National Academies Press: OpenBook

Scientific Criteria to Ensure Safe Food (2003)

Chapter: 4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products

« Previous: 3 Food Safety Tools
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 133
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 134
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 135
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 136
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 137
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 138
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 139
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 140
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 141
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 142
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 143
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 144
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 145
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 146
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 147
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 148
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 149
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 150
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 151
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 152
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 153
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 154
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 155
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 156
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 157
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 158
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 159
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 160
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 161
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 162
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 163
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 164
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 165
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 166
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 167
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 168
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 169
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 170
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 171
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 172
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 173
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 174
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 175
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 176
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 177
Suggested Citation:"4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products." Institute of Medicine and National Research Council. 2003. Scientific Criteria to Ensure Safe Food. Washington, DC: The National Academies Press. doi: 10.17226/10690.
×
Page 178

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.

4 Scientific Criteria and Performance Standards to Control Hazards in Meat and Poultry Products DESCRIPTION OF THE MEAT AND POULTRY INDUSTRY Animal production in the United States has undergone a transformation over the last 50 years from a system mainly comprised of independent animal producers to one mainly comprised of concentrated animal feeding operations. The major production animal species, beef cattle, swine, chickens, and turkeys, are produced under a variety of conditions that may have significance in regard to the presence or absence of potential foodborne pathogens. The following is a brief synopsis of animal production in the United States. Beef A major percentage of the world's beef is produced in the United States both for domestic use and for export. The U.S. fed-cattle industry is the largest in the world (ERS,2000~. Most beef produced in and exported from the United States is the grain-finished, high-quality, choice-cut variety, while imported beef is gener- ally grass-fed and is used primarily for processing as ground beef (ERS, 2002~. Red meat production is a concentrated industry. Feedlots and steer and heifer slaughter facilities are geographically concentrated in the Great Plains (MacDonald et al., 2000~. Iowa, Kansas, Nebraska, and Texas accounted for over 51 percent of the U.S. commercial red meat production in 2001 (NASS, 2002~. Since cows generally move directly to plants from dairy farms and beef cow-calf operations, cow and bull sales and slaughter plants are more widely distributed across the country (MacDonald et al., 2000~. In commercial plants, red meat 133

34 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD production totaled 45.7 billion pounds in 2001, of which beef production accounted for 26.2 billion pounds (NASS, 2002~. Four companies slaughter and process 82 percent of the beef in the United States (MacDonald et al., 2000; REAP, 2001~. Twenty percent of beef consumed originates from cull cows of the dairy industry (University of Vermont, 2003; Wallace, 2003~. Over 25 years ago, most beef was sold as whole or half carcasses that were fabricated by other processors or retailers. The advent of boxed meat (i.e., assem- bly cut and packaged meat) revolutionized the beef industry so that most fresh beef is sold as vacuum-packaged primals (large sections of a carcass cut for wholesale, such as the round, chuck, or rib) and subprimals (retail cuts) (Kinsman, 1994~. Case-ready beef (retail cuts packaged and brand labeled) is a new concept currently being embraced by some companies (Eilert and Rathje,2001~. Processed beef products (i.e., those in which the carcass identity is lost or that are subject to some treatment that affects its texture, color, and flavor) accounted for 13.9 per- cent of beef consumed in 2001 (Nalivka, 2002~. Poultry The U.S. poultry industry is comprised primarily of three segments: broilers, turkeys, and eggs. Of these three, broilers (i.e., young chickens) dominate with 66 percent of the dollar value of production (Nalivka, 2002~. The United States produced more than 8.2 billion chickens, 2.6 billion turkeys, and more than 71 billion table eggs in 2000. The U.S. broiler and turkey industries are referred to as "vertically inte- grated." The company or integrator controls all aspects of the process but con- tracts with individual landowners for growing services. The landowners furnish the poultry houses, energy, and labor, while the companies furnish the animals, feed, and technical support. The basic unit of this arrangement is the "complex," which consists of parent flocks, multiplier flocks, hatchery, feed mill, and processing plant (Figure 4.1~. Breeder farms, also called multiplier flocks, supply all of the eggs that will become the chickens for processing. For each day of processing, the hatchery must hatch enough chicks to account for losses in the field and for a standard amount of weight gain to match sales projections for the time period when these birds will be processed. The feed mill must supply feed for all of the houses within the complex to ensure that no chicken goes hungry. The complex also usually has water treatment facilities and also may have rendering capabilities for by-products. The typical complex processes over 1 mil- lion chickens per week. A typical young broiler plant can have from one to four processing lines. The maximum speed of each line is determined by the amount of inspection in place from the U.S. Department of Agriculture's (USDA) Food Safety and Inspection Service (FSIS). The categories of inspection are:

135 E I o .O a) a) C) Q ._ ~ $ ~ a) rr mace ~ 0 ~ ~ E a, o 2 `, ~~ Lo ~ Q m ~ of a) =5 a) a) IL ~a) In a) . _ Q Q U) In m o o o In IL 3 E IL o o o E IL o o o o .= s°- o

36 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD · The Streamlined Inspection System, which allows 70 birds/min with two inspectors per evisceration line (35 birds/min/inspector) The New Enhanced Line Speed, which allows 91 birds/min with three inspectors and additional plant inspection (30.3 birds/min/inspector) The New Evisceration Systems: Maestro (Meyn Poultry, Gainesville, GA) and Nu-Tech (Stork Gamco, Gainesville, GA), which allow 140 birds/ min with four inspectors per line (35 birds/min/inspector). Pork The United States is a major pork producer, second only to China. The U.S. pork industry rapidly expanded during the 1990s; more pork was produced (nearly 19 billion pounds) and more hogs slaughtered (more than 99 million head) in the United States in 1998 than ever before. Previous records in production had been set in 1992, 1994, and 1995. Approximately 85,000 pork producers are in business today compared with nearly 3 million in 1950. Farms have grown in size; over 80 percent of the hogs are grown on farms producing 1,000 or more hogs per year, while over half are grown on farms producing 2,000 or more hogs per year. These operations, which are often very technically sophisticated, are still predominantly individual family farms. The geographic location of pork production is shifting as well. While the traditional Corn Belt represents the overwhelming share of production, growth is also occurring in nontraditional hog states such as Texas, Colorado, and Okla- homa. North Carolina, which ranked fourteenth in pork production 30 years ago, now ranks second. MEAT AND POULTRY INSPECTION The Federal Inspection System Under the Federal Meat Inspection Act and the Poultry Products Inspection Act, USDA, through FSIS, inspects all domestic meat and poultry to be sold in interstate commerce in the United States (FSIS, 2001c). Approximately 6,000 meat and poultry processing plants and 130 import establishments are inspected by FSIS (FSIS, 2002c). Products inspected under FSIS authority include all products from cattle, sheep, swine, goats, horses and other equines, chickens, turkeys, ducks, geese, and guinea fowl (FSIS, 1998a). It also applies to ostriches and emus (FSIS, 2001b). Processed products containing 3 percent or more raw meat and poultry or 2 percent or more cooked meat and poultry are also included (FSIS, 2001c), with some exceptions. Products that do not cross state lines may be inspected by state rather than federal inspection agencies; there are approxi- mately 1,500 meat and poultry establishments that are inspected by state pro-

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 137 grams (GAO, 2001~. Twenty-seven states have established inspection systems equivalent to the federal system; however, products that are state-inspected can only enter intrastate commerce. To ensure the safety of imported meat and poultry products, FSIS maintains a wide-ranging system of inspection and controls. On an annual basis, FSIS evaluates the inspection systems in all foreign countries eligible to export meat and poultry to the United States to ensure that their inspection systems are equiva- lent to the U.S. system (FSIS, 2001c). This evaluation consists of a document review of the country's laws, regulations, and other written information, and an on-site review of plant facilities and equipment, laboratories, and training pro- grams. In addition, all imported meat and poultry products may be reinspected (including testing) upon entering the United States (FSIS, 2003~. The 1997 implementation of the Pathogen Reduction; Hazard Analysis and Critical Control Point Final Rule (PR/HACCP rule) initiated a significant change in the regulatory philosophy and roles of both inspectors and industry. In the past, some plants relied heavily on USDA inspectors to identify plant and process deficiencies before the company would take action to correct them. The PR/ HACCP rule defined the respective roles, tasks, and responsibilities of both industry and FSIS (FSIS, 1996~. Businesses that produce the meat and poultry products are now directly accountable for their safety (FSIS, 1998b). The introduction and implementation of the PR/HACCP rule attempted a significant change in regulatory philosophy and respective roles and responsibili- ties of industry and inspectors over a relatively short time period. The transition has not been entirely smooth; there have been some inconsistencies and setbacks in the start-up process. In response to reports published by the General Account- ing Office, USDA's Office of the Inspector General, and its own self-assessment, FSIS is taking steps to provide supplemental guidance and clarification to assist inspection staff and industry in adapting to these changes (GAO, 2002~. U.S. Department of Agriculture Inspection Models Project Pilot Program USDA began the HACCP-based Inspection Models Project (HIMP) pilot program in 1997 (FSIS, 1997, 2001a). This program was designed to explore extending HACCP and process controls to the slaughter of young animals to further improve food safety and reduce or eliminate product quality defects. A key component of HIMP includes setting performance standards by FSIS and requiring the meat and poultry processors to use process control techniques to meet the performance standards. However, the collection of the data needed to assess the effectiveness of the program has not been completed, so an evaluation of HIMP at this point would be premature. The committee supports the conclusion of previous National Academies reports (NRC, 1985b, NRC, 1987) that carcass-by-carcass inspection is ineffective from a food safety perspective. If successful, HIMP may provide a useful model

38 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD to reduce FSIS dependence on carcass-by-carcass inspection and increase the use of process control techniques to assure the safety of meat and poultry products. State Inspection Programs with Federal Oversight Twenty-seven states operate state meat and poultry inspection programs. These state programs, with federal oversight, were established with the passage of the Wholesome Meat Act of 1967 and the Wholesome Poultry Act of 1968. State meat and poultry inspection programs were required to implement the inspection system mandated by USDA in the PR/HACCP rule beginning in 1997. The transition from traditional meat and poultry inspection to the HACCP system represents a major philosophical, cultural, and procedural change for the state inspection programs. USDA provides matching funds to cover 50 percent of state program costs through the administration of renewable federal grants (WI DATCP, 2002~. State meat and poultry inspection programs are required to meet standards at least equal to the federal program, and FSIS is responsible for determining that they do so. In addition to conducting their own internal audits, state meat and poultry inspection programs are audited by USDA on a one-, two-, or four-year basis, with the frequency based on prior performance. Each state submits a state performance plan as part of an annual report for review by USDA. These plans must describe the operating practices and procedures for administering the state meat and poultry inspection programs, including laws and regulations, funding and financial accountability, resource management, staffing and training, pro- gram operations, facilities and equipment, labels and standards, in-plant review and enforcement, and laboratories (WI DATCP, 2002~. Meat and poultry plants are divided into three size categories. Large plants have 500 or more employees, small plants have 10 to 499 employees; and very small plants have fewer than 10 employees or annual sales of less than $2.5 mil- lion (FSIS, 19961. While plants under federal inspection comprise all three size categories, plants under state meat and poultry inspection programs are currently small and very small plants only (FAIM, 2002~. Consequently, the state inspec- tion programs have developed specialized expertise in working with small and very small plants. In a historical context, it was believed that state inspection programs offered economic benefits such as lower ongoing costs of state inspec- tion compared with federal inspection, greater flexibility in the scheduled time of inspection, and the ability to accommodate low-volume slaughter or processing from local livestock markets (WI DATCP, 2002~. In addition, state programs inspect and monitor custom plants, which are those that slaughter and process meat and poultry products for personal use by the animals' owners (i.e., not for subsequent sale).

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS State and Local Government Inspection of Retail Meat Processors 139 Retailers who process meat and poultry only for direct sales to consumers are subject to different inspection processes and regulations than those whose prod- ucts are sold wholesale. The Food and Drug Administration Model Food Code (FDA, 2001), implemented in 1993 and updated biennially, is a template for the regulation of retail and food service operations. As of April 2002, 49 states had either adopted or were in the process of adopting one of the biennial versions of FDA's Model Food Code. New Mexico is not pursuing adoption of the Food Code, but the state still utilizes it for guidance and interpretation (CFSAN, 2003; FDA, 2001~. The committee recommends that collaboration among USDA, FDA, and state and local governments continue, to help ensure the production of safe meat and poultry products and consumer protection in the United States. Laboratory Analysis Microbiological testing of product samples obtained by the federal and state inspection programs is conducted at USDA-approved laboratories. These are actually lagging indicators in measuring the process performance of meat or poultry plants because samples are taken after the product is prepared and pack- aged, and even with rapid methods, there is a significant lag time between the collection of the sample and the analysis of the laboratory data. By the time these data become available, the corresponding meat and poultry products often have been in the market for varying periods of time and may already have been con- sumed. Therefore, although microbiological samples provide both the plant and regulatory agency with a "score card" for plant performance, if further significant gains in the safety of the U.S. meat and poultry supply are to be realized, meat and poultry establishments need to implement more effective process control measures. As mentioned in Chapter 3, these process control measures should be linked to a systematic continuous improvement process to achieve the level of safety demanded by the U.S. consumer. The Significance of Proper Implementation and Enforcement of the HACCP System It is important to stress that any HACCP system, including one with scien- tifically valid microbiological performance standards, must be properly imple- mented to achieve its intended effect. The Government Accounting Office (GAO) audited HACCP implementation by FSIS (GAO, 2002) and concluded that there were deficiencies in the implementation process. The GAO report identified three major areas of concern. The first relates to establishment of scientifically valid HACCP plans that properly identify hazards

