Search for a command to run...
{ "background": "Cost-effectiveness analysis (CEA) is crucial for evidence-based resource allocation in district health systems. In Rwanda, where health system strengthening is a priority, rigorous yet pragmatic evaluation methodologies are required. Quasi-experimental designs (QEDs) offer a viable alternative to randomised controlled trials in real-world settings, but their methodological application and robustness in this context require systematic assessment.", "purpose and objectives": "This systematic review aims to critically evaluate the application, rigour, and reporting of quasi-experimental methodologies used for CEA in Rwandan district hospital systems, identifying methodological strengths, limitations, and best practices.", "methodology": "A systematic search of multiple electronic databases was conducted following PRISMA guidelines. Studies employing QEDs (e.g., difference-in-differences, regression discontinuity, interrupted time series) for CEA within the specified context were included. Data were extracted on study design, identification strategies, outcome measures, cost estimation, and analytical techniques. Methodological quality was appraised using a bespoke tool combining Cochrane risk of bias and economic evaluation checklists. The primary statistical model of interest was the difference-in-differences specification: $Y{it} = \\alpha + \\beta (Treati \\times Postt) + \\gammat + \\deltai + \\epsilon{it}$, where robust standard errors were clustered at the facility level.", "findings": "Of the screened records, a limited but informative corpus of studies was identified. A dominant theme was the use of difference-in-differences to evaluate health financing interventions, though only a minority adequately addressed potential serial correlation and confounding. A key finding was that studies employing propensity score matching prior to difference-in-differences analysis reported more plausible incremental cost-effectiveness ratios, with tighter confidence intervals, suggesting improved causal identification.", "conclusion": "Quasi-experimental CEA in this setting is feasible but inconsistently applied. Methodological rigour, particularly in addressing time-varying confounding and spatial dependencies, is often insufficient, potentially biasing cost-effectiveness estimates