A Bayesian Hierarchical Model for Assessing the Reliability of Public Health Surveillance Systems in Senegal: A Methodological Evaluation, 2000–2026

{ "background": "Public health surveillance systems are critical for disease control, yet their reliability is often uncertain. In many settings, including Senegal, methodological frameworks for quantifying this reliability and its spatial-temporal variation are lacking, hindering evidence-based system strengthening.", "purpose and objectives": "This study aimed to develop and evaluate a novel Bayesian hierarchical model to quantify the reliability of public health surveillance systems, with a specific application to Senegal. The objective was to provide a robust methodological tool for identifying systematic under-reporting and spatial heterogeneity in system performance.", "methodology": "We developed a Bayesian hierarchical model integrating case report data with latent true incidence. The core model structure is yit \~ (\λit \· \ρi), where yit are observed cases in region i and time t, \λit is the latent true incidence, and \ρi is the region-specific reporting reliability. We fitted the model using Markov chain Monte Carlo simulation with data from multiple disease programmes.", "findings": "The model identified substantial spatial heterogeneity in system reliability, with regional reporting probabilities (\) ranging from 0.35 to 0.92 (posterior median). A key finding was that approximately 40% of regions had a posterior probability greater than 0.9 that their true reliability was below the national target threshold of 0.8.", "conclusion": "The proposed model provides a statistically robust framework for evaluating surveillance system performance, moving beyond descriptive metrics to a probabilistic assessment of reliability. It successfully quantified significant and previously unmeasured spatial disparities in reporting completeness within the country.", "recommendations": "Implement the model as a routine analytical tool within the national surveillance division to prioritise regions for system investment. Future research should integrate socioeconomic covariates to explain the observed heterogeneity in \.", "key words": "surveillance evaluation, Bayesian statistics, hierarchical modelling, health systems, disease