"background": "Community health centres are the cornerstone of primary healthcare delivery in Rwanda, yet systematic, quantitative evaluations of their technical efficiency are lacking. Existing health systems analyses often fail to account for the hierarchical structure of health service data, where centres are nested within districts with varying resource endowments and epidemiological profiles. ", "purpose and objectives": "This protocol details a methodological framework for a multilevel regression analysis to measure and explain variations in the technical efficiency of community health centres. The primary objective is to estimate centre-level efficiency scores while quantifying the proportion of variance attributable to district-level contextual factors. ", "methodology": "We will employ a two-stage analytical design. First, technical efficiency scores for each centre will be calculated using data envelopment analysis. Second, these scores become the dependent variable in a multilevel linear regression model: y{ij = \0 + \ Xij + uj + e₈₉, where i denotes centres and j districts. The model will incorporate centre-level inputs and outputs, and district-level covariates. Inference will be based on 95% confidence intervals estimated using robust standard errors clustered at the district level. ", "findings": "As a research protocol, this paper does not present empirical results. The anticipated analysis will quantify the direction and magnitude of district-level effects on centre efficiency. For instance, it is hypothesised that a significant proportion of total variance in efficiency scores—potentially exceeding 30%—will be attributable to district-level contextual factors. ", "conclusion": "This protocol provides a novel, replicable framework for health systems research that explicitly models the multilevel determinants of technical efficiency. The methodological approach is designed to yield actionable insights for decentralised health system management. ", "recommendations": "Future applications of this protocol should ensure data collection captures both operational and contextual variables. Policymakers should utilise such multilevel analyses to tailor resource allocation and performance benchmarking, recognising the significant influence of higher-level system factors. ", "key words": "
Niyonzima et al. (Thu,) studied this question.