and substituting foods that are likely to be lower in DLCs. Because individual food-consumption patterns are generally beyond direct regulatory control, the risk-management options in this category are likely to involve education, expan- sion of food choices, incentives, and other interventions to encourage voluntary changes in what people eat. The options for reducing DLC exposure through changes in food-consump- tion patterns are likely to vary in their feasibility and effectiveness depending on the circumstances of the individual. For example, changing consumption patterns may be very difficult for indigenous populations highly dependent on wild-caught fish and other game from areas with high environmental contamination levels. It may be relatively easy for school-age children who can reduce DLC intake by replacing whole milk with low-fat (1 percent milk fat) or skim milk. The health benefit of changing consumption patterns to reduce DLC exposure may also vary considerably depending on age and other factors. For example, adolescent girls and their future offspring might benefit substantially by reducing the animal fat content of their diets and thus their DLC intake, while the benefits for adults may not be as great. These factors should be considered in designing risk-management interventions and analyzing them in the committee's proposed framework. THE RANGE OF POSSIBLE RISK-MANAGEMENT OPTIONS As suggested by the preceding discussion and presented in more detail in Chapter 7, there is a wide range of possible risk-management options for reduc- ing DLC exposure from food. The range is defined not only by the three catego- ries or stages in the food chain that are outlined above, but also by the type of intervention. These include: (1) the use of traditional food safety regulatory tools, (2) collaborative and voluntary interventions involving food safety agencies and the food industry, (3) public policy and regulatory interventions beyond tradi- tional food safety interventions, including subsidies, economic incentives, and other measures to reduce DLCs and expand dietary options, and (4) information and education interventions. All of these have been considered by the committee in identifying possible risk-management options, and they are addressed by the committee's proposed framework. Traditional Food Safety Regulatory Tools The most common risk-management intervention by the federal government with respect to food contaminants such as DLCs is the enforcement of the food adulteration provisions in the Federal Food, Drug, and Cosmetic Act (FDCA) of 1938 and other federal food safety laws that deal specifically with meat and poultry. Jurisdiction over food safety is shared between the U.S. Food and Drug Administration (FDA), the U.S. Department of Agriculture (USDA), and the U.S.
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 155 Environmental Protection Agency (EPA). As discussed in more detail below, current laws prohibit the presence in food of substances that are potentially harmful to health, and they authorize FDA to act to prevent contamination of the food supply with such substances, either by setting and enforcing standards for the amount of such substances that can be present, or by bringing case-by-case enforcement actions against the products that the government believes are con- taminated at harmful or potentially harmful levels. FDA has used these tools to restrict the presence of mercury, lead, aflatoxin, and other naturally occurring and man-made contaminants in foods. The central theme of these and other tradi- tional food safety regulatory tools is that they involve the use of the government's legal powers to prohibit harmful contamination or practices that could result in such contamination. Collaborative and Voluntary Interventions As an alternative to mandatory regulatory interventions, the food safety agen- cies have the option to initiate collaborative programs with the food industry dependent on the willingness of the industry to act voluntarily to reduce a hazard. FDA could, for example, work with animal producers to develop guidance or voluntary codes of practice that producers could follow to reduce DLCs in animal forage or feeds, or the government could provide technical or other support for industry-sponsored DLC testing programs. Such collaborative and voluntary measures assume that the industry has its own interest or incentive for taking action to reduce DLC levels in foods. In such cases, they may be viable risk- management options, especially if the data required to support the use of tradi- tional regulatory tools are unavailable. Subsidies, Economic Incentives, and Other Measures to Reduce DLCs and Expand Dietary Options Beyond the traditional regulatory or cooperative approaches to reducing DLCs, there are a number of other tools of public policy that, though rarely used to address food safety hazards, are at least theoretically available to the govern- ment to foster reductions of DLCs. These could include, for example, subsidies or other economic incentives, such as taxes, to induce reductions in the use of animal fat in animal feeds; development of alternative uses for animal fats other than feeds (e.g., biofuels); or the reduction of the fat content of processed foods. The government could also act to expand the information and choices available to consumers who might want to reduce animal fat in their diets, such as by expand- ing nutrition labeling of meat or mandating the availability of low-fat or skim milk in school lunch and other government feeding programs.
