Road Safety Risks of Cannabis Use: Sales Need to Fund Research

State cannabis policies started to become liberalized in the United States in 1996, when Californians voted to allow medical cannabis use under very liberal regulation. Medical cannabis programs have since spread to 34 states and the District of Columbia, and 11 states and the District of Columbia have legalized adult cannabis use for all purposes. Only a handful of states have retained criminal penalties for personal possession and use of cannabis.¹

A major public health concern is that these policy changes may increase the number of motor vehicle crashes involving cannabis-impaired drivers.¹ In the United States the most relevant national data come from the Fatal Accident Reporting System, but it has large amounts of missing data and there are state differences in toxicological testing and definitions of cannabis-impaired drivers.² There are also major challenges in deciding whether drivers involved in fatal accidents were cannabis impaired because cannabis metabolites can be detected hours and days after last use and do not necessarily indicate impaired driving.³

Cook et al. (p. 363) have avoided these problems by assessing the effects of medical cannabis laws and cannabis decriminalization laws on fatal motor vehicle crashes in a sample of US cities with populations of 100 000 or more. Half of the cities are in states that had or had not passed medical cannabis laws and half of the cities had or had not decriminalized personal possession and use of cannabis. They tested the hypothesis that decriminalization and medicalization would have different effects on motor vehicle deaths, specifically that after decriminalization cannabis users would be more likely to drive while cannabis impaired, thereby increasing fatal motor vehicle deaths, whereas states with medical cannabis programs that require cannabis use to take place in the home would keep users off the road and thereby reduce motor vehicle deaths.

Cook et al. used Poisson regression and a difference-in-differences approach to test these hypotheses. Their analysis controlled for the effects of differences between cities in unemployment, beer taxes, and road safety polices that may have reduced road deaths (e.g., bans on texting, sanctions for driving offenses). They also tested the robustness of their findings by using a different statistical model specification. Commendably, they have shared their analytical code and data to allow independent reanalyses.

They found a 13% decrease in road deaths after medical cannabis programs were introduced but a similar sized increase in road deaths after cannabis decriminalization. They argue that these results support their hypotheses because these opposite effects were most marked in younger male adults (who are the most likely to use cannabis and be involved in accidents) and in crashes on weekends (when cannabis use is more likely to occur).

We are cautious about accepting their interpretation for the following reasons. First, medical cannabis laws are intended to increase access to cannabis only for adults with illnesses, but we know that they also increase cannabis use among young adults because of leakage to the illicit market.⁴ One would expect this to increase rather than decrease cannabis use before driving. Second, we are unaware of any evidence that those who access medical cannabis use cannabis only at home and do not drive afterward. We think it more plausible that in states with liberal medical cannabis policies, young adults substitute cannabis for alcohol before driving, thereby reducing the risks of car crashes because cannabis produces less driving impairment than does alcohol. Some studies of the effects on cannabis liberalization have found small, short-lived increases in motor vehicle deaths after the introduction of recreational dispensaries (e.g., Lane and Hall⁵).

The increase in motor vehicle death rates after decriminalization is also puzzling. The authors’ explanation assumes that cannabis users pay attention to criminal penalties for use when criminal penