Abstract
{ "background": "The Ethiopian government has implemented a nationwide community health centre system to improve primary healthcare access. Robust evaluation of its adoption is critical for policy, yet methodological rigour in existing assessments varies significantly, with quasi-experimental designs being prominent but inconsistently applied.", "purpose and objectives": "This systematic review aims to critically appraise the application of quasi-experimental methodologies in studies measuring the adoption rates of the community health centre system, identifying common designs, analytical strengths, and limitations to inform future evaluation frameworks.", "methodology": "We systematically searched multiple electronic databases for peer-reviewed and grey literature. Studies employing quasi-experimental designs to assess adoption outcomes were included. Quality was assessed using a modified version of the ROBINS-I tool. A meta-analysis was precluded due to heterogeneity; findings were synthesised narratively. The treatment effect was commonly modelled using a difference-in-differences estimator: $Y{it} = \\beta0 + \\beta1 (Treati \\times Postt) + \\gamma X{it} + \\epsilon_{it}$, where robust standard errors were clustered at the woreda level.", "findings": "Of 42 included studies, 31 utilised difference-in-differences, though only 40% adequately tested the parallel trends assumption. A prominent theme was the frequent omission of spatial spillover effects in analytical models. The most robust studies indicated that system adoption increased facility-based deliveries by an estimated 15–22 percentage points, with wider confidence intervals in earlier evaluations.", "conclusion": "Quasi-experimental evaluations have provided valuable evidence but are often methodologically incomplete, particularly in addressing confounding and spillover. This constrains the certainty of causal inferences regarding the system's impact.", "recommendations": "Future evaluations must rigorously test and report key design assumptions, incorporate spatial econometric techniques to account for interference, and adopt synthetic control methods for settings with fewer control units.", "key words": "health systems research, impact evaluation, difference-in-differences, primary healthcare