A Bayesian hierarchical approach to identify chronic health management failures in swine holdings: Evidence from a medium-throughput italian slaughterhouse.

This study presents a risk-based surveillance framework using routine post-mortem inspection data to identify systemic health management failures in swine holdings. Analysing 339,643 slaughter records from 653 holdings across six production categories, the study assessed seven Key Performance Indicators (KPIs): Whole Carcass Condemnation (WCC), abscess, neck abscesses, liver ascaridiosis, heart pericarditis, pluck polyserositis, and pluck adhesions. To account for overdispersion inherent in slaughter inspection data, a Hierarchical Bayesian Beta-Binomial (HBB) model was applied independently for each KPI and production category, categorising holdings into four risk tiers by comparing posterior prevalence estimates against a category-specific 90th-percentile benchmark. Results demonstrated substantial pathological variation across production categories. The WCC rate in gilts (32.04‰) was 42 times higher than in heavy pigs (0.76‰). Overall, 16.2% of holdings exhibited chronic health failures (Tier G1) in at least one production category, and 3.5% were systemic outliers with chronic failures recurring across multiple production categories simultaneously. Liver ascaridiosis was the most prevalent biosecurity failure in finishing categories, reaching 135.02‰ in hogs. Conversely, welfare failures (abscess) and injection-site lesions (neck abscesses) showed substantially higher condemnation rates in breeding and rearing cohorts, with rate ratios of 3.2 and 3.5 relative to finishing categories respectively. This Bayesian framework robustly distinguishes chronic systemic management failures from sporadic or category-specific findings, enabling targeted veterinary audits and biosecurity interventions for high-risk production units and supporting the use of slaughterhouse data as a proactive herd health surveillance tool.