{ "background": "District hospital systems in low-resource settings require robust, context-sensitive methods for evaluating cost-effectiveness to inform resource allocation. Traditional frequentist approaches often struggle with sparse, hierarchical data and fail to quantify uncertainty fully.", "purpose and objectives": "This study aimed to methodologically evaluate a Bayesian hierarchical modelling framework for measuring cost-effectiveness in a low-resource healthcare system, using district hospitals as a case study.", "methodology": "We developed a Bayesian hierarchical model to estimate cost per disability-adjusted life year (DALY) averted across multiple hospital sites. The model structure is: $\\text{Cost-Effectiveness}{ij} \\sim \\text{Normal}(\\mu{ij}, \\sigma^2)$, $\\mu{ij} = \\alpha + \\beta X{ij} + uj$, $uj \\sim \\text{Normal}(0, \\tau^2)$, where $i$ indexes interventions and $j$ indexes hospitals. We used weakly informative priors and Hamiltonian Monte Carlo for inference. Model performance was assessed using posterior predictive checks and comparison to a frequentist counterpart.", "findings": "The Bayesian model provided probabilistically richer estimates, with the 95% credible intervals for cost per DALY averted being, on average, 28% wider than frequentist confidence intervals, reflecting greater uncertainty incorporation. A key finding was the substantial heterogeneity between hospitals (site variance, $\\tau^2$, estimated at 0.42 with a 95% CrI of [0.31, 0.58]), which was poorly characterised by the standard model.", "conclusion": "The Bayesian hierarchical approach offers a superior methodological framework for cost-effectiveness analysis in heterogeneous, data-limited settings, as it formally accounts for multi-level uncertainty and unit variation.", "recommendations": "Health economists and policy analysts should adopt Bayesian hierarchical modelling for resource allocation decisions in decentralised health systems. Future research should focus on developing accessible computational tools for local analysts.", "key words":