Abstract
{ "background": "Community health centres in Nigeria face systemic inefficiencies, particularly in diagnostic pathways, which compromise service delivery and health outcomes. Current evaluations often lack rigorous designs to attribute changes in system performance to specific interventions, limiting evidence for scale-up.", "purpose and objectives": "This protocol details a quasi-experimental design to evaluate the impact of a diagnostic optimisation bundle on system yield, defined as the proportion of correct clinical decisions supported by timely diagnostics. Primary objectives are to estimate the causal effect on yield and to identify key modifiable health system factors.", "methodology": "A controlled before-and-after study will be implemented across 20 centres, matched and allocated to intervention or control arms. The intervention is a bundled protocol for febrile illness management. The primary outcome is system yield, measured via clinical audit. The analysis will employ a difference-in-differences model: $Y{it} = \\beta0 + \\beta1 (Groupi \\times Timet) + \\gamma X{it} + \\epsilon{it}$, where $Y{it}$ is the yield for centre $i$ at time $t$. Inference will rely on cluster-robust standard errors to account for centre-level clustering.", "findings": "As a protocol, no empirical results are presented. The anticipated primary finding is a positive directional change in system yield for the intervention group. The analysis is powered to detect a minimum absolute increase of 15 percentage points in the yield proportion.", "conclusion": "This protocol provides a methodological framework for robust, causal evaluation of health system interventions in resource-constrained settings, moving beyond descriptive assessment.", "recommendations": "Future health systems research should adopt quasi-experimental designs with pre-specified causal models to strengthen the evidence base for policy. Implementation should be accompanied by detailed process evaluation.", "key words": "health systems research, quasi-experimental design, diagnostic stewardship, causal inference, Nigeria, primary health care", "contribution statement": "This protocol introduces a novel application of a