Drivers testing positive for cannabis were nearly twice as likely to be responsible for fatal crashes at zero blood alcohol, and over three times as likely at very low alcohol levels, compared to substance-negative drivers.
Readers who drive after using cannabis, even without drinking.
Cannabis + trace alcohol (BAC < 0.05): 3.42x higher odds of fatal crash responsibility
What the researchers found
Analyzing 4,294 drivers involved in fatal crashes in California from 1993-2009, researchers found that cannabis elevated crash responsibility even when alcohol was absent or minimal.
At zero blood alcohol concentration (BAC), cannabis-positive drivers had an odds ratio of 1.89 for crash responsibility compared to substance-negative drivers. At very low alcohol levels (0 < BAC < 0.05 g/dL, below any legal limit), the odds ratio jumped to 3.42.
The interaction between cannabis and even small amounts of alcohol appeared to amplify crash risk substantially. These findings suggest that focusing impaired driving enforcement only on high-BAC drivers misses a meaningful source of crash risk from cannabis-impaired driving.
Why it matters
As cannabis legalization spreads, understanding its independent contribution to crash risk becomes critical for traffic safety policy. This study addresses a specific gap: whether cannabis is dangerous for drivers who have consumed little or no alcohol, since previous research had mainly established the combined risk.
The numbers in context
4,294 drivers in fatal crashes. Cannabis at zero BAC: OR = 1.89 for crash responsibility. Cannabis at 0 < BAC < 0.05: OR = 3.42. Data from California, 1993-2009.
How the study worked
Retrospective analysis of fatal crashes merging California SWITRS and FARS databases from 1993-2009. Logistic regression modeled the contribution of alcohol and drugs to crash responsibility. Sample was restricted to 1993-2009 to account for changes in lab testing practices. 4,294 drivers were analyzed.
What this study cannot tell us
Retrospective analysis of crash records with limited control for confounding variables. Drug testing in fatal crashes was not systematic or standardized across jurisdictions and time periods. Cannabis detection in blood or urine does not necessarily indicate acute impairment at the time of the crash, as THC metabolites can persist for days or weeks. The study could not determine dose or timing of cannabis use.
How to read the evidence
Moderate evidence from a large retrospective database study, limited by non-standardized drug testing and inability to confirm acute impairment.
When this study was published
Published in 2017, using data from 1993-2009. Cannabis potency and use patterns have changed significantly since.
The bigger picture
The driving-under-the-influence debate has historically centered on alcohol. As cannabis becomes more accessible, this study makes the case that cannabis-impaired driving is a distinct safety concern, not merely a cofactor with alcohol. The multiplicative interaction between even tiny amounts of alcohol and cannabis is particularly concerning for public safety messaging.
Questions still open
- What is the dose-response relationship between THC blood levels and crash risk? How should per se limits for THC be set given the imperfect relationship between blood levels and impairment? Does the type of cannabis product (edible vs. smoked) affect driving impairment differently?
Common questions
Is it safe to drive after using cannabis if I have not been drinking?
Why is the combination of cannabis and even a little alcohol so dangerous?
Read the original research
Cannabis and crash responsibility while driving below the alcohol per se legal limit.
Accident; analysis and prevention, 108, 37-43
Citation
Romano, Eduardo; Voas, Robert B; Camp, Bayliss. (2017). Cannabis and crash responsibility while driving below the alcohol per se legal limit.. Accident; analysis and prevention, 108, 37-43. https://doi.org/10.1016/j.aap.2017.08.003