Purpose: Develop and validate predictive models for growth of Salmonellaspp. and non-O157 STEC in targeted food matrices.
Methods: Ground beef and shredded iceberg lettuce were inoculated with a six-serotype STEC cocktail (O26, O45, O103, O111, O121 and O145). A five-serovar Salmonellaspp. cocktail was used to inoculate reconstituted non-fat dry milk (NDM) and lettuce. Isothermal growth data was collected for each of the pathogen cocktails at various temperatures (5.0-47.5°C). A Baranyi model was used to fit the primary model. A modified Ratkowsky model was used to generate the secondary model. Two sinusoidal temperature profiles (5-15°C and 10-40°C) were used to validate the dynamic models. Mean absolute relative error (MARE) was used to judge the accuracy of the models, with 0% MARE indicating best fit.
Results: MARE values for the ground beef model for low and high temperature non-O157 STEC profiles were 8.5 and 1.7%, respectively. MARE values for growth of non-O157 STEC in lettuce were 26.7 and 4.8% for the low and high temperature profiles, respectively. Similarly, the dynamic model predicted the growth of Salmonella spp. on lettuce with MARE values of 5.4 and 6.3%, respectively, for high and low temperatures. MARE values for Salmonellain NDM were 5.8 and 6.6% for high and low temperature, respectively.
Significance: The dynamic models for both pathogens and the foods they were evaluated in resulted in low MARE% values (1.7-8.5%) at both storage temperature ranges, indicating acceptable model accuracy. MARE% for non-O157 STEC in iceberg lettuce stored at low temperatures (26.7% MARE) indicated a lack of fit of the model and a need for additional research.