Zero-inflated hierarchical models for faecal egg counts to assess anthelmintic efficacy

Area of Science:

• Veterinary Parasitology
• Statistical Modeling
• Livestock Health

Background:

• Anthelmintic resistance is a growing problem in livestock due to widespread drug use.
• Current methods for detecting resistance, like fecal egg count reduction tests, have limitations.
• These limitations include ignoring counting variability and extra zero counts.

Purpose of the Study:

• To propose and evaluate zero-inflated Bayesian hierarchical models for estimating fecal egg count reduction.
• To address the shortcomings of conventional methods in detecting anthelmintic resistance.
• To provide a more robust statistical approach for monitoring drug resistance in parasitic worms.

Main Methods:

• Development of zero-inflated Bayesian hierarchical models.
• Comparison of Bayesian models with conventional fecal egg count reduction tests, bootstrap, and quasi-Poisson regression via simulation.
• Application of the proposed model to a case study of anthelmintic resistance in Swedish sheep flocks.

Main Results:

• The proposed Bayesian models demonstrated superior robustness compared to conventional methods.
• Bayesian models performed well in terms of reducing bias and improving coverage in estimations.
• The case study illustrated the practical advantages of the Bayesian approach for real-world resistance monitoring.

Conclusions:

• Zero-inflated Bayesian hierarchical models offer a more accurate and reliable method for estimating fecal egg count reduction.
• These models effectively account for counting variability and extra zero counts, leading to better detection of anthelmintic resistance.
• The findings support the adoption of these advanced statistical models for improved livestock parasite control strategies.