140 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD and appropriate Critical Control Points (CCPs). Some establishments have failed the hazard analysis or have omitted some legitimate hazards in it and have conse- quently not provided for adequate control or interventions of these hazards (e.g., chemical residues or Salmonella'). Validation of a HACCP plan is the responsi- bility of industry personnel. FSIS inspectors are charged with verification of the Sanitation Standard Operating Procedures and HACCP plans, which may include reviewing the plan and the records and corrective actions taken a task that requires training FSIS personnel. To this effect, a recent addition to the FSIS field staff, Consumer Safety Officers, will receive more training on HACCP than the traditional inspection personnel and will be tasked with critical evaluation of HACCP plans as part of HACCP phase-2 implementation, the "Next Steps." This program is being built slowly due to budget constraints. A second area of concern mentioned in the GAO report, which if not cor- rected would make it difficult to implement scientifically valid performance standards, is the issue of corrective action if a plant experiences deviations from its HACCP plan and is deemed to be in noncompliance. Audits of these plants suggest that a majority have repetitive incidences of noncompliance without subsequent corrective action. The third concern identified in the GAO report is that if plants fail the Salmonella performance standard, regulatory action is not necessarily taken. Regulatory action letters may be delayed up to nine months. The report also indicates that, even when conditions occur that could lead to an order for suspension of inspection, orders are often put into abeyance by USDA. As shown by GAO's analysis, complex factors appear to have hampered FSIS' s ability to effectively enforce HACCP implementation in its initial phases. It is not within the charge of this committee to audit the administrative procedures involved in implementation of performance standards, but rather to comment on the scientific criteria involved in establishing them. However, the committee believes that scientific criteria, including performance standards, may be part of a HACCP program and can only be successful in reducing contamination if they are uniformly implemented, and if this implementation is enforced in a timely fashion by the responsible regulatory agency. Promulgation of new standards and establishment of rigid scientific criteria for safe food are useless if monitoring and enforcement are not ensured. To that effect, the responsibility of meat and poultry inspectors should be redefined to reflect their role within a HACCP food safety assurance system. Consistency of the Inspection Process There has been a consolidation of the meat and poultry industries in recent years. Many of the larger meat and poultry companies manage multiple process- ing plants across the United States that are regulated by both FDA and FSIS. This presents challenges to the plants and corporate management due to the inconsis- tent interpretation and enforcement of regulations, which in turn hinders imple-

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 141 mentation of consistent product safety strategies. Anecdotal stories abound in the industry about inconsistencies in the enforcement of rules and regulations be- tween plants and between districts. The committee recommends that FSIS continue its training program and the development of means to measure and evaluate the performance of its inspection team (i.e., Inspectors-in-Charge, Supervisory Veterinary Medical Officers, and inspectors), and state meat and poultry inspection teams, to ensure that regula- tions are consistently enforced across the country. Concurrently, the committee recommends that FDA also continue to develop training programs and various means to measure and evaluate the performance of FDA inspectors and state regulatory agencies that conduct FDA inspections. REVIEW OF CURRENT STANDARDS FOR MEAT AND POULTRY Current Criteria and Performance Standards USDA specifically charged this committee to develop definitions for terms such as "performance criteria" and "performance standard." The definitions of these and other relevant terms are presented in tabular form in Appendix A. The definitions adopted by the committee that are of particular relevance to the remaining sections of this chapter are those of performance standard and micro- biological criterion. Within the last decade, FSIS has established several criteria, including per- formance standards, as part of the current regulatory and inspection system for meat and poultry. These include criteria for process control and standards for pathogen reduction in raw products, adulteration, standards for cooked products, and general sanitation standards. Among these, criteria for process control and standards for pathogen reduction in raw products involve microbiological sam- pling and testing programs. The results of these testing programs are used by the agency to determine whether processors receive a "fail" or "pass." In contrast to these microbiological standards and criteria, which apply to a broad range of products, "adulteration" is very narrowly interpreted for a specific bacterium and product, Escherichia cold 0157:H7 in raw ground beef. Standards for cooked products differ from the standards for raw meat and poultry in that they require the reduction of a stated number of a specific pathogen, as well as validation of the process used to achieve that reduction, instead of a testing and sampling program. Sanitation standards (as they are specifically referred to in the Code of Federal Regulations) are less prescriptive and contain vague descriptors such as "adequate" and "sufficient." Consequently, these standards are subject to more interpretation than either the cooking process or microbiological criteria or stan- dards. Several types of standards or criteria are summarized and discussed in the following sections.

142 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD Contamination with Microorganisms; Process Control Verification Criteria and Testing; Pathogen Reduction Standards for Red Meats (9 Vol 2 C.F.R. §310.25) These criteria are part of the PRIHACCP rule and include both process control criteria for E. cold Biotype I (generic E. coli) and performance standards for a specific pathogen (salmonellae). The process control criteria are based on the quantitative level of generic E. cold on or in fresh meats. The sampling technique includes a swab or excision method for intact carcasses and a destruc- tive analysis for ground products. The sampling frequency varies both by species and by the relative size of the processing establishment (Table 4.1~. The sampling and testing protocol for the process control criteria are based on a three-class sampling program. In a three-class plan, m is the analytical value that differentiates good quality from marginally acceptable quality, M is defined as the analytical value that differentiates marginally acceptable quality from unacceptable quality, n is the number of samples taken, and c is the maximum number of samples out of n that may exceed the value set for m. For a sample set to pass, no sample may exceed the M value and no more than c samples may exceed the m value. The values for the various species are given in Table 4.2. TABLE 4.1 Sampling Frequency for Process Control Indicator (Generic Escherichia coli) for Fresh Meat Species or Size of Establishment Samples per Number of Carcasses Cattle, sheep, or horses Swine Very low-volume establishments 1 per 300 1 per 1,000 At least 1 per week, beginning June 1 of each year, until 13 in-compliance samples are collected in a row SOURCE: 9 C.F.R. §310.25. TABLE 4.2 Values for m, M, n, and c for the Process Control Indicator (Generic Escherichia coli) for Fresh Meata Species m M n c Cattle Negativeb 100 13 3 Swine 10 10,000 13 3 a m = the analytical value that differentiates good quality from marginally acceptable quality, M= the analytical value that differentiates marginally acceptable quality from unacceptable quality, n = the number of samples taken, c = is the maximum number of samples out of n that may exceed the value set for m. b Negative is defined by the sensitivity of the method used in the baseline study, with a limit of sensitivity of at least 5 cfu/cm2 carcass surface area. SOURCE: 9 C.F.R. §310.25.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS TABLE 4.3 Values for n and c for the Pathogen Reduction Standard (Salmonella Performance Standard) for Fresh Meat 143 Performance Standard Product (% positive for salmonellae) Maximum Number of Number of Positives to Achieve Samples Tested (n) Standard (c) Steers/heifers 1.0 82 1 Cows/bulls 2.7 58 2 Ground beef 7.5 53 5 Hogs 8.7 55 6 Fresh pork sausage NAa NA NA a NA = not applicable. SOURCE: 9 C.F.R. §310.25. The sampling frequency for the pathogen reduction standard for Salmonella is identical to that for the process control indicator (Table 4.1~. The sampling technique includes a swab or excision method for intact carcasses and a destruc- tive analysis for ground products. In practice, FSIS will take an initial sample set (the A set). If an establishment fails the A set, FSIS will take up to two more sample sets (the B and C sets). Failure of all three sample sets would be grounds for USDA to withdraw inspection from an establishment. The pathogen reduction standard is based on a two-class sampling plan, in which n is the number of samples taken and c is the number of samples allowed to fail the specification. The standard is based on a qualitative assay for the presence or absence of Salmonella. The values for the various species and prod- ucts are given in Table 4.3. Contamination with Microorganisms; Process Control Verification Criteria and Testing; Pathogen Reduction Standards in Raw Poultry (9 Vol 2 C.F.R. §381.94) The process control criteria and the pathogen reduction standard for raw poultry are structured in an identical manner to those for red meats. The process control criteria are based on the numerical populations of E. cold Biotype I (generic E. coli) on or in fresh poultry meats. The sampling technique includes a whole- bird rinse for intact carcasses and a destructive analysis for ground product. The sampling frequency varies both by species and by the relative size of the process- ing establishment (Table 4.4~. The sampling and testing protocols are based on a three-class sampling program, as previously described. The values for the various species are given in Table 4.5. For the pathogen reduction standard for Salmonella, the sampling frequency is identical to that for the process control indicator (Table 4.4~. The sampling

44 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD TABLE 4.4 Sampling Frequency for Process Control Indicator (Generic Escherichia coli) for Raw Poultry Species or Size of Establishment Samples per Number of Carcasses Chicken Turkeys Very low-volume establishments 1 per 22,000 1 per 3,000 At least 1 per week, beginning June 1 of each year, until 13 in-compliance samples are collected in a row SOURCE: 9 C.F.R. §310.25. TABLE 4.5 Values for m, M, n, and c for the Process Control Indicator (Generic Escherichia coli) for Raw Poultrya Species m M n c Chicken 100 1,000 13 3 Turkey NAb NA NA NA a m = the analytical value that differentiates good quality from marginally acceptable quality, M = the analytical value that differentiates marginally acceptable quality from unacceptable quality, n = the number of samples taken, c = is the maximum number of samples out of n that may exceed the value set for m. b NA = not applicable. SOURCE: 9 C.F.R. §310.25. technique includes a whole-bird rinse for intact carcasses and a destructive analysis for ground products. In practice, FSIS will take an initial sample set (the A set), and if an establishment fails the A set, FSIS will take up to two more sample sets (the B and C sets). Until recently, failure of all three sample sets would be grounds for USDA to withdraw inspection from an establishment. The pathogen reduction standard is based on a two-class sampling plan, where n is the number of samples taken and c is the number of samples allowed to fail the specification. The standard is based on a qualitative assay for the presence or absence of salmonellae. The values for the various species and prod- ucts are given in Table 4.6. Adulteration of Ground Beef; E. cold 0157:H7 (9 C.F.R. §417) USDA believes that E. cold 0157:H7 is an adulterant in raw ground beef based on its interpretation of the following section of the Federal Meat Inspection Act:

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS TABLE 4.6 Values for n and c for the Pathogen Reduction Standard (Salmonella Performance Standard) for Raw Poultry 145 Performance Standard Maximum Number of (% positive for Number of Positives to Achieve Product salmonellae) Samples Tested (n) Standard (c) Broilers 20.0 51 12 Ground chicken 44.6 53 26 Ground turkey 49.9 53 29 Turkeys NAa NA NA a NA = not applicable. SOURCE: 9 C.F.R. §310.25. (m) The term "adulterated" shall apply to any carcass, part thereof, meat or meat food product under one or more of the following circumstances: (1) if it bears or contains any poisonous or deleterious substance which may render it injurious to health; but in case the substance is not an added substance, such article shall not be considered adulterated under this clause if the quantity of such substance in or on such article does not ordinarily render it injurious to health. (4) if it has been prepared, packed, or held under unsanitary conditions whereby it may have become contaminated with filth, or whereby it may have been rendered injurious to health. (21 U.S.C. §601 (m)(l) and (41) USDA interpreted these statements to mean that the detectable presence of E. cold 0157:H7 in raw ground beef product, irrespective of the method used to detect it, would meet either of the circumstances above and, therefore, such product would be considered adulterated. Requirements for the Production of Cooked Beef Roast Beef and Cooked Corned Beef Products (9 Vol 2 C.F.R. §318.17) The previous regulations for the production of cooked meat were modified so that they are now included as performance standards within the specific HACCP plans. Using HACCP terminology, the cooking step would be a CCP and the specific requirements would be the critical limits. The cooked red meat regulation includes two performance standards specifying (1) that the cooking process achieves a certain lethality for salmonellae, and (2) a specific rate of chilling (i.e., stabilization) for control of Clostridium perfringens. These require- ments differ from the microbiological sampling programs required for raw meat and poultry in that the processor must show that its process is validated and, therefore, that it achieves the stated standard.

46 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD The standard for lethality, which must include a cooking step, specifies a 6.5-log reduction of Salmonella or an alternative lethality that achieves an equivalent probability that no viable Salmonella remain in the finished product. The standard for stabilization requires no multiplication of toxigenic micro- organisms, such as C. botulinum, and no more than a 1-log multiplication of C. perfringens. As an alternative, USDA has provided "safe harbor" processes for both lethality and stabilization standards that relieve the processor from having to validate the process. Briefly, a safe harbor process is one that has been estab- lished as accomplishing the objective. The safe harbor processes are compiled in FSIS Directives 7370.2 (FSIS, 1995) and 7110.3 (FSIS, 1989~. Requirements for the Production of Fully Cooked Poultry Products and Partially Cooked Poultry Breakfast Strips (9 Vol 2 C.F.R. §381.150) The cooked poultry meat regulations contain process control requirements similar to the standards for red meats, and these requirements also need to be included in a plant's HACCP plan. The cooking step would be a CCP and the specific requirements would be the critical limits. The standard for lethality is a 7-log reduction of salmonellae or an alternative lethality that achieves an equiva- lent probability that no viable salmonellae remain in the finished product (it must include a cooking step). For stabilization, there can be no multiplication of toxi- genic microorganisms such as C. botulinum and no more than a 1-log growth of C. perfringens. The safe harbor processes are compiled in FSIS Directives 7370.2 (FSIS, 1995) and 7110.3 (FSIS, 1989~. Animal Drug Residues The Center for Veterinary Medicine (CVM) of FDA is primarily responsible for establishing tolerances and action levels for antibiotics and hormones in the edible tissues of food-producing animals. The setting of such tolerances, and their surveillance by FSIS, was discussed earlier in the chemical risk assessment section of Chapter 3. A complete review of this area can also be found in the report The Use of Drugs in Food Animals: Benefits and Risks (NRC, 1999~. There are numerous types of drugs used in food animals. It is generally accepted in the United States that anabolic steroid hormones used to promote weight gain and feed efficiency enjoy a wide safety margin for human health when used at approved rates (21 Vol 6 C.F.R., parts 522, 556, and 558~. Anti- biotics may be used either to promote growth and feed efficiency (subtherapeutic use) or to treat actual disease (therapeutic use); the latter involves a veterinarian in the diagnosis and management of the disease. Compounds are either available over the counter or only by order of a licensed veterinarian. Veterinarians can

