Drug use and fatal motor vehicle crashes: A case-control study

Abstract

Drugged driving is a serious safety concern, but its role in motor vehicle crashes has not been adequately studied. Using a case-control design, the authors assessed the association between drug use and fatal crash risk. Cases (n = 737) were drivers who were involved in fatal motor vehicle crashes in the continental United States during specific time periods in 2007, and controls (n = 7719) were participants of the 2007 National Roadside Survey of Alcohol and Drug Use by Drivers. Overall, 31.9% of the cases and 13.7% of the controls tested positive for at least one non-alcohol drug. The estimated odds ratios of fatal crash involvement associated with specific drug categories were 1.83 [95% confidence interval (CI): 1.39, 2.39] for marijuana, 3.03 (95% CI: 2.00, 4.48) for narcotics, 3.57 (95% CI: 2.63, 4.76) for stimulants, and 4.83 (95% CI: 3.18, 7.21) for depressants. Drivers who tested positive for both alcohol and drugs were at substantially heightened risk relative to those using neither alcohol nor drugs (Odds Ratio = 23.24; 95% CI: 17.79, 30.28). These results indicate that drug use is associated with a significantly increased risk of fatal crash involvement, particularly when used in combination with alcohol.

Introduction

Drugged driving has become a safety issue of increasing public concern in the United States and many other countries (Brady and Li, 2013, Dupont et al., 2012, Legrand et al., 2013). Driving performance can be impaired by a wide array of illicit and prescription drugs. Among non-alcohol drugs, marijuana is the most frequently detected substance in the general driver population as well as in drivers being involved in crashes (Arria et al., 2011, Brady and Li, 2013, Jones et al., 2003, Kaplan et al., 2006, Kelly et al., 2004, Lacey et al., 2009a, Walsh et al., 2005). There is mounting evidence that use of marijuana impairs cognitive functions and driving performance, such as psychomotor skills, divided attention, and lane tracking (Battistella et al., 2013, Hartman and Huestis, 2013), and doubles the risk of being involved in a motor vehicle crash (Asbridge et al., 2012, Li et al., 2012). Benzodiazepines have also been frequently detected in drivers (Kelly et al., 2004, Walsh et al., 2004a) and have been consistently found to be associated with a significantly increased risk of crash involvement and crash culpability (Dassanayake et al., 2011, Dubois et al., 2008, Engeland et al., 2007, Leung, 2011, Orriols et al., 2009, Ramaekers, 2003, Rapoport et al., 2009, Smink et al., 2010, Walsh et al., 2004b). Evidence regarding the effects of stimulants on crash risk is inconsistent (Kelly et al., 2004, Engeland et al., 2007, Leung, 2011, Musshoff and Madea, 2012, Smink et al., 2010, Walsh et al., 2004a). While therapeutic use of stimulants does not appear to be related to crash risk, nonmedical use of stimulants, in high doses, or in combination with alcohol or other drugs or with lack of sleep, can pose a threat to driving safety (Bogstrand et al., 2012, Ramaekers et al., 2012, Walsh et al., 2004a). There is a paucity of empirical data about the effects of opioids on driving performance and crash risk despite the tripling of opioid prescriptions in the United States in the past two decades (NIDA, 2010a).

Previous research examining the role of drugs in motor vehicle crashes was limited primarily to case series analysis of toxicological testing data for fatally injured drivers and comparison between drivers involved in culpable crashes and those in non-culpable crashes. Although these studies could provide useful information about the prevalence of drugs in specific driver groups, they are descriptive in essence. To understand the causal role of drug use in motor vehicle crashes, empiric evidence from controlled epidemiological studies is imperative. The case-control design, which requires cases and controls to be selected through such a sampling scheme that they are representative of the respective injured and non-injured source populations (Li and Baker, 2012), has played an instrumental role in understanding the causal relationship between alcohol use and crash risk and in quantifying the dose-response effect of blood alcohol level on crash risk (McCarroll and Haddon, 1962, Zador, 1991). Using a case-control design and drug testing data from two national information systems, we assessed the association of driver drug use with the risk of fatal crash involvement and the interaction effect of drugs and alcohol on fatal crash risk.