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.
CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS CONCLUSIONS The Role of Risk Assessment Risk assessment is a specialized and systematic means for organizing and presenting information about various types of health hazards, including those associated with the consumption of broiler chickens. Because it requires explicit, consistent, and logical treatment of data and their associated uncertainties, and consideration of current scientific knowledge, risk assessment is one of the most valuabl e tools available to serve regulatory agencies e Therefore, FSIS should begin a continuing program of more formalized applications of risk assessment based on a refined risk model, such as that proposed by the committee in Chapter 3, to analyze specific risks associated with poultry and to evaluate alternative strategies for managing these risks. The data necessary to provide accurate quantitative risk assessments are not always available and vary greatly among the types of hazards presented by broiler chickens. Thus in many instances, particularly in connection with microbiological hazards, only qualitative assessments can now be done. ~ ~ ~ The consistent use of the conceptual framework and model for all assessments ensures that current information is being used in the most effective possible way to guide risk management. Although gaps in knowledge and lack of data limit the extent to which quantitative risk assessment can now be done, FSIS should consider the use of more formalized, and ultimately quantitative, risk assessment to serve as a foundation for important decisions involving issues of human health. Quantification is needed to clarify the magnitude of the various sources of human health risks and to provide a more defensible logic for decision making. Qualitative reasoning, even qualitative reasoning made more systematic by the use of a risk model, is often too vague and prone to error to serve as a satisfactory basis 158
159 for definitive decisions bearing on human health. However, this reasoning does not necessarily lead to the conclusion that FSIS needs a highly detailed, comprehensive risk-assessment system to support its decision making. Some initial efforts to quantify the framework chosen should reveal whether further quantitation will be cost effective and the extent to which risk assessment, in general, will be useful. Risk assessment is not the only component of a regulatory program. Although protection of the public from significant health risks is the ultimate goal of any FSIS program, the types and range of risk- management options are influenced by legal, regulatory, and historical precedents that cannot be ignored by decision makers. The committee has dealt only with the ways in which risk assessment can assume a more prominent role in the decisions made by FSIS in the absence of such constraints. The Risk Model The committee believes that the conceptual framework and risk model developed in Chapter 3 and examples of their application in assessing the risks presented by microbial contamination and chemical residues (Chapters 4 and 5) should serve as a guide in future FSIS data-gathering and risk-assessment activities. In Chapter 6, the committee used this analytical approach to identify risk-management options and to evaluate FSIS programs and activities. In so doing, it learned much about the adequacy of current FSIS programs. The recommendations set forth below are based on this evaluation. Current FSIS Programs The committee's use of the risk model in earlier chapters of this report was designed not only to illustrate how risk assessment can be used to evaluate current programs and to guide the development of future ones but also to answer the following question: Are current FSIS inspection techniques reasonably related to the level of public health risk associated with various components of the broiler chicken risk model? By applying the risk model developed in Chapter 3, the committee observed that the traditional postmortem inspection system cannot deal with several sources of risk at certain points in production and distribution due to limitations of FSIS authority. Among these are the assignment of acceptable daily intake (ADI) or other levels of tolerable intake for chemical residues, monitoring of poultry feed and drinking water to prevent contamination of the products, and education of the consumer regarding safe poultry preparation practices. Effective risk-management programs must include a range of activities that fall outside traditional postmortem inspection. Some of the committee's recommendations derive from this conclusion.