156 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY Information and Education Interventions The final category of possible risk-management interventions is aimed pri- marily at consumers and involves the use of information and educational cam- paigns to encourage the changes in dietary patterns that would be expected to reduce DLC exposure. Such campaigns might be aimed at indigenous popula- tions whose traditional diets are drawn from contaminated areas, or at main- stream consumers who might achieve reductions in DLC exposure by conform- ing their diets to the government's Dietary Guidelines or to the Food Guide Pyramid. LEGAL FRAMEWORK FOR THE IMPLEMENTATION OF INTERVENTIONS As noted earlier, the committee assumed for purposes of identifying and analyzing risk-management options that the government would continue to oper- ate under current statutes, although legislation to address a particular hazard is always possible. The following summary focuses on these options because they traditionally have been the primary means through which the government has acted to reduce dietary intakes of environmental contaminants such as DLCs. There are, however, constraints in the form of substantive standards and burdens of proof that have an important effect on what the government can do to address a particular hazard. These constraints are among the reasons the committee has considered a range of nonregulatory, as well as regulatory, options for reducing dietary intakes of DLCs. Food Safety Regulatory Tools Among the regulatory agencies in the federal government, FDA has the primary food safety regulatory jurisdiction over DLCs in animal feeds and human food categories other than meat and poultry. For meat and poultry, USDA's Food Safety and Inspection Service (FSIS) plays the enforcement role. FDA is the federal government's food safety standard-setting agency for environmental con- taminants, other than pesticides. EPA considers food safety and acceptable levels of DLCs in fish by setting acceptable air and water emission levels for DLCs. The FDCA provides FDA with the broad authority to control or prohibit the presence in food of contaminants such as DLCs. The authority is broad in the sense that FDA can act to control DLCs in both animal feeds and human food; can intervene at virtually any stage of food production, processing, and market- ing; can set mandatory standards applicable to broad categories of DLC-contain- ing feeds and foods; and can take case-by-case enforcement action to remove DLC-contaminated food from commerce. FDA's authority is constrained, how-
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 157 ever, by the fact that its exercise is conditioned on the agency being able to make certain factual showings concerning DLCs and the risks they pose. Statutory Adulteration Provisions FSIS has similar adulteration authorities for meat and poultry. Under the Poultry Products Inspection Act and the Federal Meat Inspection Act, FSIS may take action against meat or poultry that contains an added poisonous or deleteri- ous substance that may render the product injurious to health. Also, FSIS has authority to act against a product that is for any reason unhealthful, and against product that has been prepared under insanitary conditions whereby it may have been rendered injurious to health. Under these provisions, DLCs generally would qualify as poisonous or del- eterious substances. The "may render" adulteration standard has been interpreted by the courts to mean that there must be a "reasonable possibility" that the substance will cause harm at the levels at which it is present in food. This stan- dard does not require proof of actual harm to health, but it does require scientific evidence demonstrating that there is a reasonable possibility of harm to at least some who consume the food. Likewise, the adulteration provision that deems added but avoidable poisonous or deleterious substances to be adulterants re- quires factual evidence that the particular level of the substance in food is avoid- able using GMPs. The practical meaning of this avoidability standard of adultera- tion has not been clearly defined by FDA or the courts. As persistent and pervasive environmental contaminants, DLCs would likely be considered unavoidable at some low concentration, but determining the line between avoidability and un- avoidability for DLCs poses complex factual and policy issues, and that line has not been drawn. To enforce the first two adulteration provisions noted above, FDA could choose to proceed on the basis of case-by-case enforcement action, through infor- mal or formal standard-setting, or a combination of these two. FDA has the option, if it can prove that a particular lot of food contains an environmental contaminant at a level that violates one of the adulteration provisions, to seek court action to remove that lot from commerce and prevent its future distribution. This can be done without any prior notice or setting of standards concerning the level of the contaminant the agency considers to be in violation of one of the adulteration provisions. FDA rarely proceeds on this basis, and it would be an inefficient way to address a pervasive environmental contaminant such as DLCs. FSIS has the authority to withhold the mark of inspection from product, which is necessary for the product to enter commerce, if it is not able to find that the product is not adulterated. Once the product is in commerce, FSIS' s authority is similar to FDA's.
158 Action Levels DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY Instead, prior to taking court enforcement action, FDA typically announces in advance the level of an environmental contaminant against which it is prepared to act. FDA calls these levels action levels. They are not regulations and do not have any binding legal effect on FDA or on private parties. Technically, they are guidelines for use by FDA enforcement officials in deciding when to bring court enforcement action against a contaminated food or its purveyor under one of the adulteration provisions. In such an enforcement action, FDA still has to prove to a court's satisfaction that an adulteration provision has been violated. As a prac- tical matter, however, FDA' s action levels operate as guidance to the food indus- try and are recognized by many food producers and processors as informal stan- dards with which they seek to comply. To date, FDA has not set action levels for DLCs as a class, though it has established an action level for polychlorinated biphenyls (PCBs) in meat. Tolerances In the case of added substances that are at least to some extent unavoidably present in food, as in the case of environmental contaminants like DLCs, FDA has statutory authority to establish binding legal limits on the level of the sub- stance that may lawfully be present in food. These limits are called tolerances. They are set at the level FDA considers necessary