This Risk Assessment is a supporting document of Federal Register Notice Docket No. FSIS-2019-0023.

The Food Safety and Inspection Service's (FSIS) Strategic Plan for 2017-2021 stated that the Agency would develop Salmonella performance standards for raw pork products. This risk assessment report describes the data and analytical methods used in the development of Salmonella performance standards for raw comminuted pork and both raw intact and nonintact pork cuts.

Based on analysis of U.S. foodborne disease outbreaks, pork may be responsible for between 8 and 13 percent of roughly 1 million foodborne human salmonellosis cases each year. In the United States, the majority of pork is consumed as pork cuts (e.g., chops, roasts), with the remainder consumed as comminuted pork (e.g., sausage) and ready-to-eat products such as cooked ham.

FSIS inspects 1,110 establishments that slaughter hogs, 1,070 establishments that produce pork cuts, and 1,334 establishments producing comminuted pork. After considering the occurrence of Salmonella contamination across these industries and various resource constraints, FSIS decided to apply performance standards to establishments that produce more than 6,000 pounds of comminuted pork per day (138 establishments representing 96 percent of production); or 50,000 pounds of pork cuts per day (38 establishments representing more than 91 percent of production). Most slaughter establishments are not subject to the performance standards, but there are 45 establishments covered by these standards that account for 88 percent of all swine slaughtered in the United States.

Sampling evidence shows that Salmonella prevalence varies substantially across the pork industries. On average, 30.3 percent of comminuted pork samples were Salmonella-positive. For pork cuts samples, an average of 9.3 percent were Salmonella-positive. Contamination tends to cluster among a limited number of establishments in this industry, particularly for pork cuts.

This risk assessment uses a published model to evaluate a range of performance standard options. Each option is defined by the maximum number of allowable Salmonella-positive samples among 52 samples collected per 12-month period. For each putative performance standard, the model predicts a) the reduction in annual cases of human salmonellosis and b) the share of the respective industry that will pass the performance standard initially. For a targeted 25 percent reduction (per Healthy People 2030) in human salmonellosis cases attributed to these products, performance standards less than, or equal to, 13 and 6 provide increasing confidence in meeting this goal for comminuted pork and pork cuts, respectively.

For a comminuted pork performance standard of 13 allowable positives, the median (95 percent confidence interval) prediction is that about 8,300 (3,600 - 16,300) fewer Salmonella illnesses associated with this product will occur annually following full implementation. Also, an estimated 51 (37 - 67) percent of the production volume - and 56 (49 - 65) percent ofestablishments - will pass this performance standard initially. Following implementation of a pork cuts performance standard of 6 allowable positives, 21,600 (10,000 - 40,000) fewer Salmonella illnesses associated with these products are predicted, while an estimated 67 (61 - 74) percent of the production volume - and 61 (56 - 67) percent of establishments - will pass this performance standard initially. Of the 20 establishments that produce both pork products and are subject to these performance standards, the model predicts that about 11 will pass both standards, 2 will fail both standards and the remaining 7 will fail one standard.

The risk assessment predictions will hold only if certain assumptions are true. For example, sampling of establishments must be sufficient to classify establishments accurately; and, to generate incentives for passing the standards, information about the classification of individual establishments with respect to the performance standards must be available publicly. Following implementation of the standards, FSIS will periodically analyze the evidence it gathers to determine the accuracy of its predictions.

(Note: Download the PDF under "Related Media" for the full document)