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 3 RISK-ASSESSMENT MODEL FOR POULTRY INSPECTION: ANALYTICAL APPROACH As indicated in Chapter 2, it is generally accepted that the design and selection of inspection strategies for controlling human health risks associated with broiler chickens should be based on risk assessment. The present committee concluded that a complete quantitative risk assessment is not possible at this time because of a lack of data and limited resources. However, a qualitative assessment based on the concepts of risk assessment and the judgments of experts can be done. An analytical approach developed by the committee to conduct such an assessment is described in this chapter OVERVIEW OF THE ANALYTICAL APPROACH The analytical approach recommended for the conduct and application of risk assessment requires first a Conceptual framework and second, a risk model, For its conceptual framework, the committee adopted the well-accepted view of the role and nature of risk assessment developed in 1983 by the National Research Council's Committee on the Institutional Means for Assessment of Risks to Public Health. That committee proposed that risk assessment proceed in four steps (NRC, 1983~: Hazard identification: Determinations based on qualitative and quantitative evidence 9 of whether a particular agent (e.g., a chemical or microorganism) is or is not causally linked to particular health effects. Dose-response assessment: Determination of the relationship between the magnitude of exposure and the probability that a given health effect will occur. Exposure assessment: Determination of the extent of human exposure before or after application of regulatory controls. Risk characterization: Description of the nature and often the magnitude of human risk, including attendant uncertainty. Important in this conceptual framework is the relationship between risk management and these four steps of risk assessment. The 1983 risk assessment committee reported that: 30
31 Regulatory actions are based on two distinct elements, risk assessment . and risk management. Risk assessment is · . the use of the factual base to define the health effects of exposure of individuals or populations to hazardous materials and situations. Risk management is the process of weighing policy alternatives and selecting the most appropriate regulatory action, integrating the results of risk assessment with engineering data and with social, economic, and political concerns to reach a decis ion (NRC, 1983, p. 39. This concept leads to two important conclusions. First, the primary purpose of risk assessment is to support decisions regarding regulatory actions. Although there are nonregulatory uses of risk assessment corresponding to the full range of nonregulatory options for risk management, the regulatory orientation of the above quotation is appropriate to the responsibility of FSIS. The second conclusion is derived from the first and from the desire that risk assessments not prejudice or mislead decision makers. That is, to the greatest extent possible, risk assessments should be devoid of value judgments. Judgments are needed in risk assessment, but they should be judgments of science unbiased by the scientist's preferences for specific risk-management policies. Partitioning risk assessment into four standardized steps helps to ensure that there is a comprehensive accounting of the factors that determine risk and minimizes the policy and value judgments that might otherwise be inserted into the analysis. The key to risk characterization, the final step in the risk- assessment process, is the development and application of a risk model that guides the analyst in integrating and drawing conclusions from the first three steps: identification of the hazard, dose-response assessments, and exposure assessment. In a formal, quantitative risk assessment, the risk model consists of equations and mathematical algorithms and is often implemented as a computer code. These equations and algorithms may be developed to varying degrees of rigor, depending on the problem and the needs of the user. For qualitative risk assessment, the risk model is not formalized to permit rigorous numerical calculations. A qualitative model does, however, identify possible sources of risk and how they might be linked to health effects, and thus can be a valuable tool. By us ing such a model, investigators can conduct a formal review of the data demonstrating that a hazard exists, and organize information on dose-response relationships and exposures. Qualitative models are useful because they provide a logical framework for asking specific questions and deriving conclusions systematically. Qualitative risk
32 assessment is essential before quantitative risk assessment can be undertaken. A model to determine the risks of a process as complicated as poultry production and consumption must disaggregate the complex processes leading to the generation of those risks. The model must take into consideration points in the process at which hazard or risk agents are introduced into poultry or poultry products and modification of the quantities or characteristics of these agents by subsequent steps in the process. It can then be used to identify and determine the logical interrelationships of important critical factors that control the level of risk--that is, activities and events that introduce, alter, or determine the size of human health risks. In developing a model to assess the Herman health risks associated with poultry, the committee reviewed the principal risk agents associated with poultry production and consumption. These are summarized below and are described in more detail in Chapters 4 and 5. POULTRY RI SK AGENTS The agents responsible for nearly all the human health risks arising during the production and consumption of broiler chickens fall into two categories: . ^- THE RI SK MODEL Pathogenic microorganisms or their toxins. These agents, such as various species of Salmonella and Campylobacter, can transmit diseases to humans when present in or on infected or contaminated poultry tissues. Poultry-borne chemical residues. As described in Chapter 2, residues may be found in poultry intentionally given or exposed in other ways to chemicals before slaughter. After exposure, these chemicals may be concentrated and retained in the tissues for long periods. In general9 risk is dependent on the existence of three factors: A source from which risk agents are generated or released into the environment. For poultry production and consumption, the risk source encompasses all the activities related to poultry production, slaughtering, and processing. A route of human exposure to the risk agents, e.g., distribution and consumption of poultry products. A mechanism by which the exposures can generate adverse health effects, e.g., through microbial and chemical factors , which determine the health consequences resulting from human consumption or other contact with poultry products.