to protect public health, taking into account the extent to which the presence of the substance in food is unavoid- able. Tolerances are regulations that the FDCA requires be established under "formal" rulemaking procedures, which involve public notice and comment and formal trial-like hearings. These rulemaking procedures are costly and time con- suming. The advantage of tolerances, from an enforcement perspective, is that they are legally binding and directly enforceable through court action without any further showing by FDA that the food is unsafe or adulterated. The only toler- ances FDA has established to date are for PCBs in certain human foods (milk, dairy products, poultry, eggs, fish, infant and junior foods) and in animal feeds and feed components for food-producing animals. Regulatory Limits Due to the cumbersomeness of the formal rulemaking procedures required to establish tolerances, FDA has created an "informal" rulemaking process for es- tablishing regulatory limits for added poisonous or deleterious substances. A1- though FDA has yet to use this procedure to establish any regulatory limits, it is available in cases in which FDA is not prepared to establish a tolerance (because, for example, the levels of a contaminant may be in flux), but is prepared to define
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 159 the level at which it deems the contaminant to adulterate food because it "may render [the food] injurious to health" (Merrill and Schewel, 1980~. Regulatory limits would be established following public notice and comment, but without the formal hearings required for tolerances. It is uncertain how much deference courts would accord FDA's regulatory limits in an enforcement case. Good Manufacturing Practices The provision that declares food to be adulterated based on the conditions under which it has been prepared has been used by FDA to promulgate regula- tions establishing GMPs and other requirements for food production that are intended to prevent potentially harmful contamination. FDA has established GMP regulations for animal feeds and human food, but these currently address general practices for safe and sanitary food production; they do not establish any stan- dards or requirements for preventing or minimizing contamination with specific contaminants such as DLCs. FDA could attempt to use its GMP authority for this purpose, but the agency would have to support such regulations with evidence that establishes a link between any specific GMPs for DLCs and the prevention of situations in which the food may be injurious to health. This evidence would be difficult to provide given the lack of data on these linkages. Practical Constraints on Regulatory Options While FDA has successfully used these regulatory tools to address a wide range of food safety issues, including hazards posed by environmental contami- nants such as lead, mercury, and aflatoxin, the agency would face significant practical constraints in applying them to DLCs in feeds and food. As discussed in Chapter 2, there continues to be scientific debate and uncertainty about the level of human intake of DLCs required to cause adverse health effects in the popula- tion at large and in specific subpopulations, especially in light of the fact that the risks of DLCs are generally thought to be associated with their accumulation in human tissue over time. This uncertainty does not necessarily preclude, but would certainly complicate, any effort by FDA to establish tolerances or other regula- tory standards through rulemaking or to enforce adulteration provisions through court proceedings. In such rulemaking and court enforcement proceedings, FDA would bear the burden of proving the potential of DLCs to cause harm to human health at relatively low intake levels. In addition, effective compliance monitoring and enforcing regulatory stan- dards both require the availability of practical analytical methods and a credible level of testing to detect violations. The expense of available testing methods creates a barrier to enforcing regulatory standards for DLCs. The difficulty of meeting this legal burden, in light of the uncertainty about risk at low exposure levels and the paucity of data on actual levels of DLCs in food, is one of the
160 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY reasons the committee is not recommending any immediate action using these traditional food safety regulatory tools. The committee is aware of the recent initiative of the European Union (KU) to establish quantitative standards for DLCs in feeds and food, as discussed in Chapter 2, but notes that EU regulatory policymakers are not subject to the same legal constraints as U.S. regulators since, in the EU system of government, the adoption of the DLC policy was itself a legislative act. Voluntary Guidance, Cooperative Programs, and Subsidies to Reduce DLCs Regulatory agencies, including FDA and FSIS, commonly supplement their regulatory standard setting and enforcement programs with collaborative pro- grams to encourage and support voluntary efforts by food producers and proces- sors to address health, safety, and environmental problems. Such efforts can be productive when private entities are provided incentives to address issues when regulatory mandates are absent, and they are often pursued in circumstances in which there are legal, scientific, resource, or other practical constraints that make regulatory intervention difficult or impracticable. For example, both FDA and FSIS have begun to work with food producers to foster voluntary good agricul- tural practices (GAPs) and other measures at the food production level to address food safety hazards posed by microbial pathogens. The committee has identified a number of voluntary efforts as options to reduce DLC levels in feeds and food. They include setting nonbinding targets for reducing DLC levels in feeds and food, working with the industry to establish and encourage the adoption of voluntary GAPs and GMPs, and providing data on the current levels and sources of DLCs in feeds and food to support voluntary reduc- tion efforts. There are no legal constraints on such efforts. The committee has also identified various economic subsidies to food pro- ducers and processors as possible options for reducing DLC levels in feeds and food. These options may be more novel as a policy matter and require additional resources that are not provided for in the budgets of the food safety regulatory agencies. In addition, the regulatory agencies likely lack legal authority for such subsidies. These options would thus require congressional action to implement. Reducing Dioxin Intake by Changing Food-Consumption Patterns Perhaps the most direct way for an individual or a population to reduce dietary intake of DLCs is to reduce their consumption of dietary fat, especially from animal sources that are known to contain higher levels of these compounds. The federal government has little or no legal authority to directly control what people eat, but the committee has identified a number of options to encourage changes in food-consumption patterns that, if achieved, would reduce dietary
