{ "background": "Rural primary healthcare systems in South Africa face significant challenges in resource allocation and demand planning. Accurate forecasting of clinical outcomes is critical for improving service delivery, yet the methodological rigour of existing predictive models in this specific context remains unevaluated.", "purpose and objectives": "This systematic review aims to critically evaluate the methodological approaches used in time-series forecasting models for clinical outcomes within rural primary healthcare clinics in South Africa, and to synthesise evidence on their predictive performance and applicability.", "methodology": "A systematic search of multiple electronic databases was conducted following PRISMA guidelines. Studies employing time-series or longitudinal forecasting models for clinical outcomes (e.g., patient volumes, disease incidence) in rural primary care settings were included. Data were extracted on model specifications, validation techniques, and performance metrics. Study quality was appraised using a modified checklist for prognostic model research.", "findings": "The review identified a limited corpus of studies. A predominant theme was the reliance on autoregressive integrated moving average (ARIMA) models, represented generally as $Xt = c + \\sum{i=1}^{p}\\phii X{t-i} + \\epsilont + \\sum{i=1}^{q}\\thetai \\epsilon{t-i}$. However, fewer than 30% of studies incorporated exogenous variables related to community-level determinants. Model performance was frequently reported without confidence intervals for error metrics, limiting inference on forecast uncertainty.", "conclusion": "Existing forecasting methodologies for rural clinical outcomes are narrowly focused on temporal autocorrelation, often neglecting broader socio-environmental drivers. This constrains their utility for strategic planning in complex, resource-constrained health systems.", "recommendations": "Future research should develop and validate hybrid models that integrate traditional time-series approaches with machine learning techniques to capture exogenous factors. Reporting must include robust uncertainty quantification, such as prediction intervals, to better inform policy decisions.", "key words": "time-series analysis, forecasting, primary healthcare, clinical outcomes, rural health systems, South Africa, systematic review