160 Microbial Hazards. As stated in Chapter 3, the committee believes that the present system of inspection provides little opportunity to detect or control the most significant health risks associated with broiler chickens. Although information is not sufficient for the committee to conclude that the FSIS inspection program has no public health benefits, the weight of the evidence does suggest that the current program can not provide effective protection against the risks presented by microbial agents that are pathogenic to humans. The committee believes that alternative and potentially less costly strategies, which are described below, may provide greater benefits to public health. An integral component of these strategies is the use of a statistically based sampling plan, rather than the traditional approach of 100% inspection. Some details of a possible approach are described in the recommendations presented below under Sampling Procedures . Current inspection procedures do not significantly reduce contamination of broilers by microorganisms that are pathogenic to humans, nor were they designed to do so. Even if they were to be responsible for small reductions in contamination levels, however, there are further opportunities for increasing or decreasing microbial contamination after the products leave FSIS control. For example, cross-cont~ination can occur during distribution and food preparation; proper cooking car effective! y eliminate microbial risks . These factors affect both Salmonella and Campylobacter, which are among the major causes of food-borne morbidity among humans in the United States. Since the current system of postmortem poultry inspection is not an effective device for managing this public health problem, several types of corrective action should be taken. Furthermore, other microbial risks need to be investigated with greater vigor, and additional data should be sought so that the magnitude of the public risk associated with microbial contaminants can be assessed more accurately. These constitute the committee's central conclusions regarding FSIS inspection programs. The recommendations presented below derive from them. Chemical Hazards. Unlike microbial risks, the relationship between chemical contamination levels in poultry at the slaughterhouse and on the dinner plate is much better understood. In this case, critical factors and uncertainties in the risk model concern the nature, likelihood, and level of chemical residues and the health effects produced by the ingestion of those residues. The problem is therefore analogous to other chemical risk management problems, which have been addressed effectively through an approach based on the establishment of maximum permissible concentrations. It seems clear, however, that there has been insufficient inves- tigation of the pass ible extent of the problem of chemical res idues in poultry. A comprehens ive analysis of the potential risk of chemical
161 residues in poultry was therefore not possible, but in examining the limited data available the committee found no evidence that such residues pose a sigificant threat to public health. The committee's recommendations in this area are designed to encourage systematic investigation of the problem and to increase the efficiency of current inspection efforts. Current monitoring programs are designed primarily to detect a given frequency of tolerance violations. This objective may not be compatible with public health protection--the principal goal of FSIS inspection. It does not, for example, take into account the fact that the potential public health risks posed by chemical residues vary considerably. Several recommendations concern the means to reorient current monitoring programs. In general, it appears that FSIS programs are directed more toward detecting of macroscopic problems rather than toward investigating the origins of health hazards and finding ways to solve them. The detection of microbial and chemical contaminants as part of a quality assurance program designed to assist in revealing their sources would be a commendable addition to current efforts. The general tone of all recommendations derives from these observations. RECOMMENDATIONS General Risk assessment based on the conceptual framework and a risk model such as that provided in Chapter 3 should be used to guide FSIS risk-management programs. Examples of how these may be applied to assessments of microbiological and chemical hazards and to program evaluation are given ~ n Chapters 4, 5, and 6. Current poultry inspection programs are primarily concerned with detecting diseased and damaged poultry so that very few of the problems detected by this system are threats to public health. Because federal funding of the inspection program is intended to protect public health, it seems clear that such funds should be expended in direct proportion to the magnitude of the public health risks. Qualitative risk assessments should be used to determine where the major risks occur, how they might be controlled, and consequently to which aspects of the inspection program funds should be diverted. When needed, quantitative risk assessments may be required to elucidate complex health risks. Rather than focusing on one procedure, such as bird-by-bird inspection, as the primary component of an inspection process, FSIS should direct its efforts toward the establishment of a comprehensive quality assurance program. Such a program would consist of several components, one of which might be organoleptic inspection.
162 . An important component of any quality assurance program is a statistically based random sampling protocol. The committee recommends that FSIS establish such a sampling program as part of any future modifications of its inspection system. Emphasis should be shifted from detection to prevention of problems at the earliest feasible stage in production to increase the effectiveness of poultry risk-management activities. In the areas of risk management outside the purview of FSIS, the agency should attempt to maintain a close liaison with responsible agencies D Microbiological Contaminants . . . FSIS should use the tools of risk assessment to establish priorities for a risk-management program to protect consumers from microbiological hazards using the tools of risk assessment. Relevant activities are: Identify and evaluate potentially pathogenic microorganisms found on poultry products. Determine the potential for exposure to an infectious dose of pathogens derived from contaminated poultry products. Evaluate the potential public health impact of failure to control each of the identified microbial hazards. FSIS should identify and monitor the critical control points of the poultry sys tem at which microorganisms pathogenic