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 161 intake of DLCs. These include the provision of information to consumers about DLC levels in various foods, public education programs, and changes in govern- ment feeding programs to facilitate and encourage the consumption of foods less likely to be contaminated with DLCs. Most of these efforts could be pursued within the current statutory authorities of the food safety and other federal agen- cies. In some cases, however, changes in existing regulations would be required. ANALYTICAL TOOLS FOR HEALTH RISK DECISION-MAKING In recent years, policymakers and regulators have developed the risk analy- sis framework as a means of organizing the information needed to make deci- sions about controlling health and environmental risks. The framework starts with risk assessment, which is a method for identifying the possible harms that may result from a specified activity and estimating the probability that such harms will occur and the number of people that may be affected. It continues with risk management, which is the process of weighing policy alternatives and, if required, selecting and implementing appropriate control options. The final ele- ment of the framework is risk communication, which is the interactive exchange of information and opinions concerning risk among risk assessors, risk managers, consumers, and other interested parties. As noted previously, the sponsors of this study did not charge the committee with conducting a risk assessment for DLCs in the food supply, but rather asked it to rely on already existing information. The committee was charged to develop and assess risk-management options for reducing DLC exposure from foods. In considering these options, it was also asked to incorporate risk-risk or risk-tradeoff analysis. This approach is a method for evaluating regulatory and other options to reduce risk by highlighting the risks that may be created by these actions. It provides a means for considering multiple countervailing risks that may indicate that it is riskier to remediate a problem than to take no action. As discussed in detail below, the committee chose to employ a broader framework, termed risk- relationship analysis, which takes into consideration both the countervailing risks and the ancillary benefits that may result from a particular risk-management option. The committee also considered several aspects of risk communication, particularly for management options that involve voluntary actions by consumers or industry. The objective of the committee's approach, as presented in the framework, is to point out the types of information needed for comprehensive risk-management decisions and to explain the range of risk-management options that were consid- ered by the committee in making its final recommendations. The choice of efficient risk-management options may make use of cost- effectiveness or benefit-cost analysis. Both are methods of economic-welfare analysis that attempt to comprehensively evaluate the beneficial and adverse effects of a policy. Cost-effectiveness analysis measures the primary effects (i.e.,
162 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY the benefits) in some natural units (e.g., cases of illness prevented, Quality Ad- justed Life Years) and other consequences (i.e., the costs) in monetary units. Benefit-cost analysis measures all outcomes in monetary units. The essential idea is to choose risk-management options that provide the most risk reduction for a given cost (cost-effectiveness analysis) or offer the largest surplus of benefits over costs (benefit-cost analysis). While some elements of the application of these economic approaches to the reduction of DLC exposure from food are discussed below and in Chapters 7 and 8, the committee was not able to conduct a comprehensive analysis of this type that would allow evaluation across DLC management options because it would require resources (e.g., time, data, and money) to conduct such an analysis that was far beyond the ability of the committee. Further, the committee recognizes the importance of using comparative-risk analysis as a method of identifying government priorities among diverse health-related risks and allocating resources across those risks. Such between-risk comparisons were beyond the scope of this study. RISK-RELATIONSHIP ANALYSIS Past government efforts to reduce risks posed by environmental contami- nants in food generally have focused on the health risk posed by the contaminant. They have relied primarily on the traditional food safety regulatory tools outlined earlier, sometimes coupled with public advisories, such as ones issued to encour- age reductions in the consumption of certain fish due to high levels of mercury. Other essentially economic considerations have also played a role in regulatory choices. For example, the FDA tolerances for PCBs in fish were established on the basis of a consideration of how much fish would have to be removed from the food supply to comply with the ultimately enacted tolerance of 2 ppm. This consideration ultimately served as a practical way of assessing what levels of contamination would be deemed avoidable (above 2 ppm) and unavoidable (be- low 2 ppm). In contrast, the charge to the committee was to look beyond these traditional considerations in food safety regulation, both to consider a broader array of possible risk-management options and to consider a broader array of factors in evaluating possible options. Specifically, the study sponsors asked the committee to consider the possible nutritional consequences of dietary changes that might result from measures taken to reduce DLC exposure. The hypothesis underlying this charge was that such dietary changes, which might stem, for example, from increases in food prices due to costs imposed by new regulatory requirements or by the shifting of food-consumption patterns away from animal fat, might create countervailing health risks by depriving some individuals of needed nutrients. While the consideration of countervailing health risks is not common in food safety regulation, it is an appropriate approach in any public health context where
