Formulation of a Dose-Response Equation for Human Illness from Salmonella Enteritidis. R. M. McDowell, Risk Analysis Systems, USDA-APHIS, Riverdale, MD; and A. R. Baker, Epidemiology and Risk Assessment, USDA-FSIS, Washington, DC
The final module of the farm-to-table risk assessment of Salmonella enteritidis (SE) in eggs in USA computes the numbers of persons who become ill from ingesting specific numbers of SE bacteria contained in foods prepared from internally-contaminated eggs. The dose-response function to calculate probability of illness from such exposures was developed indirectly because no human feeding trials have been conducted with Salmonella enteritidis. In addition, the published literature catalogs the shortcomings of human feeding trial data (high minimum dose, small sample size, nature of test subjects, etc.) and the problems associated with extrapolating inferences from such data to large populations. To select an appropriate model we examined (1) dose-response studies in which feeding trial data for a variety of Salmonella serovars and other pathogens were examined and a variety of function distributions fitted to those data, and (2) epidemiologic data from foodborne disease outbreaks. The epidemiologic data—published attack rates and ingested doses in foodborne disease outbreaks from a number of Salmonella serovars—was found to be inconsistent with the dose-response functions estimated from feeding trial data using Salmonella serovars. Therefore, a surrogate pathogen/dose-response model consistent with the epidemiologic data was selected. The substantial discontinuity between the epidemiologic data and the dose-response functions estimated from feeding trial data requires further investigation before risk assessments based on data from feeding trials can be confidently used to predict human illness.
Go to . . .