Dose-response-models and their implications for quantitative risk assessment for Campylobacter infections

For some time now there have been several projects dealing with the assessment of campylobacteriosis risks for consumers. Dose-response relationships form a crucial part of such assessments, as they specify disease probabilities depending on different microbial concentrations in foods. Evaluation of such models, however, is difficult because of problems to find data on which reliable assumptions could be based. Ongoing risk analyses for Campylobacter mainly refer to a single administration study with human volunteers published by Black et al. (1988). However, whether results from this study can be transferred to target populations envisaged in risk assessments remains questionable for several reasons. In this paper some alternative dose-response models, their fit to the data of Black et al., and risk estimates resulting in a fictitious scenario are discussed and compared. Depending on the dose-response model assumed risk estimates can differ remarkably. Therefore it is hardly possible to make reliable quantifications of risks in reality, however, it can be determined how much they may vary assuming different scenarios.