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 163 interventions that are being considered to reduce a target risk could plausibly increase some other health risk. Consideration of such countervailing risks is sometimes called risk-risk or risk-tradeoff analysis. In the case of DLCs in food, however, the committee considered it at least equally plausible that dietary changes to reduce DLC exposure could have ancillary health benefits, in addition to whatever health benefit was achieved by reducing DLC intake. In particular, reductions in the consumption of animal fat to reduce DLC exposure may also be associated with a number of other health benefits, including reductions in cardio- vascular disease and cancer. The committee thus decided to adopt a framework for analysis of DLC risk-management options that is broader than a risk-risk or risk-tradeoff framework. It involves risk-relationship analysis. In the context of diet-related issues, risk-relationship analysis recognizes that most human foods provide a combination of desirable and undesirable compo- nents. In addition to attractive taste, texture, and other aesthetic attributes, desir- able components include macronutrients (protein, carbohydrate, certain fats, and fiber) and micronutrients (vitamins and minerals) that are required to support health and maintain adequate nutritional status. Other food components that may be undesirable and pose health concerns for consumers include excess levels of saturated fats and cholesterol and both naturally occurring and man-made food- borne toxins. Components that should not be present in foods include microbial pathogens, pesticides, and other hazardous contaminants. At a fundamental level, optimizing dietary choice can be viewed as choosing a mixture of foods that provides adequate and balanced qualities of desirable components, while mini- mizing health risks associated with the consumption of undesirable components or contaminants. Defining a "healthy diet" remains a matter of controversy, precisely because food is compositionally so complex, dietary patterns are so varied, and there remains considerable scientific uncertainty about the interactions of dietary com- ponents and the impact of changes in dietary patterns. Nevertheless, the U.S. Dietary Guidelines for Americans (USDA/HHS, 2000), which establishes the policy, and the Food Guide Pyramid (USDA, 1996), which provides the educa- tional component, reflect consensus recommendations about desirable intakes of micro- and macronutrients. The committee concludes that, in considering risk- management options to reduce DLC exposure, it is appropriate to conduct a risk- relationship analysis that considers whether a proposed intervention would affect dietary patterns, for better or worse, in relation to current dietary guidelines. Ideally, a risk-relationship analysis would provide a quantitative basis for considering and comparing the risk reduction likely to be achieved by the pro- posed intervention, any countervailing health risks the intervention might cause, and any expected ancillary health benefits (see Box 6-1~. Proposed interventions that have positive net health impacts, based on an aggregation of such quantita- tive assessments, would be candidates for consideration, subject to whatever
164 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY legal standards and other impact assessments that might be relevant to the selec- tion and implementation of a risk-management option. As is made clear in the discussion of specific risk-management options in Chapter 7, it may not be possible to conduct quantitative risk-relationship analy- ses, and even the qualitative outcome (i.e., are net health benefits likely to be positive or negative) of a risk-management option may be unclear. This is the case for several reasons: (1) the models and data required to predict changes in dietary patterns are extremely limited, (2) a given intervention might push diets in more than one direction; for example, reductions in the consumption of full-fat dairy products could not only reduce overall intake of undesirable animal fat, but it could also reduce the overall intake of desirable calcium if the low-fat product is not consumed, and (3) the health consequences of a given dietary change will vary among demographic groups (e.g., by age and sex) and among individuals
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 165 (depending on the individual's health status and the nature of the diet prior to the intervention). Despite these complexities and uncertainties, which inevitably make the answers produced by risk-relationship analysis uncertain, questions about coun- tervailing health risks and ancillary health benefits are worth asking. In at least some cases, they can help affirm a proposed course of action or point policy- makers in a different direction than originally proposed. OVERVIEW OF THE PROPOSED FRAMEWORK With the foregoing considerations in mind, the committee adopted a frame- work for identifying and analyzing risk-management options for reducing DLC exposure. In Chapter 7, these options are presented in three matrices correspond- ing to the three major categories of intervention options: Matrix 1: Intervention Options Considered to Reduce DLC Exposure Through Pathway I: Animal Production Systems Matrix 2: Intervention Options Considered to Reduce DLC Exposure Through Pathway II: Human Foods Matrix 3: Intervention Options Considered to Reduce DLC Exposure Through Pathway III: Food-Consumption Patterns Within these matrices, each considered option is presented in the format shown in Figure 6-1. This format captures the series of questions the committee believes government agencies should consider in identifying and analyzing risk-manage- Option Alternate/l nterim Actions Current Barriers to Implementation DLC Exposure Reduction Exposure Reduction: Risk- Relationship Analysis Ancillary Benefits: Countervailing Risks: FIGURE 6-1 Format for analyzing intervention options considered to reduce DLC exposure.
166 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY ment options for DLCs in food: (1) possible alternate or interim actions that could be taken in the short term if the proposed intervention is not immediately feasible, (2) the current barriers to implementation of a proposed option and its practical feasibility, (3) the potential target DLC exposure reduction achievable through the proposed intervention, and (4) the existence and size of ancillary benefits and countervailing risks associated with the option. The proposed framework and matrices respond directly to the committee's charge to consider options for reducing DLC exposure, taking into account the possible nutritional consequences of dietary changes that might result as a conse- quence of the measures taken. The charge to the committee and the proposed framework address the health consequences of efforts to reduce DLC exposure, not the economic consequences. The committee considers the practical feasibility of a proposed intervention to be a valid threshold question because no health benefit will accrue from an intervention that cannot feasibly be implemented. While the committee considers the economic feasibility of the options to be a key issue, it has not included benefit-cost analyses of the specific options proposed in the framework. This is primarily because consideration of the economic costs of DLC reduction options is beyond the scope of the committee's charge, except