Enhancing Safety: Restraints and its Interactions using Partial Proportional Odds Modeling

Seat belts have been unequivocally proven to save thousands of lives annually. However, despite this awareness, a concerning number of individuals persist in neglecting this critical safety precaution. Unrestrained individuals involved in road crashes face a significantly higher risk of suffering fatal injuries, making the encouragement of seat belt adoption a critical road safety imperative. The primary objective of this study is to investigate the interaction between safety restraints, seat positions, and collision types on crash severity. The crash data for the City of Charlotte, North Carolina from 2018 to 2020 was obtained from the Highway Safety Information System (HSIS) and used in the study. The crash severity was categorized into three levels, major injury, minor injury, and property damage only (PDO) crash, and considered as the dependent variable. The characteristics of each crash and its interaction with road, environment, vehicle, and driver attributes were analyzed using the Ordered Logit Model and the Partial Proportional Odds Model, utilizing the SAS 9.4 LOGISTIC procedure to develop these models. The study unveils valuable insights on the influence of safety restraints and seat positions on crash severity by the collision-type. The likelihood of a major injury is high in the case of child restraints, notably in the second left seat position. Likewise, the likelihood of a major injury is high in the case of lap belt usage alone, depending on the collision type. In contrast, the combination of shoulder and lap belts substantially reduces the likelihood of a major injury, with specific exceptions like in ran-off-road crashes. These and other findings from this study emphasize the critical role of shoulder and lap belts and seat position by collision type in enhancing occupant safety.

Keywords: Crash Severity, Partial Proportional Odds, Safety Restraints.