Hierarchical dose response of E. coli O157:H7 from human outbreaks incorporating heterogeneity in exposure

SUMMARY The infectivity of pathogenic microorganisms is a key factor in the transmission of an infectious disease in a susceptible population. Microbial infectivity is generally estimated from dose–response studies in human volunteers. This can only be done with mildly pathogenic organisms. Here a hierarchical Beta-Poisson dose–response model is developed utilizing data from human outbreaks. On the lowest level each outbreak is modelled separately and these are then combined at a second level to produce a group dose–response relation. The distribution of foodborne pathogens often shows strong heterogeneity and this is incorporated by introducing an additional parameter to the dose–response model, accounting for the degree of overdispersion relative to Poisson distribution. It was found that heterogeneity considerably influences the shape of the dose–response relationship and increases uncertainty in predicted risk. This uncertainty is greater than previously reported surrogate and outbreak models using a single level of analysis. Monte Carlo parameter samples (a, b of the Beta-Poisson model) can be readily incorporated in risk assessment models built using tools such as S-plus and @Risk.