{ "background": "Community health centres (CHCs) are critical nodes in primary healthcare delivery, yet systematic evidence on their operational reliability in low-resource settings is scarce. Current evaluations often rely on cross-sectional surveys, which lack the temporal resolution to assess system performance under routine and stressed conditions.", "purpose and objectives": "This protocol details a quasi-experimental design to quantify the reliability of CHC systems, defined as the consistency of service provision over time and across user groups. The primary objective is to estimate the system reliability index (SRI) for a network of centres. Secondary objectives include identifying key infrastructural and procedural determinants of reliability.", "methodology": "A longitudinal, controlled before-and-after study will be implemented across 20 purposively selected CHCs. The intervention arm (10 centres) will receive a standardised equipment and training package, while the control arm (10 centres) will continue routine operations. Reliability will be measured monthly over 12 months using a composite score derived from drug stock-out rates, equipment functionality, and staff absenteeism. The primary analysis will employ a difference-in-differences model: $SRI{it} = \\beta0 + \\beta1 \\text{Treatment}i + \\beta2 \\text{Post}t + \\beta3 (\\text{Treatment}i \\times \\text{Post}t) + \\epsilon{it}$, with inference based on cluster-robust standard errors.", "findings": "As a protocol, no empirical findings are presented. Anticipated results include direction and magnitude of the intervention effect on the SRI. For instance, we hypothesise a minimum 15-percentage-point improvement in the composite reliability score for the intervention group relative to controls.", "conclusion": "The proposed methodology provides a novel, structured framework for generating rigorous, operationally relevant evidence on health system performance. It moves beyond snapshot assessments to capture the dynamic stability of service delivery.", "recommendations": "Future research and national monitoring programmes should adopt