Bayesian methods for estimating pathogen prevalence within groups of animals from faecal-pat sampling.

Area of Science:

• Veterinary epidemiology
• Quantitative microbial risk assessment
• Bayesian statistics

Background:

• Foodborne pathogens like Escherichia coli O157:H7 and Campylobacter spp. are linked to food poisoning outbreaks.
• Domestic animals and wildlife act as reservoirs for these pathogens, with fecal contamination being a key transmission route.
• Accurate quantitative microbial risk models require reliable data throughout the food chain, particularly at the farm level.

Purpose of the Study:

• To develop and illustrate a Bayesian method for estimating animal pathogen prevalence from fecal samples.
• To address sampling challenges at the farm level for improved food chain risk assessment.
• To investigate sample size determination for accurate prevalence estimation.

Main Methods:

• Employed a Bayesian statistical approach to estimate animal-pathogen prevalence from fecal samples.
• Incorporated prior beliefs and accounted for uncertainties in animal-level prevalence estimates.
• Illustrated the Bayesian technique with examples considering various assumptions and conditions.

Main Results:

• The Bayesian method provides a flexible and robust alternative to classical approaches for prevalence estimation.
• This technique accommodates complexities such as unequal defecation rates, uncertain population sizes, and imperfect microbiological test sensitivity.
• The study also explored sample size requirements for achieving desired accuracy in prevalence determination.

Conclusions:

• The developed Bayesian method offers enhanced flexibility and robustness for estimating animal pathogen prevalence.
• This approach is valuable for parameterizing microbial risk models, especially when dealing with complex biological and testing scenarios.
• Accurate farm-level data, derived through advanced statistical methods, is essential for safeguarding food safety.