{ "background": "Longitudinal assessments of public health interventions in resource-limited settings are hindered by complex, multi-level data structures and non-random missingness. Existing analytical approaches often fail to adequately account for temporal dependencies and heterogeneity across community health centres, limiting robust inference on intervention efficacy.", "purpose and objectives": "This study aimed to develop and methodologically evaluate a Bayesian hierarchical model designed to estimate longitudinal risk reduction within decentralised community health systems. The objective was to provide a robust framework for quantifying intervention effects over time while handling inherent data complexities.", "methodology": "We conducted a longitudinal methodological evaluation using data from a national community health programme. The core model is specified as $y{it} \\sim \\text{Bernoulli}(\\text{logit}^{-1}(\\alpha{i} + \\beta t + \\gamma z{it}))$, with $\\alpha{i} \\sim \\text{Normal}(\\mu{\\alpha}, \\sigma^{2}{\\alpha})$, where $y{it}$ is the binary outcome for centre $i$ at time $t$, $\\alpha{i}$ are centre-specific random intercepts, and $z{it}$ represents time-varying intervention exposure. Model performance was evaluated using posterior predictive checks and comparisons to frequentist alternatives.", "findings": "The proposed model demonstrated superior handling of missing data and centre-level heterogeneity compared to generalised estimating equations. A key methodological finding was that the Bayesian approach yielded more conservative and precise estimates of temporal risk reduction, with posterior credible intervals approximately 15% narrower on average than frequentist confidence intervals under conditions of informative missingness. The model successfully identified significant variation in baseline risk ($\\sigma{\\alpha}$) across centres.", "conclusion": "The Bayesian hierarchical model provides a statistically rigorous framework for longitudinal evaluation in complex, real-world community health systems. It offers a principled approach to uncertainty quantification and inference in the presence of data challenges common in such settings.", "recommendations": "Researchers conducting longitudinal evaluations of health system