A Theoretical Framework for the Use and Interpretation of Molecular Subtyping in Foodborne Disease Surveillance.

The development of molecular subtyping for foodborne disease surveillance has significantly enhanced our ability to detect, investigate, and control common source outbreaks. Despite wide-scale use and many high profile outbreaks detected through enhanced foodborne disease surveillance, use and interpretation of molecular subtype data has been primarily intuitive. Efforts to develop interpretive guidelines have been based primarily on laboratory criteria with the relatively narrow goal of defining the likelihood of relatedness between strains. However, the significance of strain relatedness is context dependent, and additional factors are required for epidemiological inference. These factors have not been clearly defined, which has limited our ability to consistently interpret findings and effectively utilize the new tools. In this work, a theoretical framework was developed to describe the interrelationship of case definitions based on molecular or other classifications, the normal prevalence of disease or strain, frequency of exposure, specificity of exposure information, and number of cases in a common source outbreak setting. A computer model was developed to illustrate the impact of increasing case definition specificity and exposure prevalence on the association between illness and exposure. The model demonstrated that as the case definition in a case control study became more specific, measures of association increase and confidence intervals widened. Low prevalence exposures required less case definition specificity or fewer cases to reach significant association than high prevalence exposures. The model demonstrated that the optimum level of molecular subtype specificity is not fixed, but depended on the prevalence of the exposure, specificity of the exposure information, the number of available cases, and the question being asked. Furthermore, the sensitivity of exposure identification and the specificity of the exposure information have an impact on the ability to detect and resolve common source outbreaks in much the same manner as molecular subtyping. Using the key parameters identified, the inherent benefits and limitations of each of the currently available surveillance systems were compared for their ability to detect problems in the food supply. Interpretive guidelines for molecular subtype data were developed, and a practical guide to best practices was formulated.