Predictive time series modelling of under five mortality in Ghana

Under-five mortality (U5M) remains a significant public health challenge in Ghana. Although steady progress has been made, the pace of decline raises concerns about whether the country is on track to meet the global target. Hence, understanding the future trajectory of under-five mortality is critical for policy and resource allocation. To determine the trend and forecast under-five mortality rates in Ghana. This is a predictive study that applies time series analysis. Secondary data from the World Bank database was used for the study. A comparative time-series analysis was conducted using three distinct prognostic models: Error, Trend, Seasonality (ETS), Autoregressive Integrated Moving Average (ARIMA), and Prophet. The best-performing model was selected based on out-of-sample accuracy metrics, Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Corrected Akaike Information Criterion (AICc), and Bayesian Information Criterion (BIC). Time series analysis, modelling, and forecasting were performed using R. An exploratory analysis was performed to understand the characteristics of the distribution before model fitting. The best-fitted models were chosen to predict and forecast under-5 mortality rates for 7 years using ARIMA, ETS, and Prophet. Statistical significance was set at 0.05. The study showed a decreasing trend rate of 3.22(21%) per decade in U5M rates. The highest rate recorded in Ghana was 208 in 1960, decreasing to 44 per 1000 live births in 2020, and further to 37 per 1000 live births in 2023. From this analysis, ETS was the best performing model for predictive accuracy compared with ARIMA and Prophet. Using the ETS model from 2024 to 2030, U5M is forecasted to decline from 35.6 (35.4–35.9) to 28.1 (26.1–30.3). The rate would fall steadily through 34.3 (33.8–34.7) in 2025, 32.9 (32.2–33.7) in 2026, 31.7 (30.6–32.8) in 2027, and 30.4 (29.0–31.9) in 2028. By 2029, it would further reduce to 29.3 (27.5–31.1). U5M in Ghana has shown a consistent decline, with forecasts indicating a reduction by 2030. Although this trend is encouraging, it remains above the SDG target by 2030. The findings underscore the need for accelerated policy action, particularly in allocating resources to strengthen maternal and child health interventions.