{ "background": "Evaluating the efficiency of community-based health systems in low-resource settings remains methodologically challenging, particularly for capturing longitudinal effects of systemic interventions.", "purpose and objectives": "This study aims to quantify longitudinal efficiency gains within a national community health centre system, employing a quasi-experimental design to isolate the impact of a major structural reform on service delivery metrics.", "methodology": "A longitudinal, controlled interrupted time-series analysis was conducted. Panel data from a nationally representative sample of centres were analysed using a generalised estimating equations model: $Y{it} = \\beta0 + \\beta1 Tt + \\beta2 X{it} + \\beta3 (Tt \\times X{it}) + \\alphai + \\epsilon{it}$, where $Y{it}$ is the efficiency score for centre $i$ at time $t$, $Tt$ is a post-intervention indicator, and $X{it}$ are time-varying covariates. Inference is based on cluster-robust standard errors.", "findings": "Preliminary analysis of the initial longitudinal phase indicates a positive trajectory in technical efficiency, with an average increase of 12.3 percentage points (95% CI: 8.7, 15.9) observed in intervention centres relative to controls over the first major observation period.", "conclusion": "The methodological approach robustly captures the dynamic effects of systemic reform, suggesting that the intervention initiated a measurable improvement in health system efficiency.", "recommendations": "Future health systems research should adopt similar longitudinal quasi-experimental designs to attribute causality in complex, real-world settings. Policymakers should consider phased roll-outs of major reforms to facilitate rigorous evaluation.", "key words": "health systems research, efficiency measurement, quasi-experimental design, interrupted time series, longitudinal data, sub-Saharan Africa", "contribution statement": "This study provides a novel longitudinal framework for causal inference in health system efficiency evaluations, generating a unique panel dataset that tracks the