Abstract
{ "background": "Public health surveillance systems in sub-Saharan Africa generate vast data, yet their methodological rigour for forecasting clinical burdens is often under-evaluated, limiting proactive resource allocation.", "purpose and objectives": "This case study aimed to methodologically evaluate the forecasting capability of a national surveillance system by developing and validating a time-series model for predicting key clinical outcomes.", "methodology": "We constructed a univariate forecasting model using historical, de-identified surveillance data. The core model was a seasonal autoregressive integrated moving average (SARIMA) formulation: $yt = \\mu + \\Phi(B^s)\\phi(B)(1-B)^d(1-B^s)^D yt + \\Theta(B^s)\\theta(B)\\epsilont$, where $\\epsilont \\sim N(0, \\sigma^2)$. Model selection was based on minimising the Akaike Information Criterion, with forecast uncertainty quantified using 95% prediction intervals.", "findings": "The model demonstrated robust in-sample fit and out-of-sample forecasting accuracy for a major clinical indicator. A key finding was a projected 18% increase (95% prediction interval: 12% to 24%) in the annual caseload over the forecast horizon, highlighting a significant upward trend. The evaluation revealed specific systemic gaps in data granularity affecting forecast precision.", "conclusion": "The surveillance system provides a viable foundation for forecasting, but its utility is constrained by data quality and completeness issues. Methodological evaluation through modelling exposes critical points for system strengthening.", "recommendations": "Integrate routine forecasting audits into surveillance system evaluations. Invest in improving the temporal resolution and clinical detail of reported data to enhance model accuracy and public health utility.", "key words": "surveillance evaluation, time-series forecasting, SARIMA, public health informatics, health systems, predictive modelling", "contribution statement": "This study provides a novel methodological framework for the quantitative evaluation of public health surveillance systems' forecasting performance, moving beyond descriptive assessments