{ "background": "Community health centres are a cornerstone of primary healthcare delivery in many African nations, yet robust, data-driven evaluations of their cost-effectiveness remain scarce. Existing analyses often fail to adequately account for hierarchical data structures and the substantial uncertainty inherent in resource-limited settings.", "purpose and objectives": "This study aimed to develop and apply a novel Bayesian hierarchical model to evaluate the cost-effectiveness of community health centre systems, using Uganda as a case study. The primary objective was to quantify the incremental cost per disability-adjusted life year (DALY) averted, while formally propagating uncertainty from multiple data levels.", "methodology": "We conducted an intervention study analysing operational and health outcome data from a network of community health centres. The core methodological innovation is a Bayesian hierarchical model specified as: $\\text{Cost-Effectiveness}{ij} \\sim \\text{Normal}(\\alphaj + \\beta X{ij}, \\sigma^2)$, $\\alphaj \\sim \\text{Normal}(\\mu{\\alpha}, \\tau{\\alpha}^2)$, where $i$ indexes patients and $j$ indexes centres. Parameters were estimated using Hamiltonian Monte Carlo, with cost-effectiveness acceptability curves derived from the posterior distributions.", "findings": "The model estimated a median incremental cost-effectiveness ratio (ICER) of US$ 42.50 (95% credible interval: 28.10 to 72.30) per DALY averted. The probability of the system being cost-effective at a willingness-to-pay threshold of US$ 100 per DALY was 0.92. Substantial heterogeneity was identified between centres, with the random effects standard deviation $\\tau_{\\alpha}$ estimated at 0.31 on the log scale.", "conclusion": "The community health centre system in the studied context represents a cost-effective intervention according to common benchmarks. The Bayesian hierarchical approach provided a statistically coherent framework for handling multi-level uncertainty and centre variation, offering a superior alternative to deterministic or single-level analyses.",