Purpose: The objective was to quantitatively evaluate the fit of primary and secondary models on the survival/inactivation kinetics of Salmonella in multiple low-moisture foods.
Methods: Isothermal and iso-moisture survival/inactivation data for Salmonella in low-moisture foods (> 1,000 data series; 8 different products; temperatures -20 - 120°C; aw 0.11 - 0.95) were collected from prior studies and categorized. Log-linear and Weibull primary models were fit to single and grouped replicate data series and were statistically compared using the Akaike Information Criteria (AIC). Secondary models describing the effects of temperature and aw were similarly analyzed.
Results: Based on AIC, primary regression on single data series resulted in a 95% likelihood of the log-linear and Weibull models being the correct model in ~1 and 67% of the cases, respectively. Similarly, primary regression on grouped replicate series resulted in a 95% likelihood of the log-linear and Weibull models being the correct model in ~0.5 and 77% of the cases, respectively. The relative frequency of the log-linear model as the more likely correct model peaked in the 20 - 30°C or 0.90 - 0.95 temperature and aw ranges, respectively.
Significance: Based on the primary model regression results, the Weibull model best described the survival/inactivation kinetics of Salmonella during isothermal and iso-moisture conditions. However, careful consideration is required for the development or application of secondary models for dynamic temperature or moisture processes.