to the extent that economic costs imposed by an intervention affect dietary patterns and health. In addition, the traditional food safety regulatory tools do not provide for the consideration of benefit-cost analyses in setting food safety standards. Nevertheless, the committee recognizes that major government regulatory decisions, even if not controlled by cost-effectiveness, benefit-cost, or other eco- nomic analyses under their authorizing legislation, are subject to regulatory im- pact assessments under presidential Executive Orders administered by the Office of Management and Budget in the Executive Office of the President. Moreover, such economic analyses are likely to play a role in decisions on whether to mount significant new initiatives to reduce DLCs. These analyses raise their own com- plicated issues, from both methodological and policy perspectives, and these issues are beyond the scope of this report. The elements included in the committee's proposed analytical framework are described in more detail in the following sections. Alternate or Interim Actions Closely related to the issue of feasibility is the need to consider, for each proposed intervention, whether there are alternate or interim actions that could be taken in the short term to begin making progress in reducing DLCs or to make necessary preparations for interventions that may not be feasible at present. The committee finds that, while substantial public investments have been made to collect the information required for risk assessment with respect to DLCs (e.g., toxicity and epidemiological and exposure data), relatively little public invest- ment has been made to collect the data and develop the tools required for risk
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 167 management, such as rapid and inexpensive test methods. For example, there are very limited data on the actual distribution of DLCs in the production chain and food supply at the level of detail required to establish binding regulatory limits on DLCs in feeds or food, or for mandating significant changes in current production practices that affect the presence of DLCs in food. Thus, it is appropriate with respect to a number of the proposed interventions discussed in Chapter 7 to consider alternate or interim steps, such as developing improved test methods, expanding data collection programs, and fostering voluntary efforts to reduce DLCs. Current Barriers to Implementation The committee considered at length how DLCs enter the food supply, their pervasiveness, and some of the obstacles to reducing DLC exposure from food. No one desires the presence of DLCs or directly controls their entry into the food chain. DLC contamination is, for many participants in the food chain, an un- avoidable phenomenon. The widespread presence of DLCs in the environment means that DLC contamination is, to some extent, embedded in the way food is produced, beginning on the farm or in waterways. This is why, in the committee's view, continued efforts to reduce environmental emissions must remain a key element of any DLC exposure reduction strategy. The fact of widespread DLC contamination of the environment does not mean there are no feasible interventions to reduce DLC exposure through food. The committee has identified, and discusses in Chapter 7, a number of risk- management options that are more or less feasible, depending on the point in the food chain at which the option seeks to reduce DLCs and on the magnitude of the reduction being sought. Some options are clearly feasible technically, such as replacing the animal fat with high DLC levels that is contained in animal feeds with other energy sources, but they may require such significant change in cur- rent practices or may raise other issues (such as finding alternative means to use or dispose of the fat) that their practical feasibility may be legitimately ques- tioned. Other options may simply not be technically feasible, at least at present, such as the removal of DLCs from feeds or food components through processing. In any event, technical feasibility is an appropriate first question to ask in consid- ering DLC exposure reduction options. An evaluation of the economic feasibility of different risk options would require a much more extensive analysis than could be carried out by the commit- tee. Experience with evaluating other risk-reduction programs does indicate a range of private and public costs that might be associated with reducing exposure to DLCs in food. In the private sector, risk-reduction options may require in- creased expenditures in the supply chain, such as costs of monitoring, certifica- tion, substitution of higher priced inputs, or other changes in standard operations. Consumers may also be called upon to do more to protect themselves from DLC
168 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY exposure. A significant lesson from prior research is that both short- and long- term costs should be evaluated, since innovation and new approaches may cause long-term costs to diverge from short-term costs. In the public sector, risk-man- agement options incur costs related to monitoring, establishing programs, and setting and enforcing standards. Again, while not fully analyzed by the commit- tee, economic feasibility remains an important consideration in evaluating risk- management options. DLC Exposure Reduction In the committee's proposed framework, DLC exposure reduction is the declared goal, but it should be understood to be a surrogate for risk reduction. The reason to reduce DLCs in the food supply and in diets is to reduce the potential for adverse health effects resulting from exposure and accumulation of DLCs in human tissue. The estimation of the risk reduction associated with any increment of exposure reduction is, however, fraught with uncertainty, due both to the inherent difficulties of extrapolating from test conditions in which potential haz- ards have been identified and to the complexity and diversity of conditions under which humans are exposed. Numerous health effects have been ascribed to exposure to DLCs, including dermatological damage, cancer, noninsulin-dependent diabetes in adults, neuro- logical and immune system impairments in infants, and endocrine system disrup- tion (see Chapter 2~. Many of these effects were identified in individuals sub- jected to high levels of exposure (e.g., through unintended industrial releases); others have been described in specific population groups such as the Ranch Hands cohort. However, information on adverse health effects caused by low- level exposure to DLCs through foods is limited. Another source of information on the toxicity of DLCs is research on animal models. These studies, however, have inherent limitations that affect their use in human risk assessment. The dose-response effect of DLC exposure varies widely among animal species and between animal models and humans. This variability makes it difficult to establish an accurate threshold dose and leads to increased uncertainty, particularly when extrapolating toxicity findings from animal mod- els to humans. These studies are further limited by the need to use single com- pounds for exposure assessments, whereas in human exposures, multiple com- pounds and other confounding variables are present. Estimations of risk using animal models must be approached carefully and in conjunction with other re- sources, including epidemiological evidence. In addition to the uncertainties and variability inherent in the DLC health- effects data, there is wide variability in the vulnerability of human subpopula- tions. For example, fetuses and breastfeeding infants and young children may be especially vulnerable to the hazards posed by DLCs. This compounds the diffi- culty of assessing the health importance of any projected decrease in DLC expo-
