Estimating epidemiological parameters for bovine tuberculosis in British cattle using a Bayesian partial-likelihood...

Area of Science:

• Veterinary epidemiology
• Infectious disease modeling
• Bayesian statistics

Background:

• Bovine tuberculosis (bTB) poses significant challenges in cattle, exacerbated by wildlife reservoirs like badgers.
• Accurate epidemiological models are crucial for understanding and controlling bTB spread.
• Bayesian likelihood-based inference is increasingly vital for infectious disease epidemiology.

Purpose of the Study:

• To evaluate nested dynamic models of bTB transmission in British cattle.
• To infer transmission and diagnostic test parameters with minimal prior knowledge.
• To assess herd-level transmission heterogeneity and its impact on bTB dynamics.

Main Methods:

• Employed likelihood-based bootstrapping for parameter inference.
• Utilized summary data of cattle testing positive at herd outbreak onset.
• Analyzed high- and low-risk areas separately, considering nested dynamic models.
• Assessed models with and without herd heterogeneity.

Main Results:

• Models without herd heterogeneity were preferred in both high- and low-risk areas.
• Evidence suggests the presence of super-spreading cattle within herds.
• Low diagnostic test sensitivities and high within-herd basic reproduction numbers (R0) were observed.
• Current testing regimes are largely sufficient for controlling within-herd epidemics.

Conclusions:

• Summary data are effective for parameter inference in bTB models, offering an attractive alternative to data-heavy approaches.
• The findings highlight the potential for numerous unobserved infections in cattle populations.
• The modeling approach is adaptable for studying other infectious disease systems.