Early identification of high-risk older two-wheeler riders: A dual-sample approach for 30-day mortality prediction

Abstract

Older motorized two-wheeler riders face high risks of severe injuries and fatalities following motor vehicle crashes (MVCs), yet existing studies often rely on hospital-based data, potentially underestimating mortality risk. This study aims to identify predictors of 30-day mortality among older riders by leveraging both population-based and hospital-based samples. We conducted a retrospective cohort study using Taiwan’s Police-Reported Traffic Accident Registry (2019–2021), Taiwan Death Registry, and Taiwan National Health Insurance database (2016–2021). Two cohorts were established: a population-based sample of older motorized two-wheeler riders involved in crashes and a hospital-based sample of those admitted on the same day or one day after MVCs. The primary outcome was 30-day all-cause mortality. Predictive variables included injury severity, demographics, comorbidities, behaviors, and environmental factors. Logistic regression models assessed predictors, with model performance evaluated using the area under the receiver operating characteristic curve (AUC). Significant predictors of 30-day mortality included severe injury, head and neck trauma, older age, alcohol consumption, unlicensed riding, and crash location. Diabetes mellitus was a significant predictor in the population-based sample but not in the hospital-based sample, highlighting differences in data capture. Both models showed acceptable predictive ability, with AUC values of 0.80 and 0.84, respectively. A prehospital screening tool based on population-based data demonstrated effective early risk estimation at the crash scene. A population-based screening tool based on prehospital information may facilitate timely intervention and enhance trauma care for older motorized two-wheeler riders, supporting improved outcomes in aging populations.