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 147 prescribe drugs and dosages that are not specifically approved if a medical need arises. In food-producing animals, the veterinarian must also ensure that a sub- stantially extended withdrawal time is allowed to eliminate residues from the edible tissue. Some chemicals are specifically prohibited from off-label use in food- producing animals (e.g., higher dose or for indication or species not on the ap- proved label) (CVM, 2002a). These currently include chloramphenicol, clenbuterol, diethylstilbestrol, dimetridazole, ipronidazole, other nitroimidazoles, furazoli- done, nitrofurazone, sulfonamide drugs in lactating dairy cattle (except approved use of sulfadimethoxine, sulfabromomethazine, and sulfaethoxypyridazine), fluoroquinolones, and glycopeptides (21 Vol 1 C.F.R. §530.41~. The use of drugs in food animals continues to undergo regulatory review. CVM recently promulgated a revised definition of the term "no residue" when it appears in new animal drug regulations to mean that no residue is detected using an approved regulatory method (21 Vol 1 C.F.R. §500.84~. This term normally occurs in regulations where a drug is purported to be a human carcinogen, which is a toxic class that is regulated differently than other compounds. Also, CVM has issued a draft guidance to evaluate, through use of qualitative risk analysis methods, the safety of new antimicrobial animal drugs with regard to the possibil- ity of eliciting development of resistance by bacteria that are of concern to human health (CVM, 2002~. The tolerance that has already been set for some of these chemicals could be used as a performance standard. Sanitation (9 Vol 2 C.F.R. §416) The sanitation performance standards were changed from multiple, detailed, prescriptive regulations to standards. The regulations contain specific sections on grounds and facilities; equipment and utensils; sanitary operations; employee hygiene; and tagging of unsanitary equipment, utensils, rooms, or compartments. Although described as standards, the actual language includes numerous refer- ences to "adequate," "preventing sources of adulteration," and "sufficient." These regulations provide little in the way of a descriptive and objective standard and are better characterized as "guides." For example, the language of these regula- tions is sufficiently different from that of the regulations described previously as to question whether they are true standards, as defined by this committee in Appendix A. Appendix B summarizes the details of the sanitation performance standards. Using a Science-Based Approach to Develop Performance Standards and Other Scientific Criteria As described in Chapter 3, a science-based approach to developing criteria, including performance standards, entails gathering, analyzing, and utilizing the best available data. The strategy of combining a controlled study and expertise

48 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD accepts the fact that gaps in the data will always exist and that these data gaps need to be supplemented with the qualitative knowledge of (and assumptions developed by) experts in the particular subject matter. Pilot studies are the pre- ferred method of gathering data because they can be designed with the specific objective of developing performance standards. The qualitative data and assump- tions are critical issues that can affect the quality of the performance standard; transparency in describing the assumptions made becomes a critical component in the development of a standard. For example, the lethality and stabilization standard document for meat and poultry products describes the method USDA prescribed to achieve the 7-D reduction of Salmonella in ready-to-eat (RTE) poultry products and the 6.5-D reduction of Salmonella in RTE beef products (FSIS, 1998c). In this document, which also describes the scientific basis for the stabilization performance stan- dard, the validity of the data used and the assumptions made are not clear from either a mathematical or microbiological perspective. In addition, the microbio- logical and technological assumptions may not reflect actual manufacturing conditions. For example, the baseline data used were the FSIS Nationwide Micro- biological Baseline Data Collection Programs and the Nationwide Federal Plant Microbiological Surveys, published between 1994 and 1996. Because these data were gathered prior to the implementation of the PR/HACCP rule, they do not reflect improvements that were made as a result of the implementation of the rule. The authors of the performance standard assumed that the rule would not reduce the incidence of Salmonella in RTE products. Regulatory agencies need to properly set performance standards. This is a balancing act between setting a highly conservative performance standard and setting an excessively tolerant one. Although the safety margin approach is valid and useful, developing a standard that uses a safety margin based on a highly conservative worst-case scenario may lead to production of overprocessed prod- ucts of inferior quality and may place an undue economic burden on the processor, without significantly increasing product safety. Setting performance standards that are too tolerant, on the other hand, may lead to production of unsafe products. As discussed in Chapter 3, the committee stresses the importance of and recommends an increase in transparency during the development of performance standards. This transparency must include making public within limits of the Freedom of Information Act and taking into consideration confidentiality and trade secrets any analytical data used, the method used to analyze the data, and the assumptions that are made to fill in any data or technical gaps. Increasing the transparency of the process to set performance standards provides an opportunity for informed comments and input from the affected public to the regulatory agencies. This transparency is needed to increase the quality of performance standards and to provide appropriate information for conducting better reviews of the standards, either by external agencies such as GAO or by internal teams; to update the performance standard; and to meet future public health objectives.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 149 The committee also stresses the need to use proper assumptions in develop- ing performance standards. When regulatory agencies set performance standards, they need to balance a number of factors, including public health objectives, economic burden, available technologies, and the effect of the interventions on product quality. The specific standards and the basis and rationale for their implementation are discussed in subsequent sections. The Scientific Basis of Current Criteria and Performance Standards USDA discussed the rationale for the introduction and use of process control criteria and pathogen reduction standards for fresh meats in the PR/HACCP rule (FSIS, 1996~. The following sections include portions of the rule. They also include the committee' s summary and analysis of the scientific basis and rationale for each standard, as argued in the rule. Based on the analyses, the committee presents recommendations for improvements. Process Control Criteria; Generic E. cold in or on Fresh Meats (9 Vol 2 C.F.R. §310.25) In slaughter establishments, fecal contamination of carcasses is the primary avenue for contamination by pathogens. Pathogens may reside in fecal material and ingesta, both within the gastrointestinal tract and on the exterior surfaces of animals going to slaughter. Therefore, without care being taken in handling and dressing procedures during slaughter and processing, the edible portions of the carcass can become contaminated with bacteria capable of causing illness in humans. Additionally, once introduced into the establishment environment, the organisms may be spread from carcass to carcass. Because the microbial pathogens associated with fecal contamination are the single most likely source of potential food safety hazard in slaughter establish- ments, preventing and removing fecal contamination and associated bacteria are vital responsibilities of slaughter establishments. Further, because such con- tamination is largely preventable, controls to address it will be a critical part of any slaughter establishment's HACCP plan. Most slaughter establishments already have in place procedures designed to prevent and remove visible fecal contamination. There is general agreement within the scientific community that generic E. cold is the best single microbial indicator for fecal contamination. FSIS, there- fore, is requiring that establishments slaughtering livestock or poultry begin testing for E. cold (~E. coli, biotype I, nonspecific as to species, hereinafter referred to simply as E. coli) at the frequency and following the procedures described in 'Process Control Verification; E. cold Performance Criteria and Testing' section, . . ., 6 months after publication of the final rule FSIS considers the required testing to be essential for meeting current statutory requirements

150 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD for sanitation and the prevention of adulteration. This testing also will play an integral role in the successful implementation of HACCP in slaughter establish- ments. In addition, FSIS is establishing process control performance criteria for fecal contamination based on the frequency and levels of contamination of carcasses with E. coli. (FSIS, 1996, Pp. 38837-38838) FSIS is also establishing performance criteria based on national microbio- logical baseline surveys. The criteria are not regulatory standards but rather provide a benchmark for use by slaughter establishments in evaluating E. cold test results. Test results that do not meet the performance criteria will be an indication that the slaughter establishment may not be maintaining adequate process control for fecal contamination and associated bacteria. Such results will be used in conjunction with other information to evaluate and make appro- priate adjustments to ensure adequate process control for fecal contamination and associated bacteria. (FSIS, 1996, P. 38811) FSIS believes that testing for generic E. cold is the appropriate and necessary means by which meat and poultry slaughter establishments must verify their process controls. (FSIS, 1996, Pp. 38838-38839) According to a report by the National Research Council (NRC, 1985b), there are other bacteria or groups of bacteria (fecal streptococci, for example) that may serve equally well as indicators of fecal contamination as generic E. coli. How- ever, that report also stated that limits for indicator organisms were impractical because "there is no direct relationship between the presence of these types [indicator organisms] and the presence or absence of pathogens." Although argu- able, there is in fact general agreement within the scientific community that generic E. cold is perhaps the best indicator of fecal contamination. In spite of this controversy, the FSIS rationale makes reasonable assumptions and proceeds in a logical fashion. The baseline data used to develop the performance standard were collected from 1992 to 1997 as part of the FSIS Nationwide Microbiological Baseline Data Collection Programs and the Nationwide Federal Plant Microbio- logical Surveys. These programs were intended to give a general microbiological profile of a product for the selected microorganisms as a reference for further investigations and evaluations of new programs. The use of a three-class sampling protocol is appropriate for the intended purpose. The values of m, M, n, and c were established based on the national baseline data for each species and were set at levels that would allow approximately 80 percent of the establishments to pass the criteria. Because the generic E. cold limits are a guideline, industry is not obligated to have a sampling and testing program in place. Although the data collected by the industry are not within the public domain and therefore not available for review, the criteria for generic E. cold (Biotypes I and II) have been implemented in essentially all federally and state inspected establishments. The criteria have been used to detect problems and document acceptable control of the process, and anecdotal reports indicate that the criteria have served to document a reduction in

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 151 the levels of carcass contamination and have led to process improvement. An additional benefit of the generic E. cold criteria has been an increased awareness in the meat and poultry industry of the importance and significance of process control on the microbiological status of carcasses. The concept of continuous improvement is central to food safety. In principle, if populations of generic E. cold are extremely low, the sampling results from carcass data may not provide sufficient information to enable the processor to detect remaining problems and further improve operations. In situations where the populations of generic E. cold are too low to provide valuable information to the processor, the committee recommends that a reevaluation of the criteria be conducted, to identify either an alternate system of testing (i.e., sampling a larger area) or another indicator of carcass hygiene. Because the E. cold data collected by industry are not in the public domain, it is currently not possible to determine whether this is in fact a significant limitation to continuous process improvement. The committee recommends that an anonymous national database be created to collect the available generic E. cold data on carcasses so that industry and regulatory and public health agencies have benchmarks available for comparative purposes. The committee further recommends that this database be operated by a nonregulatory government agency or under contract to a university or nonprofit organization. This would allay industry concerns about potential use of such industry-generated data for regulation enforcement purposes. In addition, the committee recommends the implementation of criteria for generic E. cold in ground beef. These criteria should be developed using the generic E. cold criteria for carcasses as the model. The data from these criteria should be handled in the same manner as recommended for the E. cold criteria for carcasses (i.e., a national, anonymous database). FSIS is purposely using the term performance "criteria" rather than perfor- mance "standard" in this context because no single set of test results can demon- strate conclusively that adequate process control for fecal contamination is or is not being maintained. As explained below, if test results do not meet the appli- cable criterion, it raises questions about the adequacy of the process control. FSIS intends to consider the establishment's results and corrective actions, together with other information and inspectional observations, in evaluating whether a problem exists that requires regulatory action or other measures to protect consumers and ensure compliance with the law. (FSIS, 1996, P. 38838) FSIS has established that a "criterion" is similar to a "microbiological guide- line," as defined by the International Commission on the Microbiological Safety of Foods (ICMSF). That is, a microbiological guideline is a criterion to monitor a food process or system (ICMSF, 2002~. These criteria are usually considered advisory, but may be mandatory. In the case of criteria for process control, the

152 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD recommended levels are advisory, although FSIS clearly expects action to be taken if there is routine failure of the criteria. Pathogen Reduction Standard; Salmonella Performance Standard (9 Vol 2 C.F.R. §310.25) FSIS is also establishing pathogen reduction performance standards for Salmonella that will require all slaughter establishments to reduce the incidence of Salmonella contamination of finished meat and poultry carcasses below the national baseline prevalence as established by the most recent FSIS national microbiological baseline data for each major species. FSIS will conduct Salmonella testing in slaughter establishments to detect whether they are meet- ing the pathogen reduction performance standards, and will require corrective action or take regulatory action, as appropriate, to ensure establishments are meeting the pathogen reduction standards. Pathogen-specific performance standards for raw products are an essential component of the FSIS food safety strategy because they provide a direct measure of progress in controlling and reducing the most significant hazards associated with raw meat and poultry products. The Salmonella standards being established in this final rule, which are based on the current national baseline prevalence of Salmonella (expressed as a percentage of contaminated carcasses), are a first step in what FSIS expects to be a broader reliance in the future on pathogen-specific performance standards. FSIS plans to repeat its baseline sur- veys and collect substantial additional data through other means and, on that basis, adjust the Salmonella performance standards and possibly set standards for additional pathogens, as appropriate. Also, FSIS will continue to explore establishing pathogen-specific performance standards based on the levels of contamination (i.e., the number of organisms) on a carcass. Future FSIS efforts on such performance standards will reflect the fact that achieving the food safety goal of reducing foodborne illness to the maximum extent possible will require continuous efforts and improvement over a substantial period. (FSIS, 1996, Pp. 3881 1-38812) The stated purpose of the Salmonella performance standard is to promote a reduction in the levels of Salmonella on raw meat, hence the name Pathogen Reduction Standard. The NRC report (NRC, 1985a) stated that limits for patho- genic microorganisms in microbiological criteria for raw meats are impractical. However, since data from the USDA verification program show that the goal was achieved a reduction of the incidence of salmonellae in or on meat the com- mittee concludes that these standards are valid. In some instances, however, if the populations or incidences of salmonellae are extremely low, especially on car- casses of some production animal species, the testing may no longer be providing the information needed by the processor to continue making improvements in the process. Other testing approaches may need to be considered in such cases.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 153 Because of the importance of the baseline data, the committee recommends that a new baseline survey be conducted on a periodic basis to evaluate the microbiological status of carcass, trim, ground product, and RTE products, both at the site of production and at the retail level. This survey should evaluate the same microorganisms that were evaluated in the previous baseline surveys unless evidence for newly established pathogens is presented. The sampling design for the survey should be weighted based on the production of the establishment and account for geographical location and seasonality. Also, it is important that data for this new baseline be collected in such a way as to address two competing concerns. First, it should be possible to compare the results of the new baseline to the old baseline to determine if the situation is improving, worsening, or staying the same. Second, the new baseline should be representative and statistically valid and should correct deficiencies in the sampling plan used for the 1992 to 1997 baseline. The survey should ideally be coordinated with other baseline data collection projects, such as the Animal and Plant Health Inspection Service's National Animal Health Monitoring Survey (NAHMS). The baseline data used to develop the Salmonella performance standard were collected in the same manner as that for the E. cold process control criteria. The use of a two-class sampling protocol is appropriate for the intended purpose. The values of n and c were established based on the national baseline data for each species, and set at levels that would allow approximately 80 percent of establish- ments to pass, based on the baseline data. USDA's implementation of the PR/HACCP rule in meat and poultry plants is one of several recent control measures credited with decreasing the overall incidence of foodborne illness in the United States from 1996 to 2002 (HHS, 2002~. Data obtained from the Foodborne Diseases Active Surveillance Network (FoodNet) reveal an overall decline of 23 percent in bacterial foodborne illnesses during this 6-year period (CDC, 2002~. Since the introduction of PR/HACCP, declines in the rate of Salmonella infections in the U.S. population have coincided with declines in the prevalence of Salmonella detected in FSIS-regulated products (CDC, 2002; USDA, 2002~. Rose and colleagues (2002) reported on the preva- lence of Salmonella in raw meat and poultry, assessed on the basis of the propor- tion of inspected meat-production facilities passing the Salmonella performance standard in 1998, 1999, and 2000, compared with the defining pre-HACCP baseline prevalence data. This study consisted of 98,204 samples and 1,502 com- pleted sample sets collected from large, small, and very small processing plants that produced one of the following: broilers, market hogs, cows, bulls, steers and heifers, or ground beef, chicken, or turkey. The overall conclusion was that greater than 80 percent of the sample sets met the Salmonella performance standards of 20.0 percent for broilers, 8.7 percent for market hogs, 2.7 percent for cows and bulls, 1.0 percent for steers and heifers, 7.5 percent for ground beef, 44.6 percent for ground chicken, and 49.9 percent for ground turkey. The percentage of samples positive for Salmonella was generally lower than in the

