Assessing Injury Severity in Crashes Involving Older Drivers Using Machine Learning Techniques

Abstract: In recent years, the number of older adults holding driver’s licenses in the United States has grown significantly. While this trend reflects the desire for mobility and independence among aging populations, it also raises safety concerns. Age-related declines in cognitive, vision, and physical abilities can adversely affect driving performance, making older drivers more susceptible to severe outcomes in traffic incidents. According to the Centers for Disease Control and Prevention (CDC), in 2020 alone, over 7,500 older adults (aged 65 and above) lost their lives in traffic crashes, while more than 200,000 were treated in emergency departments for crash-related injuries. These alarming statistics emphasize the need to investigate factors that contribute to crash frequency and severity among older drivers. This study addresses this need by applying machine learning methods—namely K-Nearest Neighbor, Random Forest, Support Vector Machine, Multilayer Perceptron, and XGBoost—to identify key factors influencing injury severity in crashes involving older drivers. Utilizing three years (2017-2019) of crash data from New Jersey, the study aimed to determine which variables most significantly contribute to severe injury outcomes in these incidents and to assess the accuracy of various machine learning models in predicting injury severity. Among the methods tested, XGBoost demonstrated superior performance in accurately estimating injury severity, underscoring its effectiveness for this type of analysis. Analysis of the model’s results revealed that certain conditions, including angular crash types, involvement of multiple vehicles, and dry road surfaces, were strongly associated with higher injury severity among older drivers. These factors highlight specific crash circumstances that are particularly dangerous for older adults, offering a basis for targeted interventions. The findings of this study provide valuable insights for transportation researchers, traffic engineers, and policymakers focused on enhancing road safety for older adults. By pinpointing circumstances under which older drivers are most at risk, this research supports the development of evidence-based countermeasures aimed at reducing both the frequency and severity of crashes involving older drivers. Implementing these countermeasures has the potential to improve overall safety for aging drivers, contributing to a safer transportation environment as the population continues to age.