Integrating crash and fluids toxicology data to examine injury outcomes and associated driver behaviors

Abstract

Objectives

To examine linked data of drug- and alcohol-involved driving in the State of Connecticut and the resulting association between driver behavior and injury outcomes from motor vehicle crashes.

Methodology

Logistic regression and correlation analysis were conducted on linked toxicology (urine, blood, serum, vitreous) and crash records for the period of 2017 to 2023. Descriptive analysis and simple (Chi²) inferential tests of demographic and crash factors were also conducted. Association of injury outcomes with crash and driver behavior characteristics was measured with estimated odds ratios.

Results

Older age, speeding, lack of safety equipment, testing positive for alcohol alone or with cannabis, and other drugs were significant predictors of driver injury. Gender was not significant. Speeding, lack of safety equipment, and a driver testing positive for alcohol or cannabis alone, or in combination, or for drugs other than cannabis significantly increased the odds of injury for all crash victims; age was not a significant predictor of overall crash severity. Counterintuitively, driver errors served as protective factors for both outcome variables, suggesting other predictors may have masked true relationships.

Conclusions

Study aims have resulted in improved analysis of crash data with the addition of drug classifications. Findings indicate that research of impaired driving behaviors and crash risk can be strengthened through data linkage. While a significant relationship was identified with most predictors, lack of restraint use emerged as the strongest predictor, increasing odds of severe injury nearly 20 times. Driver errors and substance use behaviors require a more thorough examination of their relationship with injury outcomes.