Abstract
{ "background": "Evaluating the cost-effectiveness of community health centre systems in low-resource settings requires analytical methods that account for hierarchical data structures and longitudinal cost variations. Existing approaches often fail to adequately model the complex interdependencies between facility-level inputs and population-level health outcomes over time.", "purpose and objectives": "This study aims to methodologically evaluate the application of longitudinal multilevel regression for measuring cost-effectiveness in a national community health system. It assesses the model's capacity to isolate the marginal effect of system-level investments on disability-adjusted life years averted, controlling for contextual confounders.", "methodology": "A longitudinal study design was employed, analysing panel data from a national cohort of community health centres. The core statistical model is a three-level random intercepts regression: $\\text{ln}(\\text{Cost}{ijt}) = \\beta0 + \\beta1\\text{Outcome}{ijt} + \\zeta{i} + \\zeta{ij} + \\epsilon_{ijt}$, where $i$, $j$, and $t$ index district, health centre, and time, respectively. Parameters were estimated using Markov Chain Monte Carlo methods, with inference based on 95% credible intervals.", "findings": "The methodological evaluation indicates that the multilevel approach successfully partitions variance, attributing approximately 65% of cost variation to the district level. A key empirical result is that a 10% increase in supervised community health worker coverage was associated with a 3.2% reduction in cost per DALY averted (95% CrI: 1.8% to 4.5%), demonstrating the model's utility for identifying specific efficiency drivers.", "conclusion": "Longitudinal multilevel regression provides a robust methodological framework for cost-effectiveness analysis in decentralised community health systems, offering superior handling of clustered data and temporal dynamics compared to standard regression techniques.", "recommendations": "Health systems researchers should adopt longitudinal multilevel modelling for economic evaluations in similar contexts. Policymakers should