"background": "Evaluating the cost-effectiveness of health system interventions in low-resource settings is critical for policy, but robust quasi-experimental methods are underutilised. District hospital systems in sub-Saharan Africa present a complex environment for such analysis. ", "purpose and objectives": "This study provides a methodological appraisal of applying the difference-in-differences (DiD) model to measure cost-effectiveness within Ugandan district hospital systems, assessing its validity and practical implementation challenges. ", "methodology": "We constructed a DiD model, Y{it = \0 + \1 + \2 + \ (\) +, where Yit is cost per disability-adjusted life year averted. The analysis used panel data from hospital administrative records, testing for parallel trends and employing cluster-robust standard errors at the hospital level. ", "findings": "The methodological application revealed that the parallel trends assumption was violated in 30% of the matched control groups, critically undermining causal inference. The point estimate for \ suggested a cost reduction, but the 95% confidence interval was wide and included zero, indicating substantial uncertainty. ", "conclusion": "While theoretically powerful, the DiD model's application to this context requires stringent validation of its assumptions, which are often unmet in routine health system data, limiting its standalone utility for cost-effectiveness analysis. ", "recommendations": "Future evaluations should integrate DiD with complementary methods like synthetic controls or instrumental variables. Investment in higher-frequency, standardised routine data is essential to meet the model's prerequisites. ", "key words": "health economics, programme evaluation, quasi-experimental design, parallel trends, health systems strengthening, sub-Saharan Africa", "contribution statement": "This paper provides a novel, critical framework for applying difference-in-differences methodology to cost-effectiveness analysis in resource-constrained
Ssebaggala et al. (Sat,) studied this question.