Factor associated with Mortality and Its Prediction Model from Road Traffic Accidents in Health Region 8

Abstract

The road traffic accident has been a major cause of injuries and mortality among global and Thai population. An analytic study aimed to identify risk factors of road traffic mortality and to create a predictive model of mortality from the accidents in areas under Health Region 8, Thailand. The research retrieved retrospective data of injury cases and deaths from road traffic accidents in a total of 72,015 cases who registered Injury and Death Accident Information Management System during 1 October 2021 – 30 September 2022. Data analyses carried out a method of multiple logistic regression statistics and to create a model to predict the mortality from road accidents by performing the Supervised Machine Learning (Decision tree) with the Python language. The results revealed the risk factors that related to the mortality including age, sex, second party, districts with tourist attractions, road types, time of accident, safety helmet wear and seat belt buckles, alcohol drinking, and vehicle types at a significant level of
p-value<0.05. The predictive model of the mortality included all related factors from road traffic accidents revealed 92.5% of its accuracy. This model also enormously presented 89.0% of precision, 97.0% of recall, and 93.0% of F1-Score. Focusing each risk factor, personal factors had the highest accuracy and F1-Score somewhat likely at 65.4% and 67.0%, respectively. To improve safety measures for decreasing of injuries and fatalities, replication of these study’s results is reliable to create enforcement against impacts from road accidents in the future.