Substance-Involved Driving: Predicting Driving after Using Alcohol, Marijuana, and Other Drugs.

Substantial research has examined the influence of alcohol, marijuana, and other illicit drugs on driving performance; however, which psychosocial characteristics of individuals who drive while under the influence of alcohol (DUIA), marijuana (DUIM), and other drugs (DUID), how these characteristics interrelate with each other, and how they differ across degrees of substance-involved driving (SID) have not been thoroughly investigated. This article identified psychosocial predictors of SID while accounting for driving behavior and the type and level of substance use and examined the associations of psychosocial characteristics and SID with citations for traffic offenses. Telephone survey data and state driver history records for a sample of 5,244 young adults were analyzed using t-tests and logistic and multinomial logistic regression analysis to examine the correlates and predictors of substance-involved driving. Psychosocial characteristics predicted DUIA, DUIM, and DUID when tested in separate models and adjusting for driving behavior. When the substance in question was added to each model, a unique association between psychosocial characteristics and DUIA remained, but the associations between psychosocial characteristics and DUIM and DUID were completely mediated by the frequency of marijuana use and level of other drug use in their respective models. Multinomial logistic regression predicting the degree of SID, which was based on the types and combinations of SID behaviors, showed that after controlling for the use of alcohol, marijuana, and other drugs, psychosocial characteristics maintained a unique association with the degree of SID. Finally, when adjusting for driving behavior and psychosocial characteristics, the degree of SID predicted having a traffic offense. These results indicate that reducing substance use is not the only means of targeting substance-involved driving. Interventions could have enhanced effectiveness if they also targeted individual psychosocial and behavioral characteristics, either to alter these behaviors or by tailoring the intervention or program for these characteristics. (Author/publisher).