33 A risk model may be developed by identifying and linking all significant influences on these three factors. Figure 3-1 shows the major components, or submodels, of the risk mode1. Figures 3-2 through 3-6 present each submodel in greater detail, illustrating relevent factors and logical relationships. These submodels are briefly described in the following paragraphs. More detailed descriptions and references may be found in a report prepared by another National Research Council committee (NRC, 19851. Figure 3-7 combines the submodels to provide an overall account of the risk source, exposure, and health effects. The figures themselves are no more than a visual aid, and may be substantially amplified and revised as FSIS brings its full expertise to bear on risk assessment. Production Submodel The first maj or component of the risk model (Figure 3- 2) accounts for all risk factors associated with the production of live poultry. The wholesomeness and safety of poultry products depends on the health of the live birds, their feed, and the environment in which they are raised. Thus, management practices and production technologies are critical risk factors. Important production activities include breeding, hatching, feed milling, and poultry health care, each of which may affect microbiological or chemical hazards that reach the consumer . Methods of live poultry production may affect the wholesomeness of poultry products in several ways. During breeding and hatching, infection may be transmitted through the ovaries, through contaminated eggs in breeder flocks, or as a result of exposure to infectious agents in the hatchery. Infections may also occur during grow-out (Smitherman et al., 1984~. Other factors of concern are related to production facility management. For example, methods of feed storage can promote or prevent mat d growth and the production of mycotoxins , e. g., aflatoxin. Water contact ning infectious agents such as Salmonella also presents hazards (NRC, 1985, p. 128~. The condition of the feed provided during the first few days of a b~rd's life is particularly important, since it effectively establishes gut flora and the chick's ability to fight off future contaminations (Mead and Impey, 1985; Nurmi, 1985; Snoyenbos et al., 19789. The sanitation of the poultry housing facilities and methods of manure and sewage disposal can also have implications for public health. Studies of Salmonella ecology clearly establish the genetic stock, feed and feed ingredients, and environmental sources as critical points at which hazards from this microorganism may be controlled during
34 ~ _ _ ~ bye al TIC - :E X o in TIC e_ CC _ _ - o - `:L E I_ c "J ~ ~ =~ Co a D ~ ~ e_ me . _ ~ em t; ~ _, _ O "° Q He ~ ~ t O) · ·_ A) c, o me CLe O ~ m Us ~ = ~ `~ m ~ ~ _ fm N- . ~ I l C~ _ =1 ~ ._ m _ o ._ ~ S ~ `°L~ _ e. _ O E :, CD O ~ e~e V] <e~ .51~ · ~ 1 ,-1 H y o - '
~- Genetic I qual it 1 1 1 _, i | 1 Environmental I . Sanitation level contaminants ~ I, ~ ~ ~ ~ ~ Eggs _ r ~ , Chicks _ ~r, r' ~ Birds ~ Pesticides used Management practices and production technologies Microbial Food contamination additives , ~ ~ ~ ~ ~ ~ ' ~ ~ F' health ~ 3 ~ ~ . Medications Food-borne contaminants I Unhatchable I Mortality Egg-borne Infectious Infectious disease agents agents transmission PRODUCTION FIGURE 3-2 Production submodel. poultry production (Barnum, 1977; Bryan et al., 1976; NRC, 1969~. Despite precautions, however, exposures of poultry to pathogenic microorganisms such as Salmonella are difficult to avoid. Total confinement of primary breeding flocks has been studied as one means for reducing the transmission of Salmonella through eggs. fumigation techniques have been tried as well. There are methods to reduce the Salmonella contamination of processed feeds; however, feed contamination remains a widespread and serious problem (USDA, 1978> Vaccines administered to prevent economically important diseases may have public health consequences. In addition, most producers rely on a variety of prophylactic medications, especially antimicrobial agents and coccidiostats, to prevent or reduce the prevalence of infectious agents and parasites (North, 19841. The type of antibiotics as well as the quantity and time of their application on farms are critical factors, because the residues from these substances can remain in the tissues and present a health hazard to some people. Residues from drugs and medicated feeds are often found in carcasses when they are administered too soon before slaughter (USDA, 1984a). Drug
36 residues can also result from the accidental mixing of medicated and nonmedicated feeds at the feed mill, during transport to the farm, or at the farm itself. Antibiotic-resistant strains of pathogens may emerge as a result of treating birds or using antibiotics in feed. Chemical res idues can also result if the poultry are exposed to pesticides and other agricultural or industrial chemicals (Booth, 1982; Doull et al. 3 1980)e Pesticides may be applied to animals to control insects or internal parasites, but most exposures of poultry result from the application of pesticides to buildings, crops used for feed, or feed storage areas. Industrial chemicals used in electrical and mechanical equipment, e.g., polychlorinated biphenyls (PCBs) used in some electric power transformers, can also leave residues. If not detected in time, an accident, such as inadvertent contamination with PCB or hexachlorobenzene, could introduce into the food chain high levels of toxic substances that are not normally present (Booth, 1982; Doull et al., 1980)0 Thus' poultry production practices have the potential for affecting human health risks by determining whether and the extent to which various hazardous agents enter the poultry supply. They are often responsible for the diseases and contaminants that may be detected during inspection. At present, however, the Food Safety Inspection Service (FSIS) has no responsibility for monitoring the production phase . S laughter Submodel The next major component of the risk model (Figure 3-3) includes the risk factors related to slaughter and the inspection activities conducted during this process. The critical points in slaughtering operations include sanitary conditions during transport and during the slaughtering process itself, as well as antemortem and postmortem inspections and examinations for microbial and chemical contaminants. Live poultry is usually sent by truck to the slaughtering plant in specially built coops, baskets, or batteries. As noted in Chapter 2, antemortem inspection is discretionary and is designed to ensure compliance with regulations, with no apparent mechanism to selectively emphasize those regulations that relate to issues important to the public's health Nonetheless, most producers include this step in their own quality control programs, in part to provide early data on probable flock condemnation rates. Poultry raising practices are such that a lot den ivered for s laughter tends to have a common genetic and environmental background . In addition, disease incidence in a giver lot is likely to be either quite low or fairly high, resulting in distinct lot-to-lot distribution of condemnation rates . However, inspectors conducting antemortem inspection usually have no knowledge of the flock's history to aid them in identifying human health hazards.
37 - c, ° 2 _m ~ £ _ ~ _ ~ ~ .= 1 1 Yj 1 Mo 1 1 , 1 1 °1 1 -1 1 =~ 1 ~ 1 ol 1 c I I o 1 ~ 1 - oCR 1 1 ~ I 1 1 1 1 1 1 1 1 1 1 1 . =5 ~n ~ 8 Oo, _ _ c Q "c _ C E c~ ~ ~n v, ~v c O ic ~ .° _ ~ "> , e>_ _, . _ c~ <n c _ x ~ l c E ~ q, c o - - c C: a ~, . - ~ ~ c c ~:i ~ 0 ~ _ [ ._ _ ,_ ~ _ c G ~L, C ~SA- ~ ~ O ~ CJ ._ _ _ c' Q ~ ~n C ~0 a, 41 _ E e E _ t c 0 ._ _ c, ._ LU _ ~_ ~ q) 1 1 1 ~ ~ 1 c'm 1 ,= 1m' . ~:7 ~1 | ~ | 1 o°l 1 ,C C~ 1 i ~ ~ I 1 C U. 1 1 ~ '- 1 'x' 1 ,_~ 1 ~ ._ ~ m . 44 ~ o. ~ _ _ . _ 0 c ~ o o c~ _ - ~ E 1 ~ F o' ~ E 1 ~) tr LL I c 3 ,~ 1 1 o 3 1 a) ~n ~c ~' U] l H
38 During antemortem inspection, the FSIS inspector has an opportunity to observe birds between the time of arrival and when they are hung on the slaughtering line. Although birds may be rejected as a result of antemortem inspection, the criteria for condemnation are not always sensitive to the hazard posed to human health. For example, a bird developing heat stroke in transit is placed in the same public health category as one in terminal stages of septicemia salmonellosisO Birds that pass antemortem inspection are placed on a 1ine leading them to the various steps in the slaughtering process and to postmortem inspection. The first step in the slaughtering process is often, but not always, to stun the birds with an el ectrical shock Subsequently, the birds pass through steps that may affect human health risk: scalding, the removal of feathers, and the withdrawal of viscera. As in the production phase, the management practices and processing technologies used in these procedures can have a considerable impact on microbial loads. Most microbial contamination of the poultry's skin surfaces occurs during defeathering (ICMSF, 1980~. In one automated study of plant sanitation, Salmonella was isolated more often from pickers (machines provided with many rubber prongs called fingers, which remove all the feathers) than from any of the other equipment sampled, both before and after the start of processing (Campbell et al., 19841. A possible reason for this could be the complex construction of the pickers and the inherent difficulty of adequately cleaning all the picker fingers. The temperature of scald water (52°C; 126°F) and the thorough washing of poultry carcasses are critical in poultry slaughtering operations because of the potential for transfer and cross- contamination of Salmonell a and other microorganisms during defeathering (Green et al., 1982~. Perhaps the most important step of sanitary dressing is the proper removal of the gastrointestinal (GI) tract. Since Salmonella and other enteric bacteria originate in the digestive tract and fecal material of the slaughtered bird' it is extremely important to prevent contamination of the carcass by spilled GI tract contents or smeared fecal matter (ICMSF, 1980~. The likelihood of contamination increases in birds with localized or generalized diseases, infections, or contamination. After postmortem inspection, as noted in Chapter 2, the birds are passed as food or condemned in 1 of the 11 categories listed in Table 2-2. In fiscal year 1983, for all classes of poultry, less than 1% of the poultry examined were condemned during postmortem inspection (USDA, 1984c). In general, inspectors base their condemnation decisions on five criteria: condition of tissue (diseased or abnormal), type of disease (localized or generalized; acute or chronic), impairment of important body functions (e.g., uremia, icterus, toxemia), injurious to health of consumer (e.g. 9 tissues containing toxic chemical res idues or ~ nfectious agents), and
39 appearance (offensive or repugnant) (USDA, 1984b). Carcasses are usually condemned because of the presence of a visible anatomic lesion or specific condition (e.g., air sacculitis) rather than by cause (e.g., a specific infectious agent). The inspection system is not designed to detect human pathogens unless they produce an observable lesion. Neither pathogenic microorganisms that typically reside in the gastrointestinal tracts and on external surfaces of poultry nor chemical residues are generally detectable by routine organoleptic inspection procedures (i.e., sight, smell, or touch). Condemned carcasses and parts are promptly destroyed to prevent their entrance into the human food chain (Libby and Humphreys, 1975~. Info oration collected by the U.S. Department of Agriculture in previous years indicates "passed-b~rd error rates" (percentage of passed birds with gross and visible lesions) of approximately 1% to 1 . 5 % and that there are large inspector-to- inspector and day- to - day variations but small plant- to-plant variations. These estimates were obtained from 1969 to 1973, when condemnation rates were approximately 5%, and are not likely to be applicable now, when condemnation rates are roughly 1% (Booz-Allen & Hamilton, Inc ., 1977) . They indicate nevertheless, that error rates may equal a substantial fraction of condemnation rates. Carcasses contaminated with chemical residues and bacteria are generally not identifiable during inspection because these conditions are rarely visible. Therefore, these factors are not included in the error rate. Other aspects of the inspection process with public health significance are inspection of plant sanitation (Kauffman and Schaffner, 1974), monitoring of residues, and reinspection of carcasses see Chapter 2~. The residue monitoring program emphasizes the control of chlorinated hydrocarbon pesticides; however, evidence of a variety of other chemicals is also sought. During fiscal year 1983, 424 samples of young chickens were analyzed for 15 different chlorinated hydrocarbon residues. There were 73 positives, but only one that exceeded established tolerances (for chlordane) (USDA, 1986~. Some poultry producers operate their own residue programs, working with FSIS to detect violative residues and remove contaminated products from food channels. Chilling and freezing of poultry carcasses (broilers; fryers ~ at temperatures of 40~F (4°C) or less within 4 hours after slaughter are critical factors in inhibiting microbial growth (CFR, 19839. Carcass reinspection involves the sampling of carcasses that have passed routine postmortem inspection, dressing, and wash-and-chilling operations (see Chapter 2~. Final carcass washing is a risk factor influencing biological contamination levels. Germicides such as chlorine may be helpful, although the efficacy of chlorination under some conditions is not certain (NRC, 1985~. The design of final washers used on poultry evisceration lines (e.g., water pressure, nozzle type, location, and procedures) can also influence microbial counts.
