Integrating uncertainty into official control decisions: Bayesian risk-based decision support for the safe reopening of live bivalve mollusk harvesting areas.

The safety of live bivalve mollusks (LBM) depends on complex interactions between pathogen ecology, environmental variability, and regulatory monitoring outcomes. Conventional compliance-based reopening decisions for classified harvesting areas often rely on limited sampling data and deterministic thresholds, without formally accounting for prior scientific evidence or uncertainty. This study proposes a Bayesian decision-support framework that integrates peer-reviewed evidence on pathogen occurrence, accumulation dynamics, and persistence in LBM into a probabilistic regulatory model for area classification and reopening decisions. A structured literature review was conducted to identify quantitative evidence describing the prevalence and behavior of bacterial, viral, and protozoan hazards in bivalve mollusks, including Escherichia coli, Salmonella spp., norovirus, hepatitis A virus, and protozoan parasites. This evidence was translated into literature-informed prior distributions representing baseline contamination probability and environmental risk conditions. A conjugate Bayesian β-binomial model was applied to integrate official monitoring data with prior knowledge, generating posterior distributions of the probability that microbial contamination exceeds acceptable thresholds. Decision criteria were established according to precautionary regulatory tolerance levels, expressed as the posterior probability of unacceptable contamination. Numerical case studies and Monte Carlo simulations demonstrated how the proposed framework supports transparent and risk-based regulatory decisions: scenarios included reopening under low residual risk, persistent closure under high posterior probability of exceedance and borderline situations requiring additional sampling. Integration of environmental evidence and pathogen-specific persistence (notably viral contamination with slow depuration dynamics) substantially influenced posterior risk estimates and decision outcomes. The results demonstrate that identical monitoring outcomes may lead to different regulatory decisions depending on prior ecological evidence and uncertainty quantification. The proposed Bayesian decision framework integrating ecological evidence provides a scientifically robust and transparent tool for competent authorities managing LBM harvesting areas. By formally incorporating prior scientific knowledge, environmental indicators, and uncertainty, the framework operationalizes the precautionary principle while enabling adaptive, evidence-based reopening decisions consistent with European Union food safety legislation.