How does distraction affect cyclists' severe crashes? A hybrid CatBoost-SHAP and random parameters binary logit...

Area of Science:

• Road safety research
• Transportation safety analysis
• Injury epidemiology

Background:

• Cyclists are vulnerable road users facing increasing distractions.
• Distraction's impact on cyclist injury severity is understudied.
• Urban environments present unique cyclist safety challenges.

Purpose of the Study:

• To analyze the effect of cyclist distraction on crash injury severity.
• To identify factors contributing to severe injuries in distracted cyclist crashes.
• To inform targeted safety interventions for cyclists.

Main Methods:

• Analysis of U.S. Crash Report Sampling System (CRSS) data (2019-2022).
• Hybrid framework integrating CatBoost-SHAP and random parameters binary logit model with heterogeneity (RPBL-HMV).
• Statistical modeling to assess distraction and contributing factors on injury severity.

Main Results:

• Cyclist distraction is a significant factor in crash injury severity.
• Higher injury risk associated with front motor vehicle impacts, rural areas, two-way roads, higher speed limits, and weekends.
• Reduced injury risk linked to T-intersections, side/rear impacts, helmet use, and rush hour.
• Interaction effects noted, e.g., crossing actions and rush hour combined increase severe crash probability.

Conclusions:

• Cyclist distraction is a critical safety concern requiring mitigation strategies.
• Specific crash characteristics (location, time, vehicle interaction) influence severity.
• Helmet use and avoiding distraction during rush hour are key protective factors.
• Findings support targeted policies for safer cycling environments.