40 Packing and Further Processing Submodel Packing and further processing of broilers and fryers are also critical factors (Figure 3-4~. Then carcasses are cut, pathogens on the surfaces of the carcasses, including species of Salmonella, Campylobacter, Clostridium, and Staphylococcus, can contaminate workers' hands, cutting boards, knives, tabletops, saws, and other pieces of equipment, and can then be transferred directly or via cleaning cloths to other equipment. The effectiveness and extent of efforts to minimize cross -contamination at this stage and to improve the sanitation of processing equipment constitute risk factors. The temperature in rooms for deponing, slicing, and storage and the durations of storage are also critical points that determine whether contaminating organisms multiply. For frozen products the critical points up to the time of freezing are the same as those for chilled products In addition, proper packaging, rapid freezing, and the time and temperature at which products are frozen and subsequently thawed also influence the counts of microorganisms (Peterson and Gunnerson, 1974~. For vacu~m-packed poultry products, it is important to maintain anaerobic conditions in a carbon dioxide and nitrogen atmosphere so that growth of the aerobic flora that commonly spoil unpackaged raw poultry can be inhibited. For dried poultry, the moisture content should be lowered enough to provide shelf stability It is therefore essential that the drying rapidly decrease the Aw of these products to levels at which pathogens do not multiply. During the drying process, therefore, the rapidity of the procedure and temperature control are critical points. For cooked, uncured products, the quality of the raw ingredients is a critical control point as are the duration of cooking (which should be sufficient to kill yeasts, molds, parasites, and viruses), the temperature of cooking (130- 167°F, or 54- 75°C), the rate of cooling (pathogens can multiply in cooked products held too long at certain temperatures), the handling of the products after cooking (e.g., an entry point for Salmonella)(Bryan, 1980), equipment sanitation, and subsequent cold storage (when micrococci, streptococci, and other psychrophilic bacteria may multiply). For uncured canned products, acidity is important. Products with a pH of 4.6 or less are considered high-acid products and need only heating to ensure shelf stability Low-acid uncured canned foods (pH greater than 4.6) must be given time-temperature exposures that kill up to 1012 Clostridium botulinum spores. For cured canned poultry, critical factors are proper curing (including adequate salt and nitrite concentrations), quality of LAW (water activity) is the ratio of the water pressure of a food to that of pure water at the same temperature. It is the measure of water in food available for use by microorganisms that have specific cardinal requirements for Aw.
41 Poultry products Broilers/Fryers Water . availability Aw1 ~ 1. 1 Cut, processing,)_ ~ Anaerobic conditions Product pH - I and packing , I PACKING AND l FURTHER PROCESSINGJ FIGURE 3-4 Packing and further processing submodel. ingredients (with reference to microbes ), level of spore contamination, drying process (proper A`,'), container integrity, and proper cold storage. The pathogens of most concern in high Aw cured meats are Ral mon~1 1 ~ St~nhvl oc~ruc: ~ w ~ , and C . botulinum ~ Ito ~ 1974; Sebald and Jouglard, 1977 ~ . The addition of chemicals during processing of poultry is another critical factor. Violative amounts (amounts added above the set tolerance levels) could result in adverse health effects. There are aproximately 1,800 food additives, most of which are flavors and antioxidants. Probably 1% of these are used in poultry products specifically (Joint FAD/WHO Expert Committee on Food Additives, 1983~. Other factors with public health implications are concentrations of chlorine used to reduce microbiological contamination (Cantor, 1982; Mead et al., 1975; Zoeteman et al., 1982~; toxic compounds (e.g., polycyclic bydrocarbons) produced as a result of heating fresh, smoked, grilled, and cured meats (Meyer, 1960~; bone and metal fragments that enter poultry during processing ; leaching of chemicals from the packaging material into the food product (e.g., polyvinyl chloride, acrylonitrile) (Karel and Heidelbaugh, 1975; Sacharow, 1979~; and the intrinsic chemical (e.g. , polymerization, free-radical formation) and physi Cal (e . g. , consistency, texture, color) changes that occur in tissues during improper storage (Ames, 1983; Lee, 1983; Roberts, 1981) Distribution and Preparation Submodel After processing, microorganisms in poultry products can multiply, spread, and perhaps cross-contaminate other foods unless the products
42 are handled properly in the plant, during transport, in retail outlets, and by the consumer after purchase. Critical risk factors during distribution are the microbial load at the time of shipment, the internal temperature at the time of loading, and the air temperature and movement in the transport vehicle and storage warehouse ~ Figure 3-5~. Insect and rodent control may also be critical during storage of paper- and plastic-packaged dry products. In retail stores9 walk-in refrigerators and display cases and cleanliness of cutting boards and blocks, grinders, saws, tenderizers, and cutting utensils are key control points. The critical factors during preparation in food service establishments and in homes are generally beyond FSIS control, but vary with the product, equipment, and food-service system. The major opportunities for outbreaks of food-borne diseases in both food-service establishments (Bobeng and David, 1977; Bryan, 1981a,b; Bryan and McKinley, 1974; Unklesbay et al., 1977) and homes (Zottola and Wolf, 1980) are presented when cooked foods are held at or near room temperature, when cooked foods are stored in large containers during refrigerated storage, when food is prepared a day or more before serving (Bryan, 1978, 1980), and in any type of food processing plant (Bauman, 1974; DREW, 1972; Ito, 1974; Kauffman and Schaffner, 1974; Peterson and Gunnerson, 1974; WHO/ICMSF, 1982~. Other frequently identified contributory factors are inadequate cooking, inadequate reheating, contamination by infected persons, and cross-contamination by inadequately cleaned equipment (e.g., cutting boards, knives, slicers, grinders, tabletops, and storage pans). Several precautions that can be taken to help avoid these hazards are washing hands after handling raw poultry or when returning to work stations; preventing cross-contamination by thoroughly cleaning utensils, equipment, tabletops, sponges, and cleaning cloths; cooking poultry thoroughly; avoidance of holding cooked poultry at room temperature for long periods; cooling rapidly in shallow containers any leftovers or food prepared for consumption on subsequent days; and reheating leftovers thoroughly. Programs designed to provide information on such matters as Hazard Analysis Critical Control Point principles, poultry science, food microbiology, and safe food handling operations will all help in reducing exposure levels. Health Effects (Consumption) Submodel Enteric bacterial agents are primarily health hazards for the consuming public. Less commonly, they are occupational hazards in the meat packing industry. The major agents are Salmonella, Campylobacter, and Clostridium perfringens (Bryan, 1980~. Each year as many as 2 million Americans are affected by salmonellosis--an important health effect that is clearly linked to poultry (Figure 3-6) (USDA, 1978~; however, only a fraction of these cases are ever reported. Antimicrobial-resistant salmonellae account for a steadily increasing number of salmonellosis cases in the United States and can be traced to food animals, including poultry (Holmberg and Blake, 1984~. This
FIGURE 3 - S iGW4_ · Pain and suffering · Medical costs · Hospital ization cases · Lost income · outbreaks ~ 1 Health I effects | · individual | FIGURE 3 - 6 Treatment and l hospital ization costs i Other foods' I consumed I and their buffering capacity _ Determinants of Susceptibility Prior exposures CONSUMPTION 43 1 l Education/ l information 1 _ I Consumer/ commercial food-handler I awareness Meats ~ Food ~ rr preparation l ~ 1 , ~ 1 1 l T Adequacy of washing Cross-contamination of other foods D ISTR I BUTION AND PR EPARATION Time, temperature conditions Package integrity Storage ~ | Retai 1 Insect and rodent control j Poultry , products Distribution and preparation submodel. Populations · very young I · very old I · debilitated Meals Consumption submodel.
