Impact of speed on injury severity in single-vehicle run-off-road crashes: Insights from partially temporal constrained modeling approach

Zhe Wang; Chenzhu Wang; Mohamed Abdel-Aty; Lei Han; Helai Huang; Jinjun Tang

doi:10.1016/j.aap.2024.107848

Impact of speed on injury severity in single-vehicle run-off-road crashes: Insights from partially temporal constrained modeling approach

Accid Anal Prev. 2024 Nov 30:210:107848. doi: 10.1016/j.aap.2024.107848. Online ahead of print.

Authors

Zhe Wang¹, Chenzhu Wang², Mohamed Abdel-Aty³, Lei Han⁴, Helai Huang⁵, Jinjun Tang⁶

Affiliations

¹ School of Traffic and Transportation Engineering, Central South University, Changsha 410075, China; Department of Civil, Environmental & Construction Engineering, University of Central Florida, Orlando, FL 32816, United States.
² Department of Civil, Environmental & Construction Engineering, University of Central Florida, Orlando, FL 32816, United States. Electronic address: chenzhu.wang@ucf.edu.
³ Department of Civil, Environmental & Construction Engineering, University of Central Florida, Orlando, FL 32816, United States. Electronic address: m.aty@ucf.edu.
⁴ Department of Civil, Environmental & Construction Engineering, University of Central Florida, Orlando, FL 32816, United States. Electronic address: le966091@ucf.edu.
⁵ School of Traffic and Transportation Engineering, Central South University, Changsha 410075, China. Electronic address: huanghelai@csu.edu.cn.
⁶ School of Traffic and Transportation Engineering, Central South University, Changsha 410075, China. Electronic address: jinjuntang@csu.edu.cn.

PMID: 39616936
DOI: 10.1016/j.aap.2024.107848

Abstract

Single-vehicle run-off-road crashes accounts for approximately 35% of all the traffic fatalities in the U.S during the period of 2019-2021. This paper explores the association between driving speed and injury severity outcomes of single-vehicle run-off-road crashes. The single-vehicle run-off-road crash data from 2019 to 2021 on Interstate freeways in Florida are utilized, and categorized into periods of pre-, during-, and post-COVID-19 pandemic. The partially constrained temporal and temporal unconstrained random parameters logit models are developed considering three injury severity outcomes: no injury, minor injury and serious injury/fatality. Multiple variables in terms of driver, vehicle, roadway, environmental, crash, and temporal attributes are observed to significantly affect the injury severity. Moreover, temporal instability and transferability issues are validated through likelihood ratio test and out-of-sample prediction. In the partially constrained models, numerous variables such as indicators of new vehicle, male driver, and restraint-protected driving consistently yield identical parameter values across all periods, whereas various variables clearly illustrate the distinct differences across the three periods and three speed intervals. The marginal effects in the unconstrained models also display the obvious differences across three periods and three speed intervals. Moreover, the findings corroborate the increased risk outcomes linked to larger speed differences and the COVID-19 pandemic period. These results provide better understanding of the risk mechanisms underlying run-off-road crashes and furnish valuable direction for the formulation of effective safety interventions.

Keywords: Injury severity outcomes; Overspeed behavior; Partially constrained temporal model; Run-off-road crashes.