Journal Design Clinical Emerald
African Food Systems Research (Interdisciplinary - incl Agri/Env) | 27 January 2004

A Bayesian Hierarchical Model for Assessing the Reliability of Public Health Surveillance Systems in Senegal

A Methodological Evaluation, 2000–2026
A, m, i, n, a, t, a, D, i, o, p, ,, F, a, t, o, u, S, a, r, r, ,, M, o, u, s, s, a, N, d, i, a, y, e
Bayesian modellingsurveillance reliabilityhealth systemsspatial heterogeneity
Bayesian hierarchical model quantifies surveillance reliability with regional reporting probabilities from 0.35 to 0.92.
Identifies significant spatial heterogeneity in system performance across Senegal.
Provides probabilistic assessment moving beyond descriptive metrics for evidence-based strengthening.
Framework enables prioritization of regions for targeted surveillance investment.

Abstract

{ "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 $y{it} \\sim \\text{Poisson}(\\lambda{it} \\cdot \\rho{i})$, where $y{it}$ are observed cases in region $i$ and time $t$, $\\lambda{it}$ is the latent true incidence, and $\\rho{i}$ 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 ($\\rhoi$) 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 $\\rhoi$.", "key words": "surveillance evaluation, Bayesian statistics, hierarchical modelling, health systems, disease