to humans are introduced. This will assist industry in identifying production practices that result in exposure of consumers to poultry-borne pathogens. · Because fecal contamination is a primary source of the microbial load on broiler carcasses' PSIS should identify and monitor critical control points within dressing procedures to prevent soilage of carcasses by intestinal contents and carcass-to-carcass contamination. · In describing the risk model (Chapter 3), the committee identified a variety of points in the slaughtering process where operations involving machinery have the potential to influence contamination levels. Application of this risk model to microbial risks in Chapter 4 led to the conclusion that several of these operations had a critical bearing on the levels of Salmonella on carcasses. Hence, greater emphasis should be given to the operation, maintenance, and improvement of machinery (e.g., pluckers and chillers) to reduce the likelihood of microbial contamination in slaughterhouses. FSIS should begin to lay the groundwork for a shift from the organoleptic inspection of each broiler chicken to a program that is more likely to have a substantial impact on human
163 . health and disease. Further development of quantitative health risk assessment will be an essential tool in this change. Educational programs for people who handle raw broilers in slaughterhouses, at retail, and during food preparation in the home and in commercial establishments should be established or intensified to alert them of the potential risks and to instruct them in proper food handling practices. As part of this effort, all poultry products inspected by FSIS should be labeled at retail to inform consumers about the optimal ways to protect themselves from microbial hazards. Community-based surveillance of pathogens on poultry products and of food-borne disease incidence in humans should be intensified to measure the success of programs undertaken to reduce the prevalence of human pathogens on market-ready poultry and the direct impact of program activities on the public's health. FSIS should ensure that it can achieve the above objectives either by its own activities or through agreements with other agencies having appropriate authority and expertise. Chemical Residues . To protect consumers from chemical residues in poultry products, FSIS should adopt a r~sk-management program that includes the following activities (some of which would require close coordination or agreements with other agencies): Identify hazardous substances that could appear as residues in poultry products. Establish ADIs or other levels of tolerable daily intake of such substances by humans. Establish tolerances for residue levels in edible poultry products and institute enforcement programs to ensure that those levels are not exceeded. Establish levels of intake for poultry through diet, drinking water, or other sources to ensure that tolerances in poultry products are not exceeded. Establish control programs to ensure that poultry feed, drinking water, or other sources of chemical residues do not contain chemicals of concern at levels exceeding those identified in the preceding activity. Establish procedures that provide for efficient removal of contaminated poultry products from commerce. Identify priorities for each activity described above using the concepts and tools of risk assessment.
164 FSIS should categorize chemicals according to source, as described in Chapter 5, to assist in the identification of risk-management responsibilities and obj ectives . FSIS should ensure that it has in place programs to carry out all activities not clearly assigned to FDA (tolerance setting for additives and drug residues), to EPA (tolerance setting for pesticides), or to other agencies. In USDA research programs, emphasis should be given to accidental chemical contaminants, environmental contaminants, and chemicals formed during processing, storage, and heating of poultry products. The current Exploratory Testing Program should be expanded to include research on the following essential topics: the toxic properties of chemical agents (to assist In setting acceptable intake levels)9 levels in poultry products, and methods of chemical analysis. In cooperation with FDA and EPA, FSIS should learn the extent to which the existing data on chemicals intentionally administered or applied to poultry or poultry feed are adequate to support current tolerance levels. A prevention program to monitor poultry feed and drinking water should be seriously considered as the first line of defense against residues in poultry products. Such programs could be undertaken using procedures recommended by FSIS. Close cooperation with FDA may be necessary to achieve this objective. Data necessary to implement such a program (i.e., data on maximum acceptable feed and water levels) should already be available for intentionally used substances. Additional research is necessary to establish feed and drinking water limits for environmental chemicals and accidental contaminants. The committee recommends that priorities for monitoring feed, water, and poultry products and the intensity of the monitoring efforts be directly related to the relative magnitudes of risk posed by candidate substances should they escape detection and enter the marketplace. Priorities for research should be assigned in the same way. A methodology for assessing relative risks is presented in Chapter 5 (Activity 8~. FSIS should periodically assess the public health risks of chemical residues in poultry products, using data collected from its residue monitoring program together with data on toxicity Monitoring and enforcement should be increased or decreased, according to the results of these assessments. Sampling Procedures The quality and safety of poultry products are linked closely to the occurrence of chemical and microbiological contaminants that may originate from a variety of sources. To ensure the safety of retail products, contaminants must be minimized. This is most effectively accomplished by detecting them as early as possible in the poultry system and preventing their introduction by contolling the source, for
