{ "background": "Community health centres are critical nodes in Uganda's healthcare delivery, yet systematic analysis of methodologies for forecasting and improving their operational yield is lacking. A rigorous evaluation of predictive modelling approaches is required to inform resource allocation and strategic planning. ", "purpose and objectives": "This systematic review aims to critically evaluate methodological approaches used in forecasting yield metrics for Ugandan community health centres and to synthesise evidence on the performance of time-series models for predicting service output improvements. ", "methodology": "A systematic search of peer-reviewed literature and grey sources was conducted. Eligible studies were those employing quantitative forecasting models for health centre outputs. Methodological quality was appraised using a modified checklist for time-series analysis. Model performance was synthesised, with a focus on the generalised autoregressive integrated moving average (GARIMA) framework, represented as \ (B) (1-B) ᵈ yt = \ (B) \ + =1ᵏ \ x₈ₓ, where B is the backshift operator. ", "findings": "The review identified a predominant reliance on autoregressive integrated moving average models, which demonstrated moderate forecasting accuracy for patient visit volumes, with a mean absolute percentage error (MAPE) of 18. 7% (95% CI: 15. 2, 22. 3) across studies. A key thematic finding was the consistent underutilisation of exogenous variables capturing community-level determinants in model specifications, limiting predictive power. ", "conclusion": "While time-series forecasting is increasingly applied, methodological rigour is variable. The integration of community-specific covariates into robust model structures is essential for generating actionable forecasts to guide health systems strengthening. ", "recommendations": "Future research should prioritise the development and validation of hybrid models that incorporate climatic, demographic, and supply-chain covariates. Capacity building in advanced statistical modelling for health planners is urgently needed to improve forecast utility. ", "key words": "health systems, forecasting, time-series analysis, operational research, resource allocation, predictive modelling", "
Ocen et al. (2000) studied this question.