Community health centres in Kenya are pivotal for healthcare delivery, yet their effectiveness varies significantly. The current monitoring and evaluation systems often lack precision, leading to challenges in understanding clinical outcomes. A mixed-methods approach was employed, integrating observational data from existing healthcare records with predictive modelling techniques. Time-series analysis using autoregressive integrated moving average (ARIMA) model was applied to forecast future trends in clinical outcomes. The ARIMA model demonstrated a moderate success rate in predicting clinical improvement trajectories over the next six months, showing an accuracy of approximately 75% with a confidence interval of ±10%. This suggests that time-series models can be effective tools for forecasting within these settings. While preliminary results indicate potential utility, further validation and refinement are necessary to ensure the robustness of ARIMA in diverse clinical contexts. Future studies should explore additional model types and incorporate real-time feedback mechanisms to improve predictive accuracy. Community health centres, Kenya, time-series forecasting, clinical outcomes, ARIMA Treatment effect was estimated with logit (pᵢ) =₀+^ Xᵢ, and uncertainty reported using confidence-interval based inference.
Chepkoyo et al. (Thu,) studied this question.