Author
 |
Tamplin, Mark |
|
Submitted to: Meeting Abstract
Publication Type: Abstract Only Publication Acceptance Date: 9/10/2003 Publication Date: 11/8/2003 Citation: Tamplin, M.L. 2003. Strategies for applying predictive microbial models to diverse foods. U.S. Japan Natural Resources Panel. Abstract p. 105-109. Interpretive Summary: Technical Abstract: Historically, the majority of microbial models for predicting the behavior of foodborne pathogens were developed in microbiological broth systems. Currently, there is a high level of interest in applying models to various food processing unit operations in an effort to meet regulatory performance standards and to assist in the development of risk assessment. When extrapolating a model to other food matrices, it is necessary to know the bias and accuracy of the model's predictions. There are a number of intrinsic and extrinsic factors of food that can influence the robustness of models. For example, the predictive power of a model that was developed in a pure-culture, sterile broth will not likely make accurate predictions of the bacterial growth rate and maximum population density in foods that contain spoilage flora. In addition, broth-based model predictions lack accuracy at growth/no-growth boundaries. Therefore, the factors that influence the performance of a model in various food environments must be understood to define the uncertainty for different applications and to design potential approaches to adjust model parameters. If this is done, then the model may have greater predictive power and ultimately more value to the end-user. To illustrate these conditions, this presentation compares broth model predictions for Listeria monocytogenes and Escherichia coli O157:H7 with the observed growth of these pathogens in cooked, cured ham and raw ground beef, respectively. In addition, strategies for optimizing the use of models to meet microbial performance standards will be discussed. |