Cannabis legalization and driving under the influence of cannabis in a national U.S. Sample

The relationship between cannabis legalization and traffic safety remains unclear. Physiological measures of cannabis impairment remain imperfect. This analysis used self-report data to examine the relationship between cannabis legalization and driving under the influence of cannabis (DUIC). Using a cross-sectional national sample (2016-2017) of 1,249 past-30-day cannabis users, we regressed self-reported DUIC (driving within three hours of "getting high") on cannabis legalization (recreational and medical (recreational), medical only (medical), or no legal cannabis), adjusting for demographics, days of use (past 30 days), days of use*legal status, calibration weights, and geographic clustering. The risk of DUIC in recreational (risk ratio [RR] = 0.41, 95% confidence interval (CI):0.23-0.72) and medical (RR = 0.39, 95% CI:0.20-0.79) states was lower than in states without legal cannabis, with one exception. Among frequent cannabis users (≥20 days per month), there was a significantly lower risk of DUIC for those living in recreational states (RR = 0.70, 95% CI: 0.49-0.99), but not for those living in medical states (RR = 0.87, 95% CI: 0.60-1.24), compared to users living in states without legal cannabis. The risk of self-reported DUIC was lower in recreational and medical cannabis states compared to states without legal cannabis. The only exception was for frequent users in medical states, for whom there was no difference in risk compared to frequent users living in states without legal cannabis.