A Bayesian Hierarchical Model for Assessing the Reliability of Public Health Surveillance Systems in Uganda, 2000–2026

"background": "Public health surveillance systems are critical for disease control, yet their reliability in low-resource settings is often uncertain. Existing evaluation frameworks lack robust quantitative methods to integrate heterogeneous data sources and account for spatial and temporal dependencies in system performance. ", "purpose and objectives": "This study aimed to develop and apply a novel Bayesian hierarchical model to quantitatively assess the reliability of national public health surveillance, defined as the probability of correctly detecting and reporting a notifiable disease event. ", "methodology": "We developed a Bayesian latent variable model integrating case notification data, laboratory confirmation reports, and health facility survey data. The core reliability metric is modelled as () = \ + \ X{it + ui + \, where is the reliability for district i in period t, Xit are covariates, ui are spatial random effects, and \ are temporal effects. Inference used Markov chain Monte Carlo sampling. ", "findings": "Posterior estimates revealed substantial spatial heterogeneity in system reliability, with a median national reliability of 0. 72 (95% credible interval: 0. 68–0. 76). Reliability was strongly associated with health facility density (posterior probability > 0. 95) and exhibited a declining temporal trend in northern regions. ", "conclusion": "The proposed model provides a rigorous, data-driven tool for quantifying surveillance reliability, revealing significant and spatially structured gaps in performance. ", "recommendations": "Resource allocation for surveillance strengthening should be prioritised based on quantitative reliability estimates. The model should be integrated into routine performance reviews to enable targeted interventions. ", "key words": "Bayesian inference, disease surveillance, health systems, hierarchical model, latent variable, reliability, sub-Saharan Africa", "contribution statement": "This paper introduces a novel Bayesian latent variable framework for quantifying public health surveillance reliability, providing the first spatially explicit, probabilistic assessment for a national system in