Engineering-based Probabilistic Risk Assessment for Food Safety with Application to Escherichia Coli O157:h7 Contamination in Cheese

Title of Dissertation: ENGINEERING-BASED PROBABILISTIC RISK ASSESSMENT FOR FOOD SAFETY WITH APPLICATION TO ESCHERICHIA COLI O157:H7 CONTAMINATION IN CHEESE Kristin Ann Fretz, Doctor of Philosophy, 2006 Dissertation directed by: Professor Mohammad Modarres Department of Mechanical Engineering A new methodology is introduced in which engineering-based tools and techniques are adapted to quantitative microbial risk assessment (QMRA) in order to offer a more systematic solution to food safety problems. By integrating available microbial data and adapted engineering techniques within the traditional QMRA framework, this new methodology addresses some of the deficiencies of traditional approaches. Through the use of a hierarchical structure, the system is decomposed into its most basic elements so that the interrelationships and interdependences of these basic elements are captured. This hierarchical structure also identifies variability throughout the process, resulting in a risk model in which multiple scenarios can be analyzed. In addition, the engineering approach adapts methods for characterizing and propagating uncertainties. Unlike the traditional approaches in food safety, the engineering-based methodology relies on mathematical models; the uncertainties about these models (both aleatory and epistemic), as well as the uncertainties about the model parameters, are formally quantified and properly considered. This separation and characterization of uncertainties results in a more powerful risk model, so that assessments can be made as to whether additional information or changes to the physical system will reduce the total uncertainty. Finally, this research characterizes the validity of the various dose-response models. Comparison of actual outbreak observations to model predictions lends credibility and assesses uncertainty of the developed dose-response models. Thus, the results of the risk model can be used both as an absolute assessment of risk and as a relative measurement of mitigation and control strategies. As a case study, the engineering-based methodology is applied to the problem of Escherichia coli O157:H7 contamination in cheese. While it has been assumed that pathogenic microorganisms in raw milk die during cheese-making, several studies on the survival of E. coli O157:H7 in cheese have demonstrated growth during cheese manufacturing. Furthermore, E. coli O157:H7 has been linked to several outbreaks involving cheese, thereby establishing the need to investigate this route of transmission. The successful application of the engineering-based approach to the problem of E. coli O157:H7 contamination in cheese suggests that this new methodology can be applied to other food safety problems. ENGINEERING-BASED PROBABILISTIC RISK ASSESSMENT FOR FOOD SAFETY WITH APPLICATION TO ESCHERICHIA COLI O157:H7 CONTAMINATION IN CHEESE