Ambulance crash risk dynamics: a baseline (2017–2019) vs. pandemic-era (2020–2022) comparative study using a random parameter logit model
Ahmed Hossain, Swastika Barua, Subasish Das, Michael Starewich
Abstract
This research addresses the critical need to investigate ambulance-involved person injuries by comparing the baseline (2017–2019) and pandemic era (2020–2022) periods. By using crash data from Texas State (8,859 crashes) and employing the Random Parameter Logit model with the Heterogeneity in Means (RPLHM) model, this study captures unobserved heterogeneity within the crash data. Several variables showed heterogeneous impact on person injury severity during the baseline period (wet surface conditions, 1 pm – 6 pm), and pandemic era (male, dark-with-streetlight, segment). Few variables showed fixed impact on the likelihood of fatal person injury including the first harmful event as ’both vehicles going straight, and involved in a sideswipe collision’, ’collision with parked car’, and ’rural settings.’ Clear weather conditions were found to consistently reduce the likelihood of severe injuries. This study provides insights into how the pandemic influenced ambulance operations, offering a foundation for targeted safety measures and policy interventions.