Statistical Time-Series Models for Long-Range Public Health Trend Forecasting: A Review and Conceptual Framework

Long-range forecasting of public health trends is critical for evidence-based health system capacity planning, pharmaceutical supply chain management, and health equity investment. This paper presents a comprehensive review of statistical time-series modelling approaches for long-range public health trend forecasting across five domains: infectious disease incidence, chronic disease prevalence projection, healthcare utilisation and demand forecasting, pharmaceutical supply chain demand forecasting, and climate-health trend forecasting. ARIMA models, exponential smoothing state space ETS models, Bayesian structural time-series models, and hybrid statistical and machine learning approaches are systematically evaluated. An R-based implementation framework using the forecast, tidyverse, ggplot2, dplyr, and lubridate packages provides a reproducible methodological template. A responsible public health forecasting framework and comparative model performance table are proposed.