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 169 sure and underscores the need in evaluating proposed interventions to consider how the anticipated reductions in exposure will be distributed. Despite these scientific uncertainties and complexities, the assumption un- derlying the charge to the committee was that it is desirable to reduce DLC exposure through food. The committee embraces this assumption as an expres- sion of sound public health policy. The committee recognizes, however, that it will be difficult to predict the magnitude of the exposure reduction likely to result from any proposed intervention or the health benefit of the reduction, and that the benefits, whether measured in terms of direct health outcomes or the monetary value of those outcomes, will vary among individuals and subpopulations. The committee did not conduct a detailed analysis of this kind for any of the options it identified. The framework identifies exposure reduction whether, to what extent, and for whom it occurs to be an important question in evaluating risk- management options. Risk-Relationship Analysis As discussed earlier, risk-relationship analysis goes beyond consideration of the target risk reduction likely to be achieved by a proposed intervention it also considers ancillary benefits and countervailing risks potentially caused by the intervention. In the case of options for reducing DLC exposure from food, the central source of ancillary benefits and countervailing risks are potential changes in dietary patterns that may occur as a result of implementing a risk-reduction strategy. Increases in countervailing risks and ancillary benefits could occur at the same time and to the same or different people. For example, one approach to reduce exposure is to reduce the consumption of meat, poultry, fish, egg, and milk products that may contain undesirable levels of DLCs. However, these foods are sources of protein, iron, calcium, and vitamin D and, in the case of fish, omega-3 fatty acids. Therefore, decreases in intake may increase the risk for nutritional deficits and for some diseases. Alternatively or at the same time, there may be corollary benefits related to the decreased consumption of some animal foods. For example, a decreased intake of saturated fats and cholesterol is associ- ated with a decreased risk for heart disease and possibly some cancers. The committee concluded that it is appropriate to consider both the potential ancillary health benefits and countervailing health risks associated with different risk-management options, recognizing that the potential health impact of any dietary change would vary among population subgroups and individuals depend- ing on their nutritional needs and nutritional status at the time of the dietary change. As discussed in Chapter 5, the committee utilized two data resources to help predict the outcomes of certain scenarios of dietary modification that could result from efforts to reduce DLC exposure through foods. The Continuing Sur- vey of Food Intakes by Individuals (a food-consumption survey conducted by
170 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY USDA) was mapped to FDA's Total Diet Study to estimate probable DLC expo- sure by food type for various age and gender groups. The results of the analysis, summarized in Chapter 5, suggest that reductions in the consumption of animal fats can reduce DLC exposure levels to varying degrees; approximately 10 per- cent in the case of this analysis. Whether the reduction is significant for the general population is not known. However, it could be a way to reduce exposure to sensitive population groups in such a manner that negative nutritional conse- quences could be avoided. The committee also considered the possibility that regulatory interventions to reduce DLC exposure in the diet could impose costs on the food production system that would be reflected in higher consumer prices for food, which could in turn affect dietary choices and nutrient intake levels to an extent that could have negative health consequences. The impact of the price of food on nutrient avail- ability may be significant for some low-income, sensitive, or highly exposed groups. The importance of such an effect depends on three factors. The first is how much food purchases and resulting nutrient intakes respond to changes in food prices. The second is whether any resulting changes in nutrient intakes are nutri- tionally significant, that is, are they large enough to have possible health conse- quences. The third is whether the changes are adverse or beneficial. Research by Huang addresses these points. For the general population, Huang (1996) developed a methodology to mea- sure how a change in the price of a food-product category (e.g., beef) would impact the consumption of that product category (the own-price elasticity effect) and the consumption of other food-product categories (the cross-price elasticity effect). Huang then linked these changes in consumption to changes in nutrient intakes through the use of a database containing nutrient values of different food products. He found that a 1 percent increase in the price of beef would reduce per capita food energy by 0.027 percent, protein by 0.091 percent, and fat by 0.025 percent, but vitamin A would increase by 0.064 percent, with the latter change being particularly affected by cross-product effects. Thus, while a price change affects nutrient intake, the impact in percentage terms is much smaller than the change in price due to the ability of consumers to attain nutrients from substitute products. The effects of changes in nutrient intake on health depend on whether current intakes are above or below optimal levels. Moreover, while some of the changes in nutrient intake may adversely effect health (e.g., a reduction in pro- tein), others may be beneficial (e.g., a reduction in fat, an increase in vitamin A). Huang (1997) went on to assess the nutritional significance of changes in nutrient intakes resulting from price changes by evaluating them relative to FDA daily values for those nutrients. In this case, he evaluated the impact on nutrient intake of a much larger change in price of 10 percent. Even at this higher level, Huang concluded "a comparison of the nutrient quantity changes with the daily values shows that the magnitudes of nutrient quantity changes are relatively small
