Public health surveillance systems in many African nations require robust methodological frameworks to generate reliable forecasts for clinical outcomes, yet evaluations of their predictive performance are limited. This study aimed to methodologically evaluate surveillance data quality and develop a validated time-series forecasting model for key clinical outcomes to inform public health planning. We conducted a retrospective analysis of national surveillance data. A seasonal autoregressive integrated moving average (SARIMA) model, specified as SARIMA (p, d, q) (P, D, Q) ₛ, was fitted and its parameters were estimated using maximum likelihood. Model performance was assessed via rolling-origin cross-validation, with uncertainty quantified using 95% prediction intervals. The evaluation identified systematic under-reporting in specific regions, averaging 18%. The fitted SARIMA (1, 1, 1) (0, 1, 1) ₆ model provided accurate medium-term forecasts, with a mean absolute percentage error of 7. 3% (95% CI: 6. 1 to 8. 5) for the validation period. The methodological framework confirms that rigorously evaluated surveillance data can support reliable forecasting, which is critical for proactive health resource allocation. Implement routine data quality audits in identified regions and integrate the forecasting model into the national health surveillance platform for routine operational use. public health surveillance, time-series forecasting, SARIMA, clinical outcomes, Senegal, methodological evaluation This paper provides a novel, validated methodological framework for integrating data quality assessment with statistical forecasting to strengthen surveillance systems.
Marième Diop (Sun,) studied this question.