Space and time clustering of road traffic collisions among older adults in Taiwan.

Abstract

To identify and characterize space-time clusters of road traffic collisions (RTCs) involving older adults in Taiwan, with emphasis on spatial and temporal features that may inform targeted prevention strategies. We analyzed nationwide RTC data from Taiwan's Police Traffic Accident Report (PTAR) registry from 2014 to 2023, including 145,450 older adult victims aged ≥65 years. Six variables, including three spatial (urbanization level, crash location, and type of traffic signal) and three temporals (monsoon season, day of the week, and time of day), were selected to perform latent class analysis (LCA) for identifying distinct spatiotemporal crash clusters. Model fit indices (AIC, BIC, CAIC, ABIC, and entropy) guided the selection of the optimal number of clusters. Demographic and road user characteristics across clusters were compared using bivariate analyses. Three distinct clusters were identified: (1) urban intersection crashes, (2) intersection crashes in medium- and low-urbanized areas, and (3) crashes on unsignalized road segments. Collisions were more likely to occur at intersections (n = 85,247, 58.6 %) and in highly urbanized areas, (n = 61,432, 42.2 %). Most incidents took place on weekdays (n = 103,441, 71.1 %) and during morning hours (n = 70,382, 48.4 %). Significant differences across clusters were found in age, sex, road user role, and vehicle type (all p < 0.001). This study demonstrates the heterogeneity in spatiotemporal patterns of RTCs involving older adults in Taiwan. These findings highlight areas where further investigation into context-specific traffic safety measures could inform efforts to enhance mobility and reduce injury risk among older adults.