Pedestrian safety is a growing concern in the United States, with 7,500 fatalities reported in 2022, marking the highest in three decades. Connecticut followed this trend, recording 69 fatalities in the same year. This study examined factors influencing pedestrian injury severity through a multi-level statistical analysis using Connecticut crash data, NHTSA’s VIN decoder, and Canadian Vehicle Specification data.
Crashes were classified into two categories: intersection and non-intersection, and pedestrian injuries were categorized into three: severe (Fatal/K, serious/A), non-severe (Evident/B, Possible/C), or no-injury/property-damage-only (O). Separate multinomial logistic regression models were developed to identify the factors influencing pedestrian injury severity, and binary logistic regression models were developed to compare fatal and serious injuries, providing a deeper analysis of severe injury outcomes.
At non-intersection locations, pedestrian impairment (OR=3.57), driver speeding (2.85), improper crossing (2.84), driver impairment (1.88), and unlighted roadways (1.55) significantly increased the odds of severe injury. At intersections, pedestrian impairment (4.53), speeding (7.40), roadway downgrade (2.04), and unlighted conditions (1.48) were key contributors.
Binary logistic models revealed, at non-intersections, pedestrian age (3% per year), pedestrian impairment (2.03), driver impairment (1.91), and roadway upgradient (3.18) significantly increased the risks of a fatal injury versus a serious injury. At intersections, speeding (7.39) was especially critical, while passive (0.20) and active (0.61) traffic control devices substantially reduced the risk of fatal injury.
The findings provide detailed, context-specific insights to guide pedestrian safety strategies. Reducing pedestrian impairment, enforcing speed control measures, improving roadway lighting, and implementing effective traffic control devices, particularly at intersections, can substantially reduce the likelihood of pedestrian injury severity.