154 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD pre-HACCP baseline data. Data were also collected on second and third visits to plants that did not meet the performance standards on the first visit. Of the 98,206 samples collected, 6,260 were from second visits and 752 were from third visits. These results are encouraging despite some significant limitations in the data sets collected relevant to balance of the samples based on establishment size. In addition, the post-HACCP data were not designed to serve as a prevalence survey, but for verification and compliance purposes; thus, direct comparison to the pre- HACCP baseline survey is problematic. For this reason, it may not be statistically valid to compare the two data sets; however, because of the vast number of data sets collected, a decrease in Salmonella-positive samples can be clearly observed since the implementation of the Salmonella standard. The committee points out, however, that correlation and causation are two separate and distinct concepts, and while correlated, it may not be scientifically defensible to assume a cause-and-effect relationship between the PR/HACCP rule and the observed decline in the incidence of salmonellosis. The committee, recognizing the importance of measuring the public health impact of pathogen reduction performance standards, addressed this issue in Chapter 2 and recom- mended expanded foodborne disease surveillance and microbial testing of foods, linked to a comparison of microbial serotypes in isolates from animals, humans, and foods, as a means to enable regulatory and public health agencies to allocate the burden of foodborne disease to specific foods or classes of food. A number of changes have occurred coincident with HACCP implementa- tion. The positive side of this survey (Rose et al., 2002) is that Salmonella meat contamination levels were generally reduced, a finding consistent with improve- ment through HACCP implementation. As discussed elsewhere in this report, this does not mean that raw meat products are free from Salmonella, only that the performance standards based on pre-HACCP baseline prevalence targets have been met. These targets are very different across meat classes. For example, the performance standard for steers and heifers showed only one positive sample out of a sample set of 82. For ground chicken, there were 26 positive samples out of a sample set of 53. The goal of the Salmonella performance standards was to reduce the prevalence of Salmonella in raw meat and poultry products. The committee recognizes that this goal is apparently being achieved. Despite the statistical validity and possible contribution to improving public health, the Salmonella performance standards have been highly debated, espe- cially for ground products. The stated regulatory purpose of the Salmonella per- formance standard for ground products is to provide an evaluation of the HACCP plan of grinding operations. On October 7, 2002, FSIS issued a Federal Register notice informing establishments that produce raw beef product, especially intact and nonintact products in the process categories of raw product ground, raw not ground, and slaughter, about the need to reassess HACCP plans for E. cold 0157:H7 (FSIS, 2002a). Until October 2002, the raw material used to manufac- ture ground beef (boneless beef trim) may have passed all parts of the inspection

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 155 system, and may have been processed under a valid HACCP system, and yet still contain E. cold 0157:H7. Since there is no required testing of the trim itself, the possible presence of the bacterium is not detected until testing of the final product is conducted. As currently practiced, the testing performed by FSIS often does not result in detection of the bacterium until after the ground beef has been distributed, and is often already in the hands of the consumer. Until the October 2002 Federal Register notice, the regulatory burden fell solely on the producer of the ground beef, even though the actual source of the bacterium may not be within the grinding operation, but at the production of the trim. The beef grinding operations do have a responsibility to regulate the quality of the incoming raw materials, but the producers of that raw material also have a responsibility to take active measures to reduce contamination of the trim. This point has been addressed by several large companies in that they now provide purchase specification letters to their customers describing their intervention procedures on carcasses and test- ing for E. cold 0157:H7 on trimmings and in ground beef (Shire, 2003~. As a consequence of the weaknesses of the Salmonella performance standard for ground beef, enforcement of this standard has been particularly problematic. With ground beef, the pathogen may be an indication of cross-contamination; however, unless testing of the numerous sources of trimmings is performed, the standard alone cannot be appropriately used to judge the sanitary conditions of the grinding plant. The question of who is responsible for the regulatory failure when a grinding plant fails to meet the standard has not been resolved. In addition, and although the regulations state that failing the Salmonella performance standard may result in withdrawal of federal inspectors, recent liti- gation has raised questions about USDA's statutory authority for such an action. The statutory framework for government enforcement of performance standards created to assure food safety has proven to be inflexible. In Supreme Beef Proces- sors v. USDA, 275 F. 3d 432 (5th Cir. 2001), the United States Court of Appeals decided that USDA's Salmonella performance standard improperly regulated the Salmonella levels of meat entering Supreme Beef' s grinding plant and that cross- contamination of ground beef with Salmonella could not be considered an unsanitary condition rendering the product "injurious to health." Thus, in the absence of finding unsanitary conditions at the establishment, USDA could not withdraw inspection from a grinding plant that had failed the Salmonella perfor- mance standard. The Court's reading of 21 U.S.C. §601(m)~4) was that "it cannot be used to regulate characteristics of the raw materials that exist before the meat product is 'prepared, packed or held'." That is, the USDA Salmonella performance stan- dard, as applied to grinding plants, is invalid "because it regulates the procure- ment of raw materials," not the sanitary conditions of the grinding plant. Also, because ground beef can be cooked to control Salmonella and therefore may not be injurious to health, the Court decided that Salmonella is not itself considered an "adulterant" subject to the prohibition of 21 U.S.C. §601(m)~1~. In addition,

156 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD USDA's claim that the Salmonella performance standard is a proxy for the pres- ence or absence of pathogen controls was dismissed by the court, which found USDA's motivation for the performance standard to be regulation of Salmonella itself. The Supreme Beef case clearly illustrates how the legal environment in which food safety regulatory bodies operate is in conflict with the implementa- tion of current performance standards. In a more recent, high-profile case, USDA entered into a settlement with Nebraska Beef Ltd. that did not result in withdrawal of federal inspection, after issuing numerous citations against the firm for unsanitary conditions linked to the discovery of hamburger contaminated with E. cold 0157:H7. While this case was not based upon failure of the Salmonella performance standard, it sparked con- siderable discussion and concern, including an editorial in the New York Times (Becker, 2003), about whether USDA had adequate authority to protect the public health. Whether Salmonella is an adulterant under existing statutes should not be the issue. The law currently forbids the holding or processing of foods under unsani- tary conditions. The law should also ensure that foods that pose an unacceptable risk to consumers (because of either unusually high levels of pathogens or a high incidence of pathogens) are not marketed. The committee, recognizing all of the above, recommends that a Salmonella performance standard or other appropriate indicator be developed for beef trim intended for grinding (see Figure 4.2~. Such a standard could be defined as either the presence/absence of the indicator or a quantitative measurement whenever possible. In addition, the committee recom- mends that the Salmonella performance standard for ground beef be reevaluated after appropriate interventions and the trim performance standard are in place. Further research should be conducted to determine an appropriate performance standard for ground beef at the grinding operation. Furthermore, the committee recommends that all meat intended for trim for ground products, especially ground beef, be exposed to some form of verified intervention. This also applies to meat derived from heads, which currently may not be subjected to any intervention. Adulteration of Ground Beef; Escherichia cold 0157:H7 (21 U.S.C. §§601, 608, 621) FSIS interpreted the statements in the above sections of the U.S. Code to mean that the detectable presence of E. cold 0157:H7 in ground beef, irrespective of the method used, would meet either of the circumstances that would qualify this pathogen as an adulterant in ground beef. The rationale for this interpretation appears to be that ground beef contains meat from multiple carcasses, and that grinding incorporates the bacteria throughout the meat. In contrast, intact muscle cuts originate from a single carcass, and therefore, any microbial contamination that is present is only on the external surfaces of the meat. The significance of this

157 C' In C' 1 , Q IL S U) .e . _ ~ i_ ._ O ~ In ~ 0 In ~ ~ ~ ~ _ ~ ~ In Cal. m ~ 4, .8 on %0 ~ to In O ECU. O 0 cow q) ~ .~ C ~ COO Cal Cal . . . . . ~ .= Cal .= Cal . ~ Cal Cal . . o . ~ r _ o Cal o o ._. o Cal o o Sit x ~ ~ o on

158 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD is that with intact muscle cuts, cooking will destroy the bacteria on the surface- and therefore any E. cold 0157:H7 present even if the internal temperature of the cut does not reach a temperature sufficiently high to destroy this pathogen. In contrast, ground product does contain bacteria throughout the meat, and if the internal temperature does not reach a temperature sufficiently high to destroy E. cold 0157:H7, a health hazard may exist (FSIS, 1999a). This interpretation results in a situation where beef trim, if contaminated with E. cold 0157:H7, is still considered acceptable under FSIS regulations, but is considered adulterated if that trim is ground. In the United States, it is common to blend beef carcass trim from a variety of domestic and foreign sources to achieve a specific ratio of lean muscle tissue to fat, and then grind the blended trim to produce ground beef. Many independent establishments produce ground beef for both the retail and foodservice markets, and to do this they buy beef trim from various suppliers. For these independent establishments, the burden of enforce- ment falls entirely upon them. That is, an independent grinding establishment may buy beef trim that is inspected and passed by FSIS, but may be classified as adulterated after grinding if it is contaminated with E. cold 0157:H7. The grind- ing process in and of itself may not introduce the bacterium into the product; however, if the bacterium is present, it is redistributed throughout the ground meat. Because of the low infectious dose attributed to E. cold 0157:H7 and the potential severity of the disease it causes, the presence of this pathogen in foods is a serious human health hazard. However, even though E. cold 0157:H7 has been declared an adulterant in ground beef (i.e., there is a zero tolerance policy), the regulation has been insufficient to reduce the rate of human illness attribut- able to this microorganism. Thus, the corresponding human health data have shown no significant change in disease rates since 1996 (CDC,2002~. (A reported increase in the incidence of the pathogen in ground beef since 1999, as indicated by FSIS testing, is most likely the result of a change to a more sensitive analytical methodology in 1998.) It is difficult to rely on zero tolerance to achieve significant public health improvements. While it is impossible to guarantee the absence of E. cold 0157:H7 or any pathogen in food through a zero tolerance policy, the evidence indicates that either cooking to at least 160°F or irradiating to a high enough dose are reliable means of reducing the levels of E. cold 0157:H7. Irradiation, however, does not replace the need for proper cooking. The advice to cook hamburger to the recommended internal temperature of 160°F often goes unheeded by those who prepare it. Considerably more educa- tion of the public and particularly of food service managers and workers is needed. Ground beef products should bear clear and concise labels warning of the potential for harm if the product is not properly cooked.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 159 Irradiation occurs as part of the process, before distribution, so that the meat that reaches the consumer has a reduced risk of contamination. When the con- tamination is reduced before distribution, the potential for cross-contamination is significantly reduced at the level of preparation and consumption. Irradiation is applied to meats that have already passed all existing federal regulatory require- ments and is used as an additional intervention to assure the microbiological safety of the meat. The committee believes that when used, irradiation must be incorporated into the overall HACCP system; it must not be used as a substitute for existing CCPs and other interventions. Microbial contaminants have to be prevented from entering the food supply or eliminated by applying an effective intervention measure to the food. Within the HACCP concept, if there are no CCPs for a hazard, then, in the literal sense, there is no way for HACCP to control the pathogen. This is the situation with E. cold 0157:H7 in raw ground beef, for which CCPs are yet to be defined. To define CCPs, in turn, it is essential that the ecology and mode of transmission of this pathogen, from the farm to the slaughter, carcass decontamination stage, and into the trims, be understood. The assumption has been that E. cold 0157:H7 is transmitted through feces. However, recent research has suggested that the bacte- rium may also be transmitted by other means such as the oral cavity of animals (Keen and Elder, 2002~. Therefore, the committee points to the urgent need for research on the ecology of E. cold 0157:H7 and other close serotypes in beef, from the farm through transportation, lairage, slaughter, decontamination treatments, and into the trim, and recommends that USDA promptly undertake or fund such research. Parallel research to develop better interventions to prevent contaminated trim destined for ground product, especially ground beef, should be urgently con- ducted as well. In the meantime, until such information on the ecology and mode of trans- mission of this pathogen is available and effective preventive or corrective controls can be applied at the identified CCPs so that HACCP can be put into practice for ground beef, the committee urges regulatory and health authorities to (1) advise those members of the public who would prefer to minimize the risk of this product to cook irradiated and nonirradiated ground beef products to the appropriate temperature, (2) require the products to be clearly labeled with a warning of the potential for harm if not properly cooked, and (3) expand educa- tional efforts to the public and to target commercial and noncommercial food service managers and workers. Once the ecology of E. cold 0157:H7 is better understood, other technologies may prove effective to control it. For example, the concept of selective use for contaminated trim discussed in Chapter 2 (e.g., for irradiation or cooking only) could then be contemplated as an additional tool to protect consumers.

160 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD As mentioned previously, FoodNet data (CDC, 2002) suggest that the occur- rence of illness due to E. cold 0157:H7 has not declined during the past five years, raising questions as to whether the current testing of ground beef for E. cold 0157:H7 is achieving its desired goal. The committee felt that it was important to emphasize the need for testing and interventions prior to the grinding operation. If the contamination of the trim used for ground beef could be reduced, or if contaminated trim could be diverted to other processes, then the potential for contaminated fresh ground beef reaching the consumer would be reduced. The current survey testing at the retail level serves a purpose as a means of monitoring progress on this issue. However, there is also a need for more effective monitor- ing of the process itself. Adulteration of Ready-to-Eat Meats (9 Vol 2 C.F.R. §§301, 303, 317, 318, 319, 320, 325, 331, 381, 417, 430) FSIS also applies the interpretation of adulteration to the presence of any human pathogen in RTE products. RTE meats, even though some may be labeled with instructions to reheat before consumption, are generally considered adulter- ated if they contain organisms or toxins that are hazardous to the public health. As an example, the detectable presence of L. monocytogenes in RTE processed meats, such as hot dogs, would be considered adulteration. The regulations on lethality and stabilization were based on the incidence of salmonellae in precooked, ready-to-serve roast beef (FSIS, l999b). The present concerns with L. monocytogenes in RTE meats are also based on this interpreta- tion of adulteration, and the current tolerance for L. monocytogenes in RTE meats is "none detectable" within the analytical unit (FSIS, l999b). It is difficult to rely on zero tolerance to achieve significant public health improvements. This is even more evident with L. monocytogenes than with E. cold 0157:H7, because L. monocytogenes does not survive the thermal process applied in the processing of RTE meats and contaminates the meat after processing and either before or during packaging. Since L. monocytogenes is a common environ- mental bacterium, there are many potential sources of contamination, including the packaging environment and the employees themselves (FSIS, l999b). The incidence of L. monocytogenes in RTE meats in the United States is low (overall 1.82 percent [Gombas et al., 20031) and the incidence of human listeriosis is apparently declining (CDC,2002~; however, the incidence of L. monocytogenes in these products has not been reduced to zero. Canada, as well as other countries, has recognized that zero tolerance is not practically achievable and has estab- lished numerical standards for the presence of L. monocytogenes in cheeses that do not support the growth of L. monocytogenes. Unless a terminal process can be applied after RTE meat has been sealed in its final packaging, the absence of L. monocytogenes in any randomly selected package of any specific RTE meat cannot be ensured.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 161 Lethality; Standards for the Production of Certain Meat and Poultry Products (9 Vol 2 C.F.R . §§318.17, 381.150) The Lethality and Stabilization Performance Standards for Certain Meat and Poultry Products: Technical Paper (FSIS, 1998c) describe the method FSIS issued to achieve the 7-D reduction of Salmonella in RTE poultry products and the 6.5-D reduction of Salmonella in RTE beef products. The rationale given by FSIS for the lethality guidelines was based on the establishment of a worst-case population of salmonellae, by animal species, then the probability of salmonellae survival in 100 g of finished product after the specific lethality processes was calculated. Specifically, the worst case was defined as an approximate 97.5 percent upper bound for the number of salmonellae in a sample with the highest density of salmonellae from each baseline survey. Considering estimates of 2,300 salmonellae/g in raw poultry, a 30 percent recov- ery rate of salmonellae after processing, and the 97.5 percent defined upper bound, a worst-case value of 37,500 organisms/g was calculated. In a serving size of 143 g of raw product (assuming a serving size of 100 g of the cooked product), there would be approximately 5,362,500 (6.7 log~O) salmonellae. Thus, to mini- mize the risk to the consumers, a process that results in a 7-D reduction of salmonellae would be necessary. From the statistical standpoint, this approach of determining a worst-case scenario is more appropriate than using an arbitrary safety factor in that it allows FSIS to better address any uncertainty associated with the worst-case value. However, the committee believes that several of the estimates were incorrectly assumed, which resulted in an excessively conservative performance standard. For example, the worst-case definition and lethality for RTE poultry products were determined using the raw ground poultry surveys. These surveys had certain limitations, including that they did not cover all of the summer months, and therefore did not completely represent possible seasonal variations in the preva- lence and levels of salmonellae. In addition, the decimal reduction value (the Duo value) was applied on the total population instead of on a per-gram basis. A 7-D reduction would be sufficient to bring the salmonellae population from 10,000,000 to a theoretical 1 cell/g. In fact, when using the highly improbable FSIS worst- case figure of 37,500 salmonellae cells/g, the regulation should require only a 4.5-log reduction or 4.5-D process. Stabilization; Performance Standards for the Production of Certain Meat and Poultry Products (9 Vol 2 C.F.R. §§318.17, 381.150) rl~he standard for stabilization requires no multiplication of toxigenic micro- organisms, such as C. botulinum, and no more than 1-log multiplication of C. perfringens. The stabilization guidelines were derived by assuming a worst- case population for C. perfringens and assuming that at least 1 million cells are

162 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD necessary to result in illness in most cases. The worst-case population was assumed to be 10,000 for both beef and poultry, and therefore, a 1-log increase in population would still maintain a level below the 1 million/g population necessary to cause illness. This is a valid approach and provides an ample margin of safety. However, this margin may be too conservative and may force the meat processor to overprocess products, thus reducing quality. FSIS proposed to codify the chilling recommendations in FSIS Directive 7110.3 (FSIS, 1989) as safe harbors. FSIS determined that this chilling directive would constitute a safe harbor because compliance would yield cooked poultry products that would meet the stabilization performance standard and because most, if not all, establishments were already following this directive. From the statistical and the microbiological perspectives, the paper on the scientific basis for the stabilization standards (FSIS, 1998c) is very confusing and hard to use to determine the validity of either the data or the assumptions. There- fore, it is difficult to critically review this performance standard and assess the validity of the assumptions made during its development. This again illustrates the need for greater transparency in the development of food safety criteria. Cured meat products are not included in this directive and, therefore, the lethality and stabilization standards should not be applied to these products. APPLICATION OF PERFORMANCE STANDARDS WITHIN THE HACCP SYSTEM Beef and Pork The HACCP-based regulatory system is a good example of a regulatory approach that includes government, industry, and the public sector. Various com- panies throughout the food industry have been using HACCP principles since their inception to manage the risk of unsafe products entering commerce, espe- cially for foods that have a terminal process, such as commercially sterile low- acid canned foods. FDA used HACCP principles when promulgating the low- acid canned food regulations (21 C.F.R. Part 114~. The use of performance standards is different from establishing specific microbiological criteria for foods. The National Research Council Subcommittee on Microbiological Criteria addressed the subject of microbiological criteria in raw meats (NRC, 1985a). One of its summary statements was: Microbiological standards for raw meats will prevent neither spoilage nor foodborne illness and thus do not appear warranted. Instead, application of the HACCP system to the entire processing and distribution chain including the meat-processing plant, retail units, foodservice establishment, and home should be used to produce a product with satisfactory shelf-life and public health safety. (NRC, 1985a, P. 198)

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 163 The validity of this conclusion is under scrutiny. At the time of the 1985 report, only three outbreaks of E. cold 0157:H7 had been documented and much was yet to be learned about this microorganism from both a scientific and a societal perspective. The failure of a microbiological criterion to achieve its public health goal is illustrated by the zero tolerance for E. cold 0157:H7 in ground beef; outbreaks still occur. However, due to the potential severity of the resulting illness, especially in children, it may now be inappropriate to establish a level of tolerance other than zero. The PR/HACCP rule established three mandatory provisions (FSIS, 1996~. One provision mandates HACCP systems as a means of preventing or controlling contamination from pathogens. Two other provisions mandate testing for either E. cold biotype I or Salmonella. The E. cold criteria attempt to evaluate the pro- cessing efficacy at slaughter in preventing or removing fecal contamination of the carcasses. The stated purpose of the Salmonella performance standards for slaughter and for grinding operations is to verify that HACCP systems are working. The question of whether Salmonella or other microorganisms should be used to evaluate control of the slaughter process is controversial. USDA held three technical meetings between the time of the proposed rule (February 1995) and publishing the final rule (July 1996~. One of these meetings dealt with the role of microbiological testing in verifying food safety (FSIS, 1996~. Several of the presenters at this meeting advocated the use of E. cold instead of Salmonella as the organism of choice to make evaluations on control of the slaughter process. Arguments made for using E. cold were based upon (1) quantitative results as compared with qualitative results for Salmonella, (2) a much higher association with fecal contamination than Salmonella, and (3) the ability of plants to have results within 24 hours. In contrast, other speakers supported using Salmonella instead of E. coli, primarily because they believed that it would be established that HACCP was indeed reducing microorganisms of concern. It was also argued that the qualitative test for Salmonella was more appropriate because mishandling of the sample after collection would not result in a false positive, whereas mis- handling of a quantitative sample could cause the data to be much higher than at the point of sample collection. Of primary concern is the Salmonella performance standard and its link to HACCP. Within red meats, Salmonella incidence was and continues to be much lower than in poultry. The primary reason is that poultry is produced with its skin on and the skin is the main harbor for bacteria, including Salmonella. In addition, sampling for Salmonella in poultry is done on a whole-carcass rinse rather than on a comparable area of beef. Poultry The PR/HACCP rule indicated that the HACCP principles adopted by the National Advisory Committee for Microbiological Criteria for Food (NACMCF)

64 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD in 1992 would be utilized. However, many poultry companies had not identified fecal contamination as a hazard in their hazard analyses as late as January 1998 because feces did not appear to fit the definitions given by NACMCF for a biological, chemical, or physical hazard. FSIS published a Federal Register notice stating that it considered feces to be a hazard and that HACCP plans would have to have a CCP to deal with visible fecal contamination (FSIS,1997~. Plants were also sent letters giving them 72 hours to respond to this notice in writing to establishment Inspectors-in-Charge, showing that their HACCP plans included feces in the hazard analysis and that at least one CCP had been identified to control the hazard. This notice did make a direct regulatory connection between fecal contamination and the HACCP plan, which may not have been there other- wise. It also created a regulatory connection between the Salmonella perfor- mance standard and fecal contamination because the HACCP plan must address fecal contamination and the Salmonella performance standard is to evaluate the HACCP plan. The measures taken to control fecal contamination have resulted in reduced Salmonella incidence. The post-PR/HACCP directive (FSIS, 1997) where FSIS considers fecal material in prechilled carcasses to be a CCP led to significant changes in broiler processing lines. Primarily, water usage nearly doubled due to the addition of washers, which may have resulted in a dilution of pathogens. Also, continuous on-line reprocessing emerged where antimicrobial rinses were used. Therefore, although published scientific studies have failed to establish a correlation between visible fecal contamination and presence of Salmonella in raw poultry carcasses, the measures taken to control fecal contamination have resulted in reduced Salmonella incidence. Many poultry plants also did not have an identifiable CCP within their process designated to reduce Salmonella to an acceptable level because no point in the slaughter process was designed to control Salmonella incidence on poultry and, therefore, no point met the definition of a CCP (i.e., points where the identified hazard may be prevented from entering the food, eliminated from it, or reduced to acceptable levels; see Chapter 3~. This situation may not have been anticipated by FSIS because the pathogen reduction compo- nent of the rule established procedures for failing to meet the Salmonella perfor- mance standard that included evaluation of the HACCP plan on the first failure, reevaluation and an in-depth verification audit process on the second consecutive failure, and withdrawal of marks of inspection on the third consecutive failure (CDC, 2002). A concern in poultry processing is the possibility of cross-contamination. The process of preparing broilers for consumption is highly automated and there is much opportunity for the cross-contamination and spread of pathogenic micro- organisms among carcasses. This was demonstrated by Lillard (1989), who docu- mented that 3 to 5 percent of the birds were positive for Salmonella when flocks entered the process, which increased to 35 percent positive for carcasses at the end of the process. However, since that study, changes in industry practices may

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 165 have improved this scenario. For example, use of counter-current scalders and chillers, as well as chlorination, have been reported as having a dramatic effect on cross-contamination (Waldroup et al., 1992~. In the USDA baseline study cover- ing 200 broiler processing plants, the national average was down to 20 percent Salmonella-positive carcasses, with an average population of Salmonella of only 38 cfu per positive broiler carcass (Conner et al., 2001~. Another concern is that, whereas ground beef comes from using large quan- tities of lean and fat trims blended to achieve the desired fat level, ground poultry comes from either legs and drumsticks with the skin on, or from backs, necks, and frames after deponing, which may also include the skin. The skin is important to the overall product palatability as well as to the profitability, but it may add Salmonella into the system. Since HACCP implementation, several antimicrobial treatments have been approved in poultry; however, the only treatments that significantly reduce Salmonella are proper cooking or irradiation to a high enough dose. Ground Products In the production of ground products, the PR/HACCP rule acknowledges that grinding establishments cannot use the same technologies for reducing patho- gens that are used by slaughter plants, and that the establishments may have to use raw material contractual specifications to meet the performance standard (FSIS, 1996~. This, in essence, is a confirmation that the process of producing raw ground products does not reduce pathogens and that whatever pathogens are present in the raw material will remain in the finished product. While the Salmo- nella performance standard for ground products provides a guide to overall performance through the slaughter and processing continuum, it may not be appropriate to verify either the HACCP plan or the actual performance of the · 1- gnnalng process. ECONOMIC COSTS AND BENEFITS OF THE PR/HACCP RULE A large share of the recent food safety economics literature has attempted to assess impacts of the PR/HACCP rule (Unnevehr, 2000~. Discussions of the cost of compliance in firms of various sizes and on the potential for changing market structure due to the rule have led this research. This literature is based on the cost-benefit analysis accompanying the PR/HACCP rule (FSIS, 1996), a document that received criticism for its cost assumptions and hypothesized pathogen reductions. The cumulative and specula- tive nature of the cost data is inevitable and correct for the purpose of comparison against similarly aggregated and forecasted benefits of the regulation. However, no study has been able to use actual cost data linked to true plant-level hazard reductions associated with identifiable strategies adopted by firms in response to

66 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD the PRIHACCP rule including interventions targeting microbial, chemical, and physical food safety concerns to address such criticisms retrospectively. As mentioned in Chapter 2 and earlier in this chapter, the link between quantifiable reductions in foodborne illness and the direct actions of firms is not clear, and thus it is not yet possible to directly relate benefits and costs (Kuchler and Golan, 1999~. The data requirements to accurately assess the societal impacts, even for one pathogen (e.g., Salmonella), would include enumeration of each plant's reduction in prevalence following the policy linked to fixed and variable costs of the particular strategy or intervention under analysis. For example, if a plant purchased a lactic acid carcass decontamination unit only because of the perfor- mance standard, economists would need information of initial cost of $x and annual recurring costs of $y, as well as the z percent logic reduction in the incidence of Salmonella to correctly assess the impact of the standard. This information would be required of each plant and pathogen and would then still need to be linked to impacts on public health and changes in the relationships between firms at various stages of the supply chain. The most contentious cost issue in USDA's regulatory impact assessment focused on the details of process modifications required by firms to ensure com- pliance with the pathogen reduction standards. The rule established performance standards for Salmonella for all plants that slaughter and that process raw ground product. Further, all slaughter establishments are to employ a generic E. cold testing program to validate their process. Debate has centered on the additional equipment costs required by plants of various sizes, the potential structural impli- cations of the standards, and the relationship between the recurring and non- recurring elements of such process modification and other related PRIHACCP costs. The in-house review of costs of current pathogen reduction strategies per- formed by FSIS based on plant size suggests that manual hot water spraying was the most cost-effective intervention for small slaughter facilities (8¢ per carcass). Alternative strategies that were considered included a pre-evisceration acid-spray system with both a prewash spray cabinet and a sanitizing cabinet at a cost of 79¢ per carcass for low volume use, and a trisodium phosphate-based system at a cost of 85¢ per carcass. The use of steam vacuum systems, with a nonrecurring cost of $10,000 and a recurring cost of around $4,500, was also discussed. The poultry data were based on the use of trisodium phosphate rinses, estimated to cost $40,000 per line. (Large poultry establishments average two lines, small ones average one and one-half.) This translates to a cost of 0.3¢ per broiler and 1.4¢ per turkey. The use of both high and low scenarios of the costs of process modification in the final regulatory impact assessment is indicative of the methodological problem within the analysis. The low-cost scenario was based on the assumption that 10 percent of the 66 large hog and beef slaughter plants would need to install a steam vacuum system to ensure compliance with the Salmonella performance

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 167 standard. Further, "half of the 376 small establishments must install a hot water rinse at $.08 per carcass" (FSIS, 1996~. Conversely, the high-cost scenario sug- gested that 100 percent of the small and very small plants and as many as half (33) of the large plants (implying that the other plants already have such systems in place) would need to incur these costs. For facilities that do not slaughter (i.e., grinders), however, process modification costs for compliance with the Salmo- nella performance standard were not calculated; this approach suggests that these plants "must depend on the Salmonella levels of their incoming product to meet the performance standards" (FSIS, 1996~. This one clear statement made by FSIS meant that no additional costs were included or anticipated for compliance with the performance standard for grinders (which include the Supreme Beef plant). Thus, the cost-benefit analysis contained in the final rule assumed that compli- ance with the other portions of the PR/HACCP rule would lead to higher costs, but that the PR portion would not. An analysis of neither the marginal impact of the performance standard, nor its potential as a dynamic policy tool, has been attempted. A similar exercise in process modification costs for poultry suggests that the low-cost scenario would have 36 large establishments installing a trisodium phosphate-based system, with the high-cost scenario increasing this number to 182 (100 large and 82 small plants). Finally, the process modification costs for the generic E. cold sampling standard were related to the Salmonella performance standard. FSIS concluded that . . . if the low cost scenario for compliance with Salmonella standards proves to be more accurate, there will likely be more separate compliance costs for generic E. coli. As the costs for Salmonella compliance go up, the likelihood of separate E. cold costs goes down. It is important to note that under the high cost scenario, all cattle and swine slaughter establishments are using the steam vacuum system or hot water rinse and half of all poultry slaughter establishments are using TSP systems. Under this scenario, it is difficult to imagine that any establishments would still be failing to meet the performance criteria for generic E. coli. (FSIS, 1996, Pp. 38981-38982) Little consideration was given to the unique costs related to compliance with the performance standards other than to suggest the adoption of equipment that appears to have become standard in most large slaughter operations. In order to assess these estimates, Jensen and Unnevehr (2000) calculated the minimal costs of attaining a range of pathogen standards for large pork slaughter plants. Among the strategies selected were water rinses at three temperatures, with and without the application of a sanitizing spray. The per carcass costs of the wash and spray systems were found to be below the 79¢ and 85¢ estimates discussed above. Even when the most restrictive pathogen standard was simu- lated, costs were still under 50¢ per carcass, suggesting that the regulatory impact

68 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD assessment may have overestimated the costs. The authors also attempted to test which of the two FSIS cost scenarios is more appropriate. They suggested that if the selection of an intervention strategy is made on a least-cost basis, then actual process modification costs may be higher than suggested in the regulatory impact assessment for large pork-slaughter plants. Jensen and Unnevehr (2000) present a clear framework for incorporating pathogen reduction data into their assessment of least-cost interventions. How- ever, care must be taken in applying these microbiological results. As the authors admit, their data come from two separate (although small) sources. One study tested interventions in a plant environment; the other did not. One used inoculated samples; the other did not. The inoculation procedure effectively elevates patho- gen populations to an observable level, thus implying that although real-world reductions (the results of interventions) will not be of the same magnitude, they will be of the same relative order. This remains an untested hypothesis for most interventions. Without further analysis, it cannot be presumed that a certain logic reduction due to an intervention will be an improvement over current strategies; that it will be achieved in all plants at all times, regardless of the "cleanliness" of animals being presented for slaughter; or that it will lead to a risk reduction downstream at the point of consumption. Therefore, it may be more appropriate to presume that this analysis overestimated the benefits to the consumer. Broader Economic Impacts: What Needs to Be Assessed? Several potential indirect impacts should be considered in the broader eco- nomic analysis of the PR/HACCP rule. First are scale effects or implementation costs, which differ significantly by plant size. As HACCP-based regulations expand in their coverage (e.g., to the retail sector with many small and very small firms), it is argued that scale effects will be of paramount importance. The food safety system put in place by a plant can also impact nonsafety quality attributes, thus increasing overall efficiency (Unnevehr and Roberts, 1997~. That HACCP can help limit product rejection or rework, thus reducing the variability inherent to all production processes, also deserves more attention. This benefit allows for increased customer and consumer satisfaction (e.g., reduced complaints and product return); and may increase, although it is often difficult to quantify, measures of consumer confidence. Also, international trade is clearly facilitated when harmonized HACCP-based regulations are adopted (Caswell and Hooker, 1996~. Potential legal liability and insurance cost savings can arise from the use of innovative food safety controls. An advantage can be achieved by those plants and firms that are first to adopt a proven intervention. This can improve the overall company image, potentially providing a competitive and marketing advantage. Such innovation offset dynamics are discussed by Cockbill (1991)

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 169 and Hobbs and Kerr (1992) and may or may not be candidates for inclusion in future regulatory impact assessments, depending upon the details of the HACCP- based regulation under consideration. Difficulties in Forecasting Costs and Benefits for Novel Innovations The PR/HACCP rule has an admirable degree of flexibility (i.e., minimal process criteria). Further, the performance standard elements of the rule seem to have provided some incentive to promote innovation in the pathogen reduction strategies employed. However, in part due to such success in regulatory design, ex post costs may differ significantly from ex ante estimates as more plants adopt validated pathogen reduction strategies that differ from those that USDA pre- sumed would be used. This is further confounded when the selection of such strategies is not made on a least-cost basis. Limited economic research exists to provide reliable estimates of costs and resultant benefits of many food safety interventions. Several pathogen reduction strategies, particularly multiple-hurdle techniques, incorporate novel approaches for which only limited commercial applications exist, thus requiring a cautious approach to forecasting potential costs. Further, plant-level pathogen reduction benefits of multiple-hurdle interventions are not always simply additive. The potential use of novel individual interventions, as well as innovative combinations of traditional interventions, clearly make the prerule estimation of costs and benefits extremely difficult. It seems likely that in future regulatory impact assessments, the role of pilot programs to forecast real-world impacts will be expanded. Hopefully, these studies will utilize representative firms' experiences with HACCP (or whatever food safety controls are being considered) and consider all state-of-the-art interventions. Special care must be taken in estimating the impact of any novel intervention not widely adopted in the industry based on plant-level experiences and not just on laboratory or theoretical assessments. At all times, the effectiveness of novel interventions should be compared with current systems on a microbiological as well as a cost basis. THE NEED FOR ADDITIONAL APPROACHES TO REDUCE MICROBIAL HAZARDS Preventing Pathogen Contamination and Amplification Before Slaughter Pathogens, including E. cold 0157:H7, Salmonella, and Campylobacter, on hides and in internal organs of live animals arriving for slaughter are important sources of contamination of meat. Substantial surveys of pathogen prevalence in dairy herds, feedlot populations, and culled dairy cattle have been conducted. Surveys of dairy farms show that a small percentage of farms or animal feces are

170 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD positive for E. cold 0157:H7 at a single point in time but, with repeated sampling, the organism is likely to be detected on most farms (Hancock et al., 1998~. In a survey of 36 dairy herds, with repeated sampling over six months, the pathogen was ultimately detected on 75 percent of the herds, probably because carriage lasts no more than a few weeks in any animal (Hancock et al., 1997a, 1997b). The prevalence of fecal shedding of E. cold 0157:H7 was 0.9 percent among dairy cows and 2.9 percent among dairy cows about to be culled; these data suggest that culling either selects for animals likely to be contaminated or contributes to their contamination. On average, 24.2 percent of dairy operations had at least one positive animal; this prevalence was seasonal, increasing in the summer months. Surveys of beef cattle in feedlots show a similar pattern, though the prevalence of contamination is generally higher (Veterinary Services, 2001a). Lately, methods based on immunomagnetic separation have allowed better detection of animals shedding low levels of E. cold 0157:H7 (Besser et al., 2001~. Due to the higher sensitivity methods, it is currently believed that the prevalence of E. cold 0157:H7 is higher than previously thought. Given that about 23 percent of the nation's dairy herd is culled and sent for slaughter annually (APHIS, 1996) and that much of it becomes ground beef (Troutt et al., 2001), the committee concludes that prevalence data on E. cold 0157:H7 in culled animals is needed. Better understanding of the circumstances associated with the presence of pathogens could lead to targeted efforts to miti- gate or prevent their circulation among live animals. E. cold 0157:H7 is a hardy pathogen, able to survive in damp cattle manure for up to 70 days, to survive and multiply in the sediment of cattle water troughs for months, to rapidly grow in moist cattle ration, and to be carried by wild deer (Keene et al., 1997; LeJeune et al., 2001; Lynn et al., 1998; Wang et al., 1996~. Epidemiological studies that link the presence or absence of the organism in a herd to various management practices have suggested stronger association with using corn-based feed or feeding barley than with feeding soy meal or spreading fresh manure on forage crops (Dargatz et al.,1997; Hancock et al., 1997b; Herriott et al., 1998~. The rumen of a fasted animal may be more hospitable to growth of Salmonella and E. cold 0157, and it has been suggested that the common practice of fasting animals preslaughter may increase the shedding and spread of E. cold 0157:H7 (Hancock et al., 1998; Rasmussen et al., 1993~. Salmonella are also commonly present among dairy herds and feedlots. The 1996 NAHMS survey of dairy cattle reported a prevalence of 5.4 percent among animals and 27 percent among dairy operations sampled a single time (Wells et al., 1998~. As with E. cold 0157:H7, the data also suggest that the level is higher in culled animals. Among feedlot cattle, the prevalence of Salmonella was 6.3 percent in animals, 22.3 percent in pens, and 51 percent in feedlots (Veterinary Services, 2001b). Factors associated with the presence of Salmonella on farms have not been examined as thoroughly as for E. cold 0157:H7. Nevertheless, some general principles of control of Salmonella among cattle herds have been

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 171 defined and are also applicable to the control of Salmonella Typhimurium DT104 (Dargatz et al., 1998) and other important animal-borne illnesses such as Johne's Disease (Groenendaal and Galligan, 1999; Wells et al., 1999~. In addition to the prevalence on the farm, other factors that may increase the risk of pathogens in meat relate to the transportation of herds in trucks from a pasture or barn through auction yards, feedlots, and holding pens, where they are exposed to fecal or other means of contamination from animals previously or currently there. The 1996 NAHMS survey of dairy cattle reported that 15 percent of feces from individual culled dairy cattle were positive for Salmonella at market and that 67 percent of markets had at least one animal shedding Salmonella (Wells et al., 1998~. Furthermore, a recent systematic national survey of 5,000 culled dairy cattle reported that 23 percent of animals carried Salmonella at the point of slaughter, with a range of O to 93 percent on a given day at a given establishment (Troutt et al., 2001~. Similarly, the prevalence of E. cold 0157:H7 among culled dairy cattle at market in the NAHMS study was 1.8 percent, twice that on the farm, and, when tested a single time, 31 percent of the markets had a positive animal (Wells et al., 1998~. In a recent survey of cattle in 29 pens in 5 major feedlots, and based on a single fecal sample from each animal, 23 percent of individual animals and 100 percent of feedlot pens were positive for E. cold 0157:H7 (Smith et al., 2001~. The environmental conditions in the pen (e.g., muddy grounds after a rain) were associated with the likelihood of finding the pathogen. The final point of potential introduction and amplification of live-animal contamination with pathogens is the holding pens immediately before slaughter- ing. Two recent studies suggest that, for swine and cattle, the abattoir terminal holding pen is a significant point of contamination with E. cold 0157:H7 and that, therefore, sanitation of the terminal holding pen is likely to be an important control point for this pathogen (Avery et al., 2002; Hurd et al., 2001~. In summary, the committee concludes that efforts to reduce preslaughter contamination are likely to be an important part of a farm-to-table food safety strategy, not only to reduce pathogen load at the slaughter plant, but also to prevent the hazard from direct contact with infected animals, from runoff on feedlots and farms, and from contaminated water supplies (Crump et al., 2002; Hilborn et al., 1999; Kassenborg et al., 1998; Martin et al., 1986; O'Brien and Adak, 2002; PPHB,2000~. This prevention process, beneficial to both animal and human health, comprises on-farm management practices that may reduce the spread and amplification of pathogens, as may sanitation practices during trans- portation and in feedlots, final holding pens, and slaughter boxes. Moreover, measures that increase the resistance of animals to intestinal contamination in the last days of their lives should be examined and evaluated through formal inter- vention trials.

