Abstract
{ "background": "District hospitals in Ethiopia face persistent systemic shocks from climate, conflict, and disease, yet a robust, quantitative methodology for forecasting their operational resilience and evaluating risk reduction interventions is lacking. Existing assessments are often cross-sectional and descriptive, failing to model temporal dynamics and forecast future system states under stress.", "purpose and objectives": "This protocol details a methodological framework for evaluating the resilience of district hospital systems through a time-series forecasting model. The primary objective is to develop and validate a model that forecasts key service capacity indicators, enabling the simulation of intervention impacts on risk reduction over a multi-year period.", "methodology": "We will employ a longitudinal, quantitative design using retrospective administrative data. The core analytical model is a seasonal autoregressive integrated moving average (SARIMA) model, specified as $\\phi(B)\\Phi(B^s)\\nabla^d\\nablas^D yt = \\theta(B)\\Theta(B^s)\\epsilont$, where $yt$ represents a hospital output metric. Model parameters will be estimated via maximum likelihood, with forecast uncertainty quantified using 95% prediction intervals. The model will be used to simulate counterfactual scenarios for specific resilience-building interventions.", "findings": "As a research protocol, this paper does not present empirical results. However, the proposed methodology is designed to generate forecasts of hospital service outputs (e.g., monthly outpatient attendance). A key anticipated output is the simulated impact of a specific intervention, such as estimating the proportion of service degradation mitigated by a backup power supply during forecasted drought periods.", "conclusion": "This protocol establishes a novel, evidence-based methodological approach for proactively assessing and strengthening health system resilience. The forecasting framework moves beyond descriptive analysis to provide a tool for pre-emptive policy planning and resource allocation.", "recommendations": "We recommend the adoption of this time-series forecasting approach by health planners and researchers for routine resilience monitoring. Future work should integrate this model with early warning systems to enable dynamic, data-driven responses to emerging threats.", "key words":