Effect of Production Facility Factors on the Prevalence of Salmonella in Small and Very Small Meat Processing Plants

NON-TECHNICAL SUMMARY: A. The unique challenges of smaller meat processors heighten the need for a detailed investigation to both describe the production facility characteristics and assess what characteristics are associated with bacterial contamination of meat products. A. This study aims to evaluate the characteristics of small meat processors to establish sources and currently used control measures for Salmonella in meat and poultry products. Correlation between these characteristics and the plants' microbial tests will then evaluate which characteristics are effective preventive measures and which are risk factors for positive test results.

<P>
APPROACH: Investigators will compose a questionnaire designed to determine processing plants characteristics. The topics to be encompassed include: physical features, sanitation practices, good manufacturing practices, sampling program and methods, HACCP plans, measurement practices, and range of processing and methods. Once the instrument is developed, it will be tested with the assistance of the Center for Survey Research at The Ohio State University. Then, a case-control study of small and very small meat processors will be designed. Approximately 45 plants will be randomly chosen as cases. We will survey 4 control plants per every case plant. The control plants will be matched to the case plant based on 2 criteria: FSIS inspection district and size of the establishment. We will survey approximately 45 processing plants which have failed a Salmonella test set in 2006 and 180 matching plants that did not fail any sets in 2006. A phone interview will be arranged with a member of each establishments HACCP team. To prevent bias, the interviewer will be blinded as to the case or control status of the processor. The microbial test results will come from a source outside of the plant, eliminating bias stemming from self-reported results. During the study, all information will be blinded by establishment number, and kept confidential by the investigators. All published characteristics and practices will be presented as general information and not tied to any establishment; no establishments will be recognized individually. The association of production facility factors with failed Salmonella test sets will be modeled. Using logistic regression, the dependant variable, Salmonella test failure, will be coded 1 and 0 for cases and controls respectively. A step-wise model building process will be used. Univariate logistic regression models will be constructed with each of the questions or subject areas of the survey. The p-value of entry is 0.25, with the independent variables being added in from the lowest p-value. Meaningful interactions will be investigated and all independent variables of significance will be reported. The investigators will establish a reference work for small and very small meat processors. The report will identify risk factors and preventive measures for Salmonella. An electronic version will be made available on the Ohio State University Meat Science website. The report will be printed and sent to the processors that participated in the study, and a letter will be sent to all other small and very small processors as a notice to the availability. Finally, the material will be incorporated into the extension short courses that Dr. Knipe offers to meat processors and his formal college class. As with any program designed to assist a group of people there needs to be an assessment to evaluate the validity and usefulness of the program. We will evaluate the results achieved, practice change, knowledge, attitude, and skills (KAS) change, reactions recipients involved, activities performed and the inputs made.

<P>

PROGRESS: 2007/09 TO 2008/09<br>
OUTPUTS: Using results from the FSIS 2005 Salmonella test sets, a case-control study was designed. The small and very small plants that failed test set A were selected as cases (n=32) and controls were matched to the case plants by inspection district and size, 4:1. Control plants had completed and passed the A set tests. A survey instrument was created in 3 parts: General Hazard Analysis and Critical Control Points (HACCP), Slaughter, and Raw Product Processing, and mailed to the meat processors. These processors were queried regarding their processing and food safety methods, in an attempt to identify trends relative to methods used by those who had passed and those who had failed their Salmonella test sets. Data was analyzed, results were compiled and conclusions made in the Ph.D. dissertation of Mary Kay Folk. <br>PARTICIPANTS: Mary Kay Folk, graduate student who conducted the survey and analyzed the data. C. Lynn Knipe, student's advisor, supervised the development of the survey instrument, conduction of the survey, and analysis of the data. Julie A. Funk, Michigan State University, supervised the development of the survey instrument, conduction of the survey, and analysis of the data. <br>TARGET AUDIENCES: Owners or technical service employees of small and very small meat processors involved with raw processing of beef, pork or poultry. <br><br>
IMPACT: 2007/09 TO 2008/09<br>
The survey response rates were 40% and 38% for case and control plants, respectively. Other than variables representing plant size, such as numbers of employees and volume of production, there were few significant differences between small and very small respondents. Differences between cases and controls were found in animal washing before slaughter, type of poultry evisceration and percentage of raw product from in-house slaughter. Most of the plants (71%) operated under 2 or 3 processing categories; with the majority of the plants processing 10 or more products (60%). Seventy-six percent processed raw products daily. Only 34% of the respondents slaughtered animals with 57% of those slaughtering daily. Gloves were worn during processing in 88% of the plants that processed raw products; however, almost a quarter of those reported no policy requiring glove use. Ninety-five percent of plants reporting woven glove use have policies to launder or dispose of gloves. About 36% of the plants had Sanitation Standard Operating Procedures or Good Manufacturing Practices in place to specifically address Salmonella. Almost 28% have determined Salmonella contamination as a hazard likely to occur in their processes. Consistent use and knowledge of sanitation protocols were lacking in many cases. Additional details from these plants could provide more useful information for Salmonella control in smaller processing facilities. Results from this research will help focus and expand specific Extension programs for small and very small meat processors.
<br><br>PROGRESS: 2006/09/15 TO 2007/09/14<br>
There is a critical need for evaluation of production facility factors that affect contamination of meat products especially in plants with less than 500 employees. The unique challenges of these plants, such as lack of technical support, financial and physical flexibility, and highly variable operations, heighten the need for a detailed investigation to describe the production facility characteristics and assess what characteristics are associated with bacterial contamination of meat products. Specifically, this study will be comparing these plant characteristics with FSIS' Salmonella test set results. Raw meat processing plants are tested for Salmonella as an indication of cleanliness and adequate sanitation. In regard to food safety, according to CDC data, Salmonellosis is the only foodborne illness that has not shown a decreasing trend in the last decade. The first objective of this project is to create a survey instrument that will elicit descriptive data regarding the production facilities of small and very small meat processing plants in the US. To this point in the project the instrument has been created and tested using a small group of local meat processors. These processors have worked with us frequently in the past, and gave input not only about the questions asked, but the wording, the subjects and topics that were missing. Since the survey was finalized, it is being sent to the sample set in waves. The set was chosen based on 2005 Salmonella test set results. All 32 plants that failed their first Salmonella test set are "cases" in the sample. For each of those cases 4 control plants were chosen at random from the same size plants in the same inspection district. At this point 104 surveys have been sent to the sample set. Phone contact has been made with all plants that have working phone numbers, and current response rate is 29 percent. <br><br>

IMPACT: 2006/09/15 TO 2007/09/14<br>
No outcomes or impacts have resulted yet. The project is still in the data gathering phase.