72 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD Therefore, the committee recommends that USDA conduct or fund research on the role of nonfocal carriage and commingling prior to and after slaughter to elucidate the factors that contribute to the microbial pathogen contamination of live animals, carcasses, and products. The committee also recommends a research focus on intervention trials at all stages of the production process of meat and poultry products. The committee further concludes that the level of contamination of animals coming to slaughter is likely to be associated with the contamination of the meat; therefore, monitoring levels of contamination on and in the incoming animals is likely an important measurement of the level of risk and could help determine or require the use of mitigation steps. More importantly, measures that may reduce such contamination, such as changing what animals are fed in the last week of life, reducing fecal contamination on hides in the muddy seasons, or sanitizing the terminal holding pen and kill box, should be rapidly evaluated so that the level of contamination at the slaughter plant may be reduced. Consequently, the committee recommends that industry and regulatory agen- cies continue to place greater emphasis on contamination prevention rather than rely on inspection and end-product testing to ensure the safety of meat. Monitoring Pathogen Contamination of Herds and Flocks to Assign Raw Foods to Further Processing The nature of foodborne hazards has changed dramatically over the last century since the first federal meat inspection system was created. The hazard posed by diseased and dying animals has been replaced by hazards that are more difficult to detect. Common zoonotic pathogens such as Campylobacter in broilers, S. Enteritidis in layers, E. cold 0157:H7 in cattle, and Yersinia enterocolitica in pork cause no apparent illness in the food animals that harbor them, yet can contaminate the foods produced from these animals. Public health surveillance and investigations have attempted to measure the human illness burden that these and other foodborne pathogens cause, and have traced them back to food animal reservoirs. In the absence of grossly visible markers for contamination of live animals with microbial pathogens, the effectiveness of new systems for control may depend on such measures as accurate separation of higher-risk flocks or herds from others. The Pennsylvania Egg Quality Assurance Program, for example, is an S. Enteritidis control program in layer flocks that began in 1992 (FSIS, 2002b). Routine monitoring of flocks for the presence of S. Enteritidis is part of this program and is linked to vigorous efforts to prevent contamination of the next generation of birds that will enter the farm, as well as to the diversion to pasteurization of eggs from contaminated flocks. The result has been a slow but steady decline in the proportion of egg-producing facilities that have S. Enteritidis, from 25.7 percent in 1994 to 7.3 percent in 1998 (PFMA, 2000~.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 173 A review of the change in prevalence of the four most common Salmonella serotypes found in broiler chickens in the United States indicated that all four declined substantially and significantly after the PR/HACCP rule was imple- mented (RTI, 2002~. In other countries, even more dramatic declines have been achieved by using microbial monitoring to drive farm- or flock-based control efforts. Sweden has largely controlled S. Enteritidis in chicken-rearing operations (Wierup et al., 1995~. This achievement, however, has come at a high cost derived from destruction of contaminated flocks. The European Union, in turn, issued a directive in 1992 mandating the screening of flocks and herds for S. Enteritidis and S. Typhimurium with a view to subsidized destruction of those found to be contaminated (EC, 1992~; Denmark, Finland, Sweden, and Ireland joined the program by 1999 (Murder and Schlundt, 1999~. However, given the vast differ- ence in the scale of poultry production between the United States and European countries, such an approach would need to be structured differently in the United States. In 2001, Norway launched a national control program for Campylobacter based on the testing of chicken flocks and of finished carcasses; chickens from positive flocks are slaughtered after the negative flocks to minimize cross- contamination, and the carcasses are either sent for supervised cooking or are frozen (Norwegian Zoonosis Centre, 2002~. It is too soon to tell whether carcass contamination with Campylobacter has actually been reduced as a result of this program. DO MEAT AND POULTRY PERFORMANCE STANDARDS IMPROVE PUBLIC HEALTH? The committee recognizes that substantial declines in four bacterial foodborne diseases observed in the United States via FoodNet surveillance since 1996 indicate that the collective efforts to improve food safety are having an effect (CDC, 2002~. As the most prominent declines are in infections caused by the meat-associated pathogens Campylobacter, Listeria, and Y. enterocolitica- 27, 35, and 49 percent declines, respectively it is likely that the PR/HACCP rule is contributing to this effect, although concurrent changes in distribution, retail, and consumer behavior could also be important in decreasing infections due to such pathogens (CDC, 2002~. The fact that no sustained decline has been observed yet in infections caused by E. cold 0157:H7 may mean that the estab- lished zero tolerance for this pathogen does not offer added protection, perhaps because the principal determinants of contamination are preslaughter, or perhaps because it was effective and blunted what otherwise would have been an increase. The data needed to distinguish between these possibilities are lacking. The de- cline in listeriosis is particularly noteworthy. Listeriosis declined between 1988 and 1995 and had appeared to reach a plateau. Further industry efforts, including formulation and process changes, stimulated by a large outbreak associated with

74 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD hot dogs in 1999, as well as efforts to educate high-nsk populations, may have resulted in an additional 35 percent decline (CDC, 2002) in human cases. A persistent challenge is that attributing such changes to any one factor is difficult because many food safety measures may be taking place at the same time, and because a given infection may have multiple possible food and nonfood sources. As was recommended in Chapter 2, measuring changes in consumer behavior, as well as microbial subtyping of pathogen strains from different food sources and comparison with isolates from human infections, could help conquer this challenge. REFERENCES APHIS (Animal and Plant Health Inspection Service). 1996. Dairy '96. Part 1: Reference of 1996 Dairy Management Practices. Online. U.S. Department of Agriculture (USDA). Available at http://www.aphis.usda.gov/vs/ceah/cahm/Dairy_Cattle/dr96desl.pdf. Accessed July 19, 2002. Avery S. Small A, Reid CA, Buncic S. 2002. Pulsed field gel electrophoresis characterization of Shiga-toxin producing Escherichia cold 0157:H7 from hides of cattle at slaughter. J Food Prot 65: 1 172-1 176. Becker E. 2003, January 23. Government in showdown in bid to shut beef processor. New York Times. P. A16. Besser TE, Richards BL, Rice DH, Hancock DD. 2001. Escherichia cold 0157:H7 infection of calves: Infectious dose and direct contact transmission. Epidemiol Infect 127:555-560. Caswell JA, Hooker NH. 1996. HACCP as an International Trade Standard. Am JAgric Econ 78:775- 779. CDC (Centers for Disease Control and Prevention). 2002. Preliminary FoodNet Data on the inci- dence of foodborne illnesses Selected sites, United States, 2001. Morb Mortal Wkly Rep 51 :325-329. CFSAN (Center for Food Safety and Applied Nutrition). 2003. Real Progress in Food Code Adop- tions. Online. Food and Drug Administration (FDA). Available at http://vm.cfsan.fda.gov/~ear/ fcadopt.html. Accessed April 10, 2003. Cockbill C. 1991. The Food Safety Act: An introduction. Br Food J 93:4-7. Conner DE, Davis MA, Zhang L. 2001 Poultry-borne pathogens: Plant considerations in poultry meat processing. In: Sams AR, ed. Poultry Meat Processing. Boca Raton, FL: CRC Press. Pp. 137-158. Crump J. Sulka A, Angulo FJ. 2002. An outbreak of Escherichia cold 0157:H7 infections among visitors to a dairy farm. N Engl J Med 347:555-560. CVM (Center for Veterinary Medicine). 2002a. Animal Medicinal Drug Use Clarification Act of 1994 (AMDUCA). Online. FDA. Available at http://www.fda.gov/cvm/index/amducca/ amducatoc.htm. Accessed February 14, 2003. CVM. 2002b. Guidance for Industry. Evaluating the Safety of Antimicrobial New Animal Drugs with Regard to Their Microbial Effects on Bacteria of Human Health Concern. Online. FDA. Avail- able at http://www.fda.gov/OHRMS/DOCKETS/98fr/80040a56.pdf. Accessed May 13, 2003. Dargatz D, Scott W. Thomas LA, Hancock D, Garber L. 1997. Factors associated with the presence of Escherichia cold in feces of feedlot cattle. J Food Prot 60:466-470. Dargatz D, Wells S. Fedorka-Cray P. Akkina, J. 1998. The veterinarian's role in diagnosis, treat- ment, and prevention of multidrug resistant Salmonella Typhimurium DT104. Bovine Practi- tioner 32: 1-6.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 175 EC (European Commission). 1992. Council Directive 92/117/EEC of 17 December, 1992 concerning measures for protection against specified zoonoses and specified zoonotic agents in animals and products of animal origin in order to prevent outbreaks of food-borne infections and intoxi- cations. Off J Eur Communities L062:38-48. Eilert SJ, Rathje NM. 2001. Case ready red meat: Demand and technology. Proc Recip Meat Conf 54:59-61. ERS (Economic Research Service). 2000. Cattle: Background. Online. USDA. Available at http:// www.ers.usda.gov/Briefing/Cattle/Background.htm. Accessed June 25, 2002. ERS. 2002. Cattle: Trade. Online. USDA. Available at http://www.ers.usda.gov/Briefing/Cattle/ Trade.htm. Accessed June 25, 2002. FAIM (Field Automation and Information Management Division). 2002. FAIM Division Home Page. Online. Food Safety and Inspection Service, USDA. Available at http://www.fsis.usda.gov/ OFO/FAIM/fainmain.htm. Accessed May 15, 2002. FDA (Food and Drug Administration). 2001. Food Code 2001. Online. Available at http:// www.cfsan.fda.gov/~dms/foodcode.html#getOl. Accessed January 21, 2002. FSIS (Food Safety and Inspection Service). 1989. Time/Temperature Guidelines for Cooling Heated Products. Online. USDA. Available at http://www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/ FSISDir7110-3Revl.pdf. Accessed February 3, 2002. FSIS. 1995. Compilation of Issuances on Cooked Meat and Poultry Product Temperatures. Online. USDA. Available at http://www.fsis.usda.gov/OPPDE/rdad/FSISDirectives/FSISDir7370.2.pdf. Accessed February 3, 2002. FSIS. 1996. Pathogen reduction; Hazard analysis and critical control point (HACCP) systems; Final rule. Fed Regist 61: 38805-38855. FSIS. 1997. Livestock carcasses and poultry carcasses contaminated with visible fecal material. Fed Regist 62:63254-63255. FSIS. 1998a. Importing Meat and Poultry to the United States. Online. USDA. Available at http:// www.fsis.usda.gov/oa/programs/import.htm. Accessed June 26, 2002. FSIS. 1998b. Key Facts: Role of the Inspector Under HACCP. Online. USDA. Available at http:// www.fsis.usda.gov/oa/background/keyrole.htm. Accessed February 3, 2002. FSIS. 1998c. Lethality and Stabilization Performance Standards for Certain Meat and Poultry Prod- ucts: Technical Paper. Online. USDA. Available at http://www.fsis.usda.gov/OPPDE/rdad/ FRPubs/95-033F_tech%20paper.pdf. Accessed June 26, 2002. FSIS. 1999a. Beef products contaminated with Escherichia cold 0157:H7. Fed Regist 64:2803-2805. FSIS. l999b. Performance standards for the production of certain meat and poultry products. Fed Regist 64:732-749. FSIS. 2001a. HACCP-Based Inspection Models Project (HIMP): Young Chicken Inspection. Online. USDA. Available at http://www.fsis.usda.gov/OPPDE/Nis/HIMP/Docs/YNGChk_Drf6.pdf. Accessed July 19, 2002. FSIS. 2001b. Mandatory Inspection of Ratites and Squabs. Online. USDA. Available at http:// www.fsis.usda.gov/OPPDE/rdad/FRPubs/01-045IF.htm. Accessed May 23, 2003. FSIS. 2001c. Protecting the Public From Foodborne Illness: The Food Safety and Inspection Ser- vice. Online. USDA. Available at http://www.fsis.usda.gov/oa/background/fsisgeneral.htm. Accessed June 26, 2002. FSIS. 2002a. E. cold 0157:H7 contamination of beef products. Fed Regist 67:62325-62334. FSIS. 2002b. Egg Quality Assurance. Online. USDA. Available at http://www.fsis.usda.gov/OPPDE/ animalprod/Publications/overview/qa/a84.htm. Accessed July 19, 2002. FSIS, 2002c. FAIM Division Home Page. Online. USDA. Available at http://www.fsis.usda.gov/ OFO/faim/faimmain.htm. Accessed May 15, 2002. FSIS. 2003. FSIS Import Reinspection. Online. USDA. Available at http://www.fsis.usda.gov/ OPPDE/IPS/Importing.htm#Reinspection. Accessed May 10, 2003.