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 171 and, therefore, unlikely to be of nutritional significance" (p. 18~. He cautioned, however, that the nutrient effects could exacerbate existing nutritional problems if they occurred over prolonged periods of time. A risk-management option to reduce DLC exposure might affect the prices of more than one category of food products at the same time, potentially increas- ing the impact of the option on nutrient intakes. Nevertheless, it is the committee's judgment that the potential countervailing risk stemming from a causal link be- tween increased production costs, increased food prices, changed dietary intakes, and poorer health is not likely to be an important consideration in evaluating risk- management options for reducing DLC exposure. Indeed, some of the changes in dietary intake are likely to be beneficial for most consumers (e.g., a reduction in fat). Unless price increases were quite large (e.g., greater than 10 percent) and widespread across food product categories, it is likely that the general population would be able to attain sufficient nutrients by substituting other food products for those that became more expensive. However, the impact of price changes for the nutritionally vulnerable could require closer scrutiny. The committee also discussed a separate strand of economic analysis, health- health analysis, which focuses on expenditure tradeoffs made by consumers (Chapman and Hariharan, 1994, 1996; Keeney, 1990, 1997; Lutter et al., 1999; Viscusi, 1994~. This approach argues that when risk-reduction options instituted by the government cause consumers' costs to increase, consumers must spend less on other goods, some of which promote health and safety. This decrease in spending on health and safety could result in increases in other risks that could offset any decrease in risk resulting from the government policy. The committee had diverse views on whether this countervailing risk scenario was likely to be an important factor in analyzing options to reduce DLC exposure from foods and did not include further discussion of it in Chapter 7. RESEARCH NEEDS The framework for discussing intervention options ends with research needs related to each option. The committee found that these needs were numerous and diverse, such as testing methodologies and techniques, sampling for DLC pres- ence in the food-supply chain, and analyses of the impacts of dietary changes on DLC exposure. In discussing each option in Chapter 7, the committee lists key research needs related to the possible implementation and assessment of the option. CONCLUSION The proposed framework summarized in this chapter essentially poses a series of questions that the committee considers relevant in identifying and evalu- ating possible DLC exposure reduction options. In-depth analysis of the kind that
72 DIOXINS AND DIOXIN-LIKE COMPOUNDS IN THE FOOD SUPPLY would likely be necessary to support government action would require both a detailed description of the proposed intervention and its implementation, as well as good data and information to answer the questions posed by the framework, especially on the specific intervention's feasibility and likely contribution to exposure reduction. The committee did not attempt to develop and analyze op- tions at this level of detail. The framework can also be used, however, to identify and test options pre- lim~nanly to consider whether they are plausible and deserving of further refine- ment and evaluation. In Chapter 7, how the framework can be used for this purpose by applying it to some possible options is illustrated, and some of the issues and data needs that would have to be addressed to more fully develop and evaluate these options are discussed. REFERENCES AEA Technology. 1999. Compilation of EU Dioxin Exposure and Health Data. Prepared for the European Commission DO Environment. Oxfordshire, England: AEA Technology. ATSDR (Agency for Toxic Substances and Disease Registry). 1998. Toxicological Profile for Chlo- rinated Dibenzo-p-dioxins. Atlanta, GA: ATSDR. Chapman KS, Hariharan G. 1994. Controlling for causality in the link from income to mortality. J Risk Uncertain 8:85-94. Chapman KS, Hariharan G. 1996. Do poor people have a stronger relationship between income and mortality than the rich? Implications of panel data for health-health analysis. J Risk Uncertain 12:51-63. EPA (U.S. Environmental Protection Agency). 2000. Exposure and Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. Draft Final Report. Washington, DC: EPA. Fiedler H. Hutzinger O. Welsch-Pausch K, Schmiedinger A. 2000. Evaluation of the Occurrence of PCDD/PCDF and POPs in Wastes and Their Potential to Enter the Foodchain. Prepared for the European Commission DO Environment. Bayreuth, Germany: University of Bayreuth. Huang KS. 1996. Nutrient elasticities in a complete food demand system. Am JAgr Econ 78:12-19. Huang KS. 1997. How Economic Factors Influence the Nutrient Content of Diets. Technical Bulletin No. 1864. Washington, DC: U.S. Department of Agriculture, Economic Research Service. IARC (International Agency for Research on Cancer). 1997. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 69: Polychlorinated Dibenzo-para-Dioxins and Poly- chlorinated Dibenzofurans. Lyon, France: World Health Organization. Keeney R. 1990. Mortality risks induced by economic expenditures. Risk Anal 10:147-159 Keeney R. 1997. Estimating fatalities induced by the economic costs of regulations. J Risk Uncertain 14:5-23. Lutter R. Morrall IF III, Viscusi WK. 1999. The cost-per-life-saved cutoff in safety-enhancing regu- lations. Econ Inq 37:599-608. Merrill RA, Schewel M. 1980. FDA regulation of environmental contaminants of food. Va Law Rev 66:1357. Scientific Committee on Food. 2000. Opinion of the Scientific Committee on Food on the Risk Assessment of Dioxins and Dioxin-like PCBs in Food. Brussels: European Commission. Scientific Committee on Food. 2001. Opinion of the Scientific Committee on Food on the Risk Assessment of Dioxins and Dioxin-like PCBs in Food, Update. Brussels: European Commission.
DEVELOPMENT OF POLICY OPTIONS TO REDUCE EXPOSURE 173 USDA (U.S. Department of Agriculture). 1996. The Food Guide Pyramid. Home and Garden Bulle- tin No. 252. Washington, DC: USDA. USDA/HHS (U.S. Department of Health and Human Services). 2000. Nutrition and Your Health: Dietary Guidelines for Americans, 5th ed. Home and Garden Bulletin No. 232. Washington, DC: U.S. Government Printing Office. Viscusi WK. 1994. Mortality effects of regulatory costs and policy evaluation criteria. Rand J Econ 225:94-109.