Statistical modelling of injury mortality in Gauteng and Mpumalanga Provinces of South Africa

Abstract

Injury is a truly global health issue with massive societal and economic consequences. In this study, Negative-Binomial integer-valued generalised autoregressive conditional heteroscedasticity (NB-INGARCH) and autoregressive integrated moving average (ARIMA) techniques have been compared and used to build models for both Mpumalanga and Gauteng monthly injury mortality data. The best model was chosen using the root mean square error (RMSE). The best model is the one with the lowest RMSE value. The ARIMA(1, 1, 1) × (1, 1, 1)12 model had the lowest RMSE, making it the most suitable model for both MP and GP monthly injury mortality data. The results identified ARIMA(1, 1, 1) × (1, 1, 1)12 as an appropriate model for predicting Mpumalanga and Gauteng monthly injury mortality with the lowest root mean square error. ARIMA(1, 1, 1)× (1, 1, 1)12 model is applied to forecast the injury mortality for the next two years. Furthermore, the forecasted results of ARIMA(1, 1, 1) × (1, 1, 1)12 model show a decrease of injury mortality in 2020 as compared to 2019. A multifaceted approach to reduce injury mortality is needed. Regulating alcohol sales and raising alcohol prices prevent all forms of violence, while improving drinking environments prevent youth violence. A Graduated Driver Licensing system could benefit the youth driver population to reduce transport accidents.