Human salmonellosis: estimation of dose-illness from outbreak data

Area of Science:

• Microbiology
• Epidemiology
• Quantitative Risk Assessment

Background:

• Accurate dose-response models are vital for quantitative risk assessment.
• Human feeding trials have limitations, often excluding low doses and specific populations.
• Epidemiological data offers value but faces challenges with uncertainty.

Purpose of the Study:

• To model the dose-illness relationship using Salmonella outbreak data.
• To investigate differences in host susceptibility and pathogen-food matrix interactions.
• To develop robust models accounting for variability and uncertainty.

Main Methods:

• Utilized generalized linear mixed models and fractional polynomials for dose-illness modeling.
• Modified fractional polynomial models to align with established dose-illness properties.
• Incorporated host susceptibility as fixed effects and pathogen/food matrix as random effects.
• Employed bootstrap procedures to address stochastic variability and data uncertainty.

Main Results:

• Identified higher illness probability in susceptible populations at low doses for virulent pathogen-food combinations.
• Observed increased illness probability in susceptible populations at high doses for less virulent combinations.
• Indicated the presence of immunity in the normal population, contrasting with the susceptible group.

Conclusions:

• Host susceptibility significantly influences illness probability across varying dose levels and pathogen virulence.
• Immunity appears to be a factor in the general population but not in susceptible individuals.
• The developed models provide a more nuanced understanding of microbial risk assessment.