165 example, by sampling and testing the water and feed of growing chickens. However, preinspection components of the poultry system are outside FSIS jurisdiction. Therefore, any sampling program adopted by FSIS incorporating, for example, an effort to monitor microbial counts during slaughter should be complemented by collaborative efforts with FDA and EPA to monitor the microbial load of feed that might have been contaminated by rodents or that might contain improperly sterilized animal by-products. One possible scheme for monitoring chickens during slaughter is presented below using the format of ~ three-stage sampling scheme. The committee recommends that FSIS develop and maintain close liaison with programs that monitor the acceptability of feed and water supplied to growing chickens and substantially expand its own sampling program to assess the quality of poultry products and, most importantly, the human health hazards of freshly slaughtered birds. To accomplish this, FSIS should consider the following three-stage sampling scheme developed by the committee. In the first stage, a random number generator (functioning independently from plant quality control programs), a counter, and a signal attached to the line could flag birds with a probability (~) of removal from the line and mark them for detailed inspection. The sampling fraction ~ might be set at 1/500 or 1/5,000 (or higher or lower), depending on plant history, past performance of the grower, and other factors. Plant inspection staff should have a role only in determining how many birds are chosen, not which birds are chosen. Sampled birds would be marked to identify flock and time of inspection, for example, and hung on a rack for organoleptic examination in groups (perhaps by flocks) in a manner substantially more detailed than at present. This sample inspection would of course not preclude additional inspection of the present type or other types that might be deemed appropriate by the inspector in charge; it provides a minimum for FSIS involvement, not a maximum. This first-stage inspection would have several purposes: to maintain awareness of the quality of the product at this stage, including detection of gross disease and malfunctioning equipment; to ensure that FSIS can fulfill its statutory mandate; to provide direction for further studies, especially laboratory studies of microbial and chemical hazards; to exert pressure on poultry processors to apply inspection procedures that conform to acceptable standards; and to provide data that could serve as a teas is for further modification of the sampling system, e.g., in the size of the samples, in inspection techniques, and in objectives. The committee recognizes the appearance of a disjunction between earlier statements that organoleptic inspection has no demonstrated health value and this recommendation to continue organoleptic
166 inspection of a small percentage of slaughtered birds. There is no contradiction, however, since the recommended inspection of the sample is not intended to detect and condemn defective birds individually but, rather, to provide a continuing source of information about how the poultry industry responds to the shift from treatment to prevention of problems, to ensure that plant inspection does not deteriorate to the extent that new health hazards are presented, and to maintain awareness of changes in a flock's condition or in equipment performance that may have health impli cations . It may in time become apparent that this phase of sampling can be omitted under some conditions, for example, when a producer has a very strong quality control program of its ohm and a long history of minimal problems in the samples, but the committee does not recommend the complete omission of this step under exis tiny conditions . In the second stage of sampling, most birds selected in the first stage would be returned to the line, but a random subsample would be retained for still more detailed study, including simple laboratory studies for microbial load and chemical residues at the plant. The sample must again be strictly random, perhaps by rack location (randomized anew each time). Sampling fractions might be anywhere from 1/20 to 1/500, depending again on the past record of the supplier and the past performance of the plant. The purpose of this stage is surveillance for specs fic, direct microbial and chemical risks to health to ensure a timely response may be fully processed and shipped available, but the primary goal of and correction as needed. Flocks before any of the test results are this sampling would have been met, that is, to develop a needed information base, not now available, on health hazards as they appear at this very critical point in the process. Recall of birds or parts found to have very serious problems might be feasible. Thus, an effort should be made to find and implement rapid testing procedures that will make it possib] e to remove contaminated birds before they reach the retail market. In the third stage, a further random subsample of the second stage would be frozen or otherwise preserved and sent to a central laboratory for studies not feasible at the plant. This would in essence continue and expand the present scope of NRP sampling, and methods and sampl ing fractions should be adjusted to both the information need and the work load that NRP can accommodate. The purpose and justification would be similar to those supporting present NRP activities . Chemical testing would probably dominate this stage, but the committee believes that opportunities for further microbial testing should be examined by FSIS. This subsample should be subjected to morphologic and etiologic diagnoses to help build a data base against which future program activities can be measured. The co~Tunittee also recommends that the present size and composition of the NRP sample be reexamined in light of data needs and resource allocations within the overall inspection program proposed by the committee. With the present random sampling of 300 chickens per year,
167 some plants and some large producers may not be sampled for years at a stretch and even major problems that are localized in space or time (accidents) may be overlooked and thus not resolved in a timely fashion. In short, the stated goal of a 95% probability of finding a chemical contaminant present in 1% of birds slaughtered during a single year may not be sufficiently protective of the public health to meet current expectations. However, it would be unrealistic to sample the entire U.S. population of chickens as if they are a homogeneous group. The sample size required would be unmanageably large. Instead, FSIS should develop an appropriate sampling structure based, for example, on homogeneous and useful units like rearing groups, or random lots of birds characterized by some common feature such as their point of origin. (Recommendations for matching sampling rate to relative risks are presented above under Chemical Residues.) Sampling of feed, water, and processed products is likely to be less complex than sampling of birds on the line, and FSIS already has programs in place or models to follow. The committee emphasizes, however, that expert technical support should be provided to ensure that sampling plans meet program goals in a technically sound and cost-effective manner.