The purpose of this review was to determine if the FSIS Risk Assessment for Listeria monocytogenes (LM) in Deli Meats model worked as described and to examine the impact of alternative model input assumptions on model calibration, intervention options and conclusions.
In general, the FSIS model works as described in the FSIS report. The formulas used to model the mass balance approach are correctly implemented. The distribution used in the calibration to represent listeria concentrations in deli meats at retail correctly simulates the data in FDA/FSIS risk assessment. The number of iterations used in the risk assessment (1,000,000 iterations) is sufficient for the model output to stabilize. However, the distribution used by FSIS to represent the amount of listeria added during a contamination event is not necessarily the distribution that resulted in the best fit when compared to that based on the data in FDA/FSIS risk assessment. Correlation between the duration of a contamination event, the interval between contamination events, or the number of Listeria organisms transferred to the FCS is not allowed in the FSIS in-plant model.
Alternative parametric distribution for specific variables may be needed, or there may be other model construct limitations. FSIS’s conclusions about the relative effectiveness of various intervention scenarios remain questionable.