44 disease accounts for much pain and suffering and is costly in terms of medical care, hospitalization, and lost income through absence from work. Campylobacter presents a somewhat s imilar health problem . Although this microorganism tends not to multiply in food at room temperatures (Skirrow, 1982), it can survive on chilled carcasses for months (Oosterom et al., 1983~. Since antemortem and postmortem inspection can rarely identify Campylobacter or Clostridiu~ perfringens (CDC, 1983) in infected poultry, prevention of carcass contamination by fecal matter is a critical control point, as it is for salmonellae. Toxic chemicals, including carcinogens, may also contribute to adverse health effects in humans. For example, polycyclic aromatic hydrocarbons (PAHs ) are commons y found in fresh, smoked, grilled, and cured poultry (NBC, 1985~. Of the more than 100 PAH compounds identified, 5 are known carcinogens when given orally, and 3 of these 5 are part of the average U. S . diet. The impact of these compounds on humans at ordinary dietary levels throughout a lifetime is not known (Meyer, 1960 ~ . USE OF THE RISK MODEL The risk model described above accounts for all maj or elements involved in bringing poultry products to the consumer and some specific factors that may influence the safety and wholesomeness of the products and the consequent health impacts on consumers (Figure 3 - 7 ~ . The model provides a basis for risk assessment in that it serves as a logical road map that can be used in the evaluation of current or proposed strategies for reducing and controlling risks. The method of assessment consists of evaluating the extent to which the control strategy affects each critical control point and then determining the implications of these effects. To the extent that quantitative data exist, a well-defined, quantitative measure of risk may be determined through application of logic inherent in the risk model. In the absence of such data, qualitative judgments may still be used to develop a logical and rigorous evaluation of alternative strategies. The risk model also suggests a generalized logic for investigating and comparing alternative strategies for controlling human health risks O Specifically, evaluation of the relative effectiveness of alternative controls for a specific risk requires the following steps: 1. Assess the magnitude of the health risk of concern. For example, how serious is the risk of ingesting microbial contaminants in poultry? 20 Review the risk model to determine its adequacy for the intended purpose, and use it to identify and organize the critical factors that determine or influence the magnitude of
; I_. ~ ~ - ~ To 3 1 1 i: ~ 1 2 v E 1 o ' o U) : t: ~ _~ o C _ A- do. - C ~ ~ ~ ~ ~ _ y .... - o U a En 2t · ~) ~.0 ' .- l EM
46 the risk. For example, the number and severity of potential health effects may be influenced by the level of microbial contamination of feed, the degree of contamination of the carcass during evisceration, the extent of cross-contamination during cutting, and other factors. 3. Compare the degree to which alternative strategies can affect the critical factors. 4. Use the conceptual risk model to determine how impacts on critical factors relate to incremental changes in the likelihood that health effects will occur in consumers and the magnitude of those effects. Data are currently insufficient to permit the last step to be based on formal quantitative models. Nevertheless, this task can be performed qualitatively by using the risk model to guide the professional judgments necessary to reach a qualitative conclusion about the impacts on human health. As appropriate quantitative data are collected, the various steps outlined above may be used to produce quantitative measures of risk. Regardless of whether risk estimates are quantitative or qualitative, such assessments are essential to the identification and selection of FSIS strategies for protecting public health. Described below is a generalized logic for incorporating risk assessment into the planning process. USING RISK ASSESSMENTS TO PROTECT HEALTH A substantial assurance that health risks do not reach unacceptable levels is surely a critical function of the FSIS program. Potential adverse health consequences vary in frequency and severity, and evidence concerning them ranges from nearly nothing to conclusive. An optimal health-protection program (including inspection) can exploit these variations both to increase effectiveness of inspection procedures and to reduce costs of their implementation. Put simply, it is important to determine what types of risk and evidence pertaining to that risk justifies what set of actions. Limited information and uncertainties create difficulties in each step of risk assessment. Hazard identification is often uncertain; e.g., there may be only poor evidence on the clinical significance of certain strains of Salmonella found in broiler chickens or on the carcinogenicity of a chemical residue. Dose-response relationships are hard to establish with precision and may vary from time to time, place to place, and (especially) consumer to consumer. It is also difficult to assess exposures with any degree of completeness or certainty, since only a small percentage of broiler chickens can be tested for specific chemical or microbial hazards. Because of these and countless similar problems knowledge about the health risks attributable to broiler chickens is somewhat fragmentary and uncertain.
47 The current FSIS inspection program is relatively inflexible and is not well suited for providing information needed to resolve these uncertainties. For example, every bird must pass organoleptic inspection, and line speeds are fixed at specific rates. Sample sizes for testing and surveillance of chemical residues are also fixed; no opportunity is provided for adjusting the detection sensitivity to changing perceptions regarding the magnitudes of the hazards or the likelihood that residues will be found. Opportunities for resolving uncertainties are missed; for example, birds condemned during organoleptic inspection are not examined to learn what they can teach regarding the assessment and abatement of health risks. The inspection itself cannot ordinarily distinguish contaminated from uncontaminated birds, whether the contamination is microbial or chemical. Information might be collected to build a base of knowledge for improving inspection effectiveness. For example, if the first, approximately random 10% of a flock is unusually healthy, the remaining 90% is likely to be healthy too. An infection in one grow-out house may well be present in an adjacent house. If a prohibited practice is detected at one place or time in a slaughter operation, it is pass ible that there may be other prohibited practices at other places and times in that operation. Indeed, inspectors often believe quite strongly that they can distinguish good flocks and good operators from bad ones. This continuum in knowledge and the parallel continuum in risk are not, on the whole, reflected in either the strategies of FSIS for inspection or the more general control of health hazards. The methods used cannot be tailored to changing situations and do not maximize the potential for learning. Thus, an opportunity exists to improve both the ruble c health and the cost-effectiveness of the FSIS inspection procedures by adopting procedures capable of account' ng for substantial gradations in both risk and knowledge about risks in general and about circumstances contributing to risk. To protect health, inspection should be deliberately and objectively targeted to maximize, for a given expenditure of resources, the return in reduced morbidity and mortality. This return will vary from one activity to another, from one producer to another from one time to another, and in other ways. Formal risk assessment is the key to capturing ~ organizing ~ and interpreting the evidence that can be brought to bear on the optimum use of inspection resources. It will identify problems in a clear and organized way; produce the best possible estimate of the likelihood that a risk exists, and if so, its size; and make clear the kinds and degrees oF uncertainty attached to that estimate. This information can then be translated into a detailed strategy to protect human health, including the development of specific regulations and instructions to inspection staff. A comprehensive strategy for reducing the health risks attributable to poultry should be based on a broad conceptual model of how those risks arise and on assessments of the nature and magnitude of those
48 risks made at the highest level of precision attainable. Furthermore: the risk reduction strategy should be cost effective and designed to take advantage of the full range of tools available to FSIS. Any health protection program is likely to involve several steps, but there is no agreed-upon classification of these steps, such as there is for the four steps in risk assessment. Some of these activities are listed below. Their order should not be construed as an implication of their priority · Establish objectives and set priorities . · Identify and analyze alternatives for achieving priority obj ectives such as the following: e . g., 100% organoleptic inspection or a requirement for a withdrawal period after use of a drug. Identify potential hazards and set tolerances or action levels (targets and goals) for each, including considerations of possible synergistic interactions. Select and implement a control program. Conduct monitoring and surveillance and interpret results. Take appropriate steps to ensure compliance (by producers) and enforcement (by FSIS). Conduct research to improve the model' the risk assessments, or the data on which they are based. These kinds of activities should be developed as an integrated package; they are substantially less independent than the steps in risk assessment. Several categories of such activities are discussed below. Establish Priorities Any feasible program for risk management will require establishment of priorities for reducing or eliminating factors that contribute to risk. Such priorities and the frequency and intensity of monitoring should be based on risk assessment, but for most kinds of FSIS interventions it is only the relative risks of the various alternatives that are of concern. For example, in choosing among postmortem inspection strategies, it is the potential increase ox decrease from current levels of health risks that is most important. Thus, there would be much value in having a scheme for relative ranking of all poultry-borne risks, including microbial hazards, even if the absolute magnitudes of those risks remain uncertain. In Chapter 5, for example, it is shown how different carcinogenic hazards can be ranked on a common qualitative scale. A broader system of ranking may require a substantial research program as well as field testing. A scheme for assessing relative risks need not include estimation of the absolute risk of any of the substances to be ranked It is necessary only that the scheme incorporate in a systematic way some measures of both pathogenicity and exposure. Both of these components are essential and must be determined to an adequate degree of
49 accuracy. For example, the data on Class 1, 2, and 3 chemicals (see Chapter 5) vary widely in quality and content, and these differences must be taken into account in a systematic way. The primary purpose of a relative risk assessment is to ensure that two major risk-management activities--monitoring of feed and water and monitoring of poultry products (including whole birds)--are given a level of emphasis reflecting the probability that the risk factor will be found in food intended for human consumption, the level of risk that may result if it escapes detection, and the extent to which the risk can be reduced. Identify Problems of Risk Management, and Set Acceptable Levels of Risk After analyzing a broad range of hazards, Lowrance (1976) concluded that ''a thing is safe if its risks are judged acceptable." This is not a tautology; perfect safety is a chimera, but very small risks may be deemed acceptable when the costs (in the broadest sense) of their further reduction exceed the expected benefits. For example, FDA does not in general concern itself with carcinogens that are believed to produce cancer risks of less than one per million persons exposed at the maximum allowable dose over a lifetime. Acceptability of risk depends on many things, including the type of outcome (e.g., skin rash vs. cancer), whether the risk is already common or familiar, and whether the risk is known to exposed persons and assumed voluntarily (Fischhoff et al., 1978~. Perceptions of risk to health are important whether or not they are in line with the best quantitative estimates. Both the definition of risk and the determination that certain risks cannot be eliminated at an acceptable cost are generally very difficult. Risk management, including the setting of acceptable levels of risk, is a political rather than scientific task and hence outside the committee's purview. The committee notes, however, that risk management should ordinarily be based on the best available scientific risk assessment and an objective analysis of the appropriate role of risk assessment in risk management. Furthermore, risk-management goals should be precisely stated in quantitative, evaluable terms. For example, a maximum tolerance level below the lowest detectable level (microbial or chemical) would be unenforceable. Monitoring and Surveillance There will be a continuing need for monitoring and surveillance to ensure that FSIS program goals are met--goals that can be partitioned into structure, process, and outcome. It will also be necessary to fine-tune inspection activities and to update the system to keep pace with changing risks and production practices. Of particular importance for the monitoring program is the selection of sampling rates. Statistical sampling strategies can be devised to ensure, with a specified degree of confidence, that products
so containing excessive levels of chemical residues are identified for removal from the food supply. The desirable degree of confidence for potentially high risk substances should be greater than for other substances. In Chapter 5 of this report, the committee recommends specific criteria for chemical residues posing different levels of potential risk to public health. Chapter 6 describes a risk ranking scheme, which can be used not only to develop monitoring strategies, but also to aid other data gathering efforts. Risk assessment plays several important roles in a program to control or eliminate health hazards posed by broiler chickens. All are based on applying, with varying degrees of rigor, the elements of a conceptual model of risk and rely on the use of specific types of data. An effective risk management scheme will require each of these program elements, although not all need to be within the direct control of FSIS (indeed, some are already established at FDA and EPA). Nevertheless, it is important that FSIS ensure adequate coverage of the full range of activities needed in risk management and that the agency acquire substantial knowledge of the adequacy of each activity. REFERENCES Ames, B. N. 1983. Dietary carcinogens and anticarcinogens: Oxygen radicals and degenerative diseases. Science 221:1256-1264. Barnum, D. A., ed. 1977. Proceedings of the International Symposium on Salmonella and Prospects for Control held at the University of Guelph , June 8 - 11, 1977 . University of Guelph , Guelph , Ontario, Canada. 200 pp. Bauman, H. E. 1974. The HACCP concept and microbiological hazard categories . Food Technol . 28: 30, 32 9 34, 74. Bobeng, B. J., and B. D. David. 1977. HACCP models for quality control of entree production in food service systems e J. Food Protect. 40:632-638. Booth, N. H. 1982. Drug and chemical residues in the edible tissues of animals. Pp. 1065-1113 in N. H. Booth and L. E. McDonald, eds. Veterinary Pharmacology and Therapeutics, 5th ado Iowa State University Press, Ames, Iowa. Booz-Allen & Hamilton 9 Inc. 1977. Study of the Federal Meat and Poultry Inspection System, Vol. II--Opportunities for Change-An Evaluation of Specific Alternatives. Prepared for the U.S. Department of Agriculture. Contract No. 53-3142-6-3614. U.S. Department of Agriculture, Washington, D.C. 351 pp.
51 Bryan, F. L. 1978. Factors that contribute to outbreaks of foodborne disease . J . Food Protect . 41: 816- 827 . Bryan, F. L. 1980. Foodborne diseases in the United States associated with meat and poultry. J. Food Protect. 43:140-150. Bryan, F. L. 1981a. Hazard analysis critical control point approach: Epidemiologic rationale and application to foodservice operations. J . Environ. Health 44: 7-14. Bryan, F. L. 1981b. Hazard analysis of food service operations. Food Technol. 3S:78-87. Bryan, F. L., and T. W. McKinley. 1974. Prevention of foodborne illness by time-temperature control of thawing, cooking, chilling and reheating turkeys in school lunch kitchens. J. Milk Food Technol. 37:420-429. Bryan, F. L., H. Anderson, R. K. Anderson, K. J. Baker, H. Matsuura, T. W. McKinley, R. Swanson, and E. Todd. 1976. Procedures to Investigate Food-borne Illness, 3rd ed. International Association of Milk, Food and Environmental Sanitarians, Ames, Iowa. 74 pp. Campbell, D. F., R. W. Johnston, M. W. Wheeler, K. V. Nagaraja, C. D. Szymansaki, and B. S. Pomeray. 1984. Effects of the evisceration and cooling processes on the incidence of Salmonella In fresh dressed turkeys grown under Salmonella-controlled and uncontrolled environments. Poult. Sci. 63:1069-1072. Cantor, K. P. 1982. Epidemiological evidence of carcinogenicity of chlorinated organics in drinking water. Environ. Health Perspect. 26:187-195. CDC (Centers for Disease Control). 1983. Foodborne Disease Surveillance, Annual Summary 1981. HHS Publ. No. (CDC) 83-8185. Centers for Disease Control, Public Health Service, U. S . Department of Health and Human Services, Atlanta. 41 pp. CFR (Code of Federal Regulations). 1983. Title 9, Animals and Animal Products; Section 381.66, Temperatures and chilling and freezing procedures. U. S. Government Printing Office, Washington, D.C. DREW (U.S. Department of Health, Education, and Welfare). 1972. Proceedings of the 1971 National Conference on Food Protection held in Denver, Colorado, April 4-8, 1971. Sponsored by American Public Health Association. DREW Publ. No. (FDA) 72-2015. U.S. Food and Drug Administration, U.S. Department of Health, Education, and Welfare, Washington, D.C. 242 pp.
52 Doull, J., C. D. Klaassen, and M. 0. Amdur, eds. 1980. Casarett and Doull' s Toxicology- -The Basic Science of Poisons, 2nd ed. Macmillan, New fork. 778 pp. Fischhoff , B ., P . Slovic , S . Lichtenstein, S . Read, and B . Combs . 1978. How safe is safe enough? A psychometric study of attitudes towards technological risks and benefits. Policy Sci. 9:127-152. Green, Se S. ~ A. B. Moran, R. W. Johnston, P. Ubler, and J. Chiu. 1982. The incidence of Salmonella species and serotypes in young whole chicken carcasses in 1979 as compared with 19670 Poult. Sci. 61:288 - 293 . Holmberg, S. D., and P. A. Blake. 1984. Staphylococcal food poisoning in the United States: New facts and old misconceptions. J. Am. Med. Assoc. 251:487-489. ICMSF (International Commission on Microbiological Specifications for Foods ~ . 1980 . Microbial Ecology of Foods , Vol . II: Food Commodities. Academic Press, New York. 664 pp. Ito, K. 1974. Microbiological critical control points in canned foods ~ Food Technol. 28 :46, 48. Joint FAD/WHO Expert Committee on Food Additives. 1983. Evaluation of Certain Food Additives and Contaminants. 27th Report of the Joint FAD/WHO Expert Committee on Food Additives, Technical Report Series No. 696. World Health Organization, Geneva. 47 pp. Karel, M., and No D ~ Heidelbaugh. 1975. Effects of packaging on nutrients. Pp. 412-462 in R. S. Harris and E. Karmas, eds. Nutritional Evaluation of Food Processing, 2nd ed. AVI Publishing, Westport, Conn. Kauffman, F. L., and R. M. Schaffner. 1974. Hazard analysis, critical control points and good manufacturing practices regulations (sanitation) in food plant inspections. Pp. 402-407 in the Proceedings of the IV International Congress on Food Science and Technology held in Madrid, Spain, September 22-279 1974. Instituto de AgroquLmica y Technologia de Alimentos, Valencia, Spain. Lee, F. A. 1983. Basic Food Chemistry, 2nd ed. AVI Publishing a Westport' Conn. 564 pp. Libby, J. A., and MO R. Humphreys. 1975. Post-mortem dispositions. Pp. 8S-186 in J. A. Libby, ed. Meat Hygiene, 4th ed. Lea & Febiger, Philadelphia. Lowrance, W. W. 1976. Of Acceptable Risk. Science and the Determination of Safety. William Kaufmann, Los Altos, Calif. 180 PP ~
53 Mead, G. C., and C. S. Impey. 1985. Control of Salmonella colonization in poultry flocks by defined gut-flora treatment. Pp. 72-79 in G. H. Snoyenbos, ed. Proceedings of the International Symposium on Salmonella held in New Orleans, Louisiana, July 19-20, 1984. American Association of Avian Pathologists, Kennett Square, Pa. Mead, G. C., B. W. Adams, and R. T. Parry. 1975. The effectiveness of in-plant chlorination in poultry processing. Br. Poult. Sci. 16:517-526. Meyer, L. H., ed. 1960. Food Chemistry. Reinhold Organic Chemistry and Biochemistry Textbook Series. Reinhold, New York. 385 pp. North, M. O. 1984. Commercial Chicken Production Manual, 3rd ed. Animal Science Textbook Series. AVI Publishing, Westport, Conn. 714 pp. NRC (National Research Council). 1969. An Evaluation of the Salmonella Problem. Report of the Committee on Salmonella, Division of Biology and Agriculture. National Academy of Sciences, Washington, D . C . 207 pp . NRC (National Research Council). 1983. Risk Assessment in the Federal Government: Managing the Process . Report of the Committee on the Institutional Means for Assessment of Risks to Public Health, Commission on Life Sciences. National Academy Press 9 Washington, D.C. 203 pp. NRC (National Research Council). 1985 . Meat and Poultry Inspection : The Scientific Basis of the Nation's Program. Report of the Committee on the Scientific Basis of the Nation's Meat and Poultry Inspection Program, Food and Nutrition Board. National Academy Press, Washington, D.C. 209 pp. Nurmi, E. 1985. Use of competitive exclusion in prevention of salmonellae and other enteropathogenic bacteria infections in poultry. Pp. 64-71 in G. H. Snoyenbos, ed. Proceedings of the International Symposium on Salmonella held in New Orleans, Louisiana, Jul y 19 - 20 , 1984 . American Association of Avian Pathologists, Kennett Square, Pa. Oosterom, J ., G . J . A o De Wilde ~ E . De Boer , L. H . De Blaauw, and H . Karman. 1983 . Survival of CampYlobacter j ejuni during poultry processing and pig slaughter~ng. J. Food Protect. 46:702-706. Peterson, A. C., and R. E. Gunnerson. 1974. Microbiological critical control points in frozen foods. Food Technol. 28:37-44.
54 Roberts, H. R., ed. 1981. Food Safety. Wiley, New York. 339 pp. Sacharow, S . 1979 . Packaging Regulations . ANTI Publishing, Westport, Conr`. 207 pp. Sebaid, M., and J. Jouglard. 1977. Aspects actuels du botulisms. Rev. Prat. 27:173-176. Skirrow, M. B. 19820 Campylobacter enteritis°-The first 5 years. J. Hyg. 89: 175-184. Smitherman, R. E., C. A. Genigeorgis , and T. B. Farver. 1984. Preliminary observations on the occurrence of Campylobacter iej uni at four Cali fornia chicken ranches . J O Food Protect . 47: 293 - 298 . Snoyenbos, G. H., 0. M. Weinack9 and C. F. Smyser. 1978. Protecting chicks and poults from salmonellae by oral administration of "normal gut microfloraO '' Avian Dis. 22: 273-287 . Unklesbay, N. F., R. B. Maxcy, M. E. Knickrehm, K. E. Stevenson, M. L. Cremer, and M. E. Matthews. 1977. Foodservice systems: Product flow and microbial quality and safety of foods. Research Bulletin 1018. College of Agriculture, Agriculture Experiment Station, University of Missouri, Columbia, Mo. 36 pp. USDA (U.S. Department of Agriculture). 1978. Recommendations for Reduction and Control of Salmonellosis. Report of the U.S. Advisory Committee on Salmonella. Food Safety and Quality Service, U.S. Department of Agr~culture, Washington, D.C e 30 pp ~ USDA (U.S. Department of Agriculture). 1984a. FSIS Facts: The National Residue Program. FSIS-18. Food Safety and Inspection Service, U.S. Department of Agriculture, Washington9 D e C ~ 4 pp. USDA (U.S. Department of Agriculture). 1984b. Livestock Carcass Disposition Review. Progr~ Training Division, Meat and Poultry Inspection Technical Ser~rices, Food Safety and Inspection Service, U.S. Department of Agriculture, Denton, Tex. 74 pp. USDA (U.S. Department of Agriculture). 1984c. Statistical S~mmary: Federal Meat and Poultry Inspection For Fiscal Year 1983. FSIS-14. Food Safety and Inspection Service, U.SO Department of Agriculture, Washington, D.C. 34 pp. USDA (U. S . Department of Agriculture) . 1986 . Domestic Residue Data Book: National Residue Progra~n 1983. Science Program, Food Safety and Inspection Service, U.S. Department of Agriculture, Washington, D.C. 95 pp.
55 WHO/ICMSF (World Health Organization/International Commission on Microbiological Specifications for Foods). 1982. Report of the WHO/ICMSF Meeting on Hazard Analysis: Critical Control Point System in Food Hygiene. World Health Organization, Geneva. Zoeteman, B. C. J., J. Hrubec, E. de Greef, and H. J. Kool. 1982. Mutagenic activity associated with by-products of drinking water disinfection by chlorine, chlorine dioxide, ozone and W -irradiation. Environ. Health Perspect . 46: 197 - 205 . Zottola, E . A., and I. D. Wolf. 1980 . Recipe hazard analysis--RHAS--a systematic approach to analyzing potential hazards In a recipe for food preparation-preservation. J. Food Protect. 44:560-564.