176 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD GAO (General Accounting Office). 2001. Food Safety: Overview of Federal and State Expenditures. GAO-01-177. Washington, DC: GAO. GAO. 2002. Meat and Poultry: Better USDA Oversight and Enforcement of Safety Rules Needed to Reduce Risk of Foodborne Illnesses. Appendix II: Comments from the U.S. Department of Agriculture. GAO-02-902. Washington, DC: GAO. Gombas DE, Chen Y. Clavero RS, Scotta VN. 2003. Survey of Listeria monocytogenes in ready-to- eat foods. J Food Prot 559-569. Groenendaal H. Galligan D. 1999. Economical Consequences of Johne's Disease Control Programs. Kennet Square, PA: Center for Animal Health and Productivity, University of Pennsylvania. Hancock D, Besser T. Rice DH, Herriott DE, Tarr PI. 1997a. A longitudinal study of Escherichia cold 0157 in fourteen cattle herds. Epidemiol Infect 118:193-195. Hancock DD, Rice DH, Herriot DE, Besser TE, Ebel ED. 1997b. Effects of farm manure-handling practices on Escherichia cold 0157 prevalence in cattle. J Food Prot 60:363-366. Hancock D, Besser T. Rice DH. 1998. Ecology of Escherichia cold 0157:H7 in cattle and impact of management practices. In: Kaper J. O'Brien A, eds. Escherichia cold and other Shiga toxin- producing E. cold Strains. Washington, DC: ASM Press. Pp. 85-91. Herriott D, Hancock D, Ebel ED, Carpenter LV, Rice DH, Besser TE. 1998. Association of herd management factors with colonization of dairy cattle by Shiga toxin-positive Escherichia cold 0157. J Food Prot 61 :802-807. HHS (U.S. Department of Health and Human Services). 2002. Foodborne Illnesses Post Dramatic Six-Year Decline. Online. HHS Press Release 4/18/2002. Available at http://www.hhs.gov/ news/press/2002pres/200204189.html. Accessed December 15, 2002. Hilborn E, Mermin J. Mshar PA, Hadler JL, Voetsch A, Wojtkunski C, Swartz M, Mshar R. Lambert- Fair MA, Farrar JA, Glynn MK, Slutsker L. 1999. A multistate outbreak of Escherichia cold 0157:H7 infections associated with consumption of mesclun lettuce. Arch Intern Med 159: 1758-1764. Hobbs JE, Kerr WA. 1992. Costs of monitoring food safety and vertical coordination in agribusiness: What can be learned from the British Food Safety Act 1990? Agribus Internat J 8:575-584. Hurd H. McKean J. Wesley IV, Karriker LA. 2001. The effect of lairage on Salmonella isolation from market swine. J Food Prot 64:939-944. ICMSF (International Commission on Microbiological Specifications for Foods). 2002. Micro- organisms in Foods 7. Microbiological Testing in Food Safety Management. New York: Kluwer Academic/Plenum Publishers. Jensen HH, Unnevehr LJ. 2000. HACCP in pork processing: Costs and benefits. In: Unnevehr LJ, ed. Economics of HACCP: Costs and Benefits. St. Paul, MN: Eagen Press. Pp. 29-44. Kassenborg H. Hedberg C, Evans M, Chin G. Fiorentino T. Vugia D, Bardsley M, Slutsker L, Griffin P. 1998. Case-control study of sporadic Escherichia cold 0157:H7 infections in 5 FoodNet sites. In: Proceedings of the International Conference on Emerging Infectious Diseases. Atlanta: CDC. P. 50. Keen JE, Elder RO. 2002. Isolation of shiga-toxigenic Escherichia cold 0157 from hide surfaces and the oral cavity of finished beef feedlot cattle. J Am Vet Med Assoc 220:756-763. Keene W. Sazie E, Kok J. Rice DH, Hancock DD, Balan VK, Zhao T. Doyle MP. 1997. An outbreak of Escherichia cold 0157:H7 infections traced to jerky made from deer meat. JAMA 227:1229- 1231. Kinsman DM. 1994. Historical perspective and current status. In: Kinsman DM, Kotula AW, Breidenstrin BC, eds. Muscle Foods: Meat, Poultry and Seafood Technology. New York: Chapman and Hall. Pp. 1-24. Kuchler F. Golan E. 1999. Assigning Values to Life: Comparing Methods for Valuing Health Risks. Agricultural Economics Report No. 784. Washington, DC: ERS, USDA. LeJeune J. Besser T. Hancock DD. 2001. Cattle water troughs as reservoirs of Escherichia cold 0157. Appl Environ Microbiol 67:3053-3057.

CONTROLS FOR HAZARDS IN MEAT AND POULTRY PRODUCTS 177 Lillard HS. 1989. Factors affecting persistence of Salmonella during the processing of poultry. J Food Prot 52:829-832. Lynn TV, Hancock DD, Besser TE, Harrison JH, Rice DH, Stewart NT, Rowan LL. 1998. The occurrence and replication of Escherichia cold in cattle feeds. J. Dairy Sci 81:1102-1108. MacDonald JM, Ollinger ME, Nelson KE, Handy GR. 2000. Consolidation in U.S. Meatpacking. Agricultural Economics Report No. 785. Washington, DC: Agricultural Research Service, USDA. Martin DL, Shipman LD, Wells JO, Potter ME, Hedberg K, Wachsmuth IK, Tauxe RV, Davis JP, Arnoldi J. Tilleli J. 1986. Isolation of Escherichia cold 0157:H7 from dairy cattle associated with two cases of haemolytic uraemic syndrome. Lancet 2:1043. Mulder RWA, Schlundt J. 1999. Safety of poultry meat: From farm to table. In: Molins RA, Corry J. eds. International Consultative Group on Food Irradiation (ICGFI). Vienna: International Atomic Energy Agency. Nalivka J. 2002. Meat and Poultry Facts Book 2002. Mt. Morris, IL: Watt Publishing. NASS (National Agriculture Statistics Service). 2002. Livestock Slaughter: 2001 Summary. Online. USDA. Available at http://usda.mannlib.cornell.edu/reports/nassr/livestock/pls-bban/lsanO302.txt. Accessed October 9, 2002. Norwegian Zoonosis Centre. 2002. Action plan—Campylobacter sp. in Norwegian broilers. Summary of results from the surveillance of broiler flocks 2002; Survey of broiler meat products 2002. Online. Available at http://www.vetinst.no/Zoonosesenteret/campyl-report-action-2002.pdf. Accessed September, 11, 2002. NRC (National Research Council). 1985a. An Evaluation of the Role of Microbiological Criteria for Foods and Food Ingredients. Washington, DC: National Academy Press. NRC. 1985b. Meat and Poultry Inspection: The Scientific Basis of the Nation's Program. Washing- ton, DC: National Academy Press. NRC. 1987. Poultry Inspection: The Basis for a Risk-Assessment Approach. Washington, DC: Na- tional Academy Press. NRC. 1999. The Use of Drugs in Food Animals: Benefits and Risks. Washington, DC: National Academy Press. O'Brien S. Adak G. 2002. Escherichia cold 0157:H7 Piecing together the jigsaw puzzle. N Engl J Med 347:608-609. PFMA (Pennsylvania Food Merchants Association). 2000. Pennsylvania Presents Results of Egg Quality Assurance, January 2000. Online. Available at http://www.pfma.org/media/advisor/ JAN00/peqap.html. Accessed January 23, 2003. PPHB (Population and Public Health Branch). 2000. Waterborne outbreak of gastroenteritis associ- ated with a contaminated municipal water supply, Walkerton, Ontario, May-June 2000. Can Commun Dis Rep 26-20:170-173. Rasmussen M, Cray W. Casey TA, Whipp SC. 1993. Rumen contents as a reservoir of entero- hemorrhagic Escherichia coli. FEMS Microbiol Lett 114:79-84. REAP (Research-Education-Advocacy-People). 2001. REAP General 2001 Report on the Meat Packing Industry. Online. Available at http://www.reapinc.org/ REAP%202001%20REPORT%200N%20MEAT%20PACKING.html. Accessed June 26, 2002. Rose BE, Hill WE, Umholtz R. Ramnsom GM, James WO. 2002. Testing for Salmonella in raw meat and poultry products collected at federally inspected establishments in the United States, 1998 through 2000. J Food Prot 65:937-947. RTI (Research Triangle Institute). 2002. Revised Final Draft: PR/HACCP Rule Evaluation Report. Effect of the 1996 PR/HACCP Rule on Foodborne Illness: Young Chicken Salmonella Study. Research Triangle Park, NC: RTI. Shire B. 2003. Do You Own/Operate a Raw Beef Plant? Here's How to Reassess Your HACCP Plan for E. cold 0157:H7. AAMP Special Report. Elizabethtown, PA: American Association of Meat Processors.

78 SCIENTIFIC CRITERIA TO ENSURE SAFE FOOD Smith D, Blackford M, Younts S. Moxley R. Gray J. Hungerford L, Milton T. Klopfenstein T. 2001. Ecological relationships between the prevalence of cattle shedding Escherichia cold 0157:H7 and characteristics of the cattle or conditions of the feedlot pen. J Food Prot 64: 1899-1903. Troutt H. Galland J. Osburn BI, Brewer RL, Braun RK, Schmitz JA, Sears P. Childers AB, Richey E, Mather E, Gibson M, Murthy K, Hogue A. 2001. Prevalence of Salmonella spp. in cull (market) dairy cows at slaughter. JAm Vet Med. Assoc 219: 1212-1215. University of Vermont. 2003. Beef Cattle. Online. Available at http://asci.uvm.edu/course/asciO01/ beef.html. Accessed January 15, 2003. Unnevehr LJ, ed. 2000. The Economics of HACCP: Studies of Costs and Benefits, 2000. St. Paul, MN: Eagan Press. Unnevehr L, Roberts T. 1997. Improving cost/benefit analysis for HACCP and microbial food safety: An economist's overview. In: Caswell JA, Cotterill RW, eds. Strategy and Policy in the Food System: Emerging Issues. Amherst, MA: University of Massachusetts. Pp. 225-229. USDA (U.S. Department of Agriculture). 2002. USDA Data Show a Reduction of Salmonella in Raw Meat and Poultry. Online. Available at http://www.usda.gov/news/releases/2002/04/0154.htm. Accessed June 11, 2002. Veterinary Services. 2001a. Info Sheet: Escherichia cold 0157 in United States Feedlots. Online. APHIS, USDA. Available at http://www.aphis.usda.gov/vs/ceah/cahm/Beef_Feedlot/coli.PDF. Accessed May 14, 2002. Veterinary Services. 2001b. Info Sheet: Salmonella in United States Feedlots. Online. APHIS, USDA. Available at http://www.aphis.usda.gov/vs/ceah/cahm/Beef_feedlot/99sal.PDF. Accessed May 14, 2002. Waldroup AL, Rathgeber BM, Forsythe RH. 1992. Effects of six modifications on the incidence and levels of spoilage and pathogenic organisms on commercially processed postchill broilers. J Appl Poult Res 2:111-116. Wallace RL. 2003. Market Cows: A Potential Profit Center. mini Dairy Net. Online. Available at http://traill.outreach.uiuc.edu/dairynet/paperDisplay.cfm?DyanBizInfoID=984. Accessed Janu- ary 15, 2003. Wang G. Zhao T. Doyle MP. 1996. Fate of enterohemorrhagic Escherichia cold 0157:H7 in bovine feces. Appl Environ Microbiol 62:2567-2570. Wells J. Wagner A, Dargatz A. 1999. Factors associated with Johne's disease on U.S. dairy opera- tions. In: Manning EJB, Collins MT, eds. Proceedings of the 6th International Collegium of Paratuberculosis. Madison, WI: International Association for Paratuberculosis. P. 13. Wells SJ, Fedorka-Cray PJ, Russell R. Besser T. McDonough P. Smith B. 1998. E. cold 0157 and Salmonella—Status on U.S. Dairy Operations. Center for Epidemiology and Animal Health, Veterinary Service, APHIS, USDA. Online. Available at http://www.aphis.usda.gov/vs/ceah/ cahm/Dairy_Cattle/ecosalm98.htm. Accessed July 19, 2002. WI DATCP (Wisconsin Department of Agriculture, Trade and Consumer Protection). 2002. Wisconsin's Meat Safety & Inspection Program: Analysis and Evaluation. Madison, WI: WI DATCP. Wierup M, Engstrom B. Engwall A, Wahlstrom H. 1995. Control of Salmonella Enteritidis in Swe- den. Int J Food Microbiol 25:219-226.

Next: 5 Scientific Criteria and Performance Standards to Control Hazards in Seafood »
Scientific Criteria to Ensure Safe Food Get This Book
×
Buy Paperback | $68.00 Buy Ebook | $54.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Food safety regulators face a daunting task: crafting food safety performance standards and systems that continue in the tradition of using the best available science to protect the health of the American public, while working within an increasingly antiquated and fragmented regulatory framework. Current food safety standards have been set over a period of years and under diverse circumstances, based on a host of scientific, legal, and practical constraints.

Scientific Criteria to Ensure Safe Food lays the groundwork for creating new regulations that are consistent, reliable, and ensure the best protection for the health of American consumers. This book addresses the biggest concerns in food safety—including microbial disease surveillance plans, tools for establishing food safety criteria, and issues specific to meat, dairy, poultry, seafood, and produce. It provides a candid analysis of the problems with the current system, and outlines the major components of the task at hand: creating workable, streamlined food safety standards and practices.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!