JMIR Publications

ABSTRACT

Background:

The legalization of cannabis in several U.S. states and Canada has raised concerns over cannabis-impaired driving. However, a paucity of data exists on cannabis consumption patterns and factors that affect risky behaviors associated with cannabis use. As a result, policy makers, insurers, and industry stakeholders have limited quantitative evidence to assess the severity of the problem.

Objective:

The objective of this preliminary study was to quantify the prevalence of cannabis-impaired driving and understand the factors that determine the propensity to drive impaired from users of a digital health educational intervention for cannabis use.

Methods:

Data were analyzed from 1,140 participants who completed “Check Your Cannabis” (CYC) between March and December 2019. The CYC asks a brief set of questions about an individual’s cannabis use, as well as questions about personal beliefs and behaviors. An ordered probit model was used to test relationships between cannabis use, demographics and driving behaviors.

Results:

While gender and age were not statistically significant factors in respondents reporting cannabis-impaired driving, high-risk behaviors were significant determinants of the probability of cannabis-impaired driving. Every 5-point increase in the ASSIST score increased the probability of sometimes driving after cannabis use by 4% (P<.001). Polysubstance use was also a statistically significant determinant of cannabis-impaired driving. Compared to the base group of participants who reported never drinking alcohol or using other substances with cannabis, those who sometimes drink or use other substances with cannabis were 13% (P<.001) more likely to sometimes or always drive after using cannabis. The largest amount spent on cannabis any given day was also a statistically significant predictor of cannabis-impaired driving, however, this effect was small. For example, an increased maximum expenditure on cannabis of $500 increased the probability of reporting sometimes driving after cannabis use by 5% (P=.02).

Conclusions:

To our knowledge this is the first study to examine associations between self-reported cannabis use and driving behaviors. Our analysis indicates that contrary to current research and public perceptions, age and gender were not factors. However, largest amount spent on any given day, higher ASSIST scores, and polysubstance use was positively and significantly associated with driving under the influence of cannabis. Based on these results, public health campaigns and other interventions may have greater impact if they focus resources on problematic cannabis users rather than youth or the general population. Future research may investigate if spending patterns may give insight on those who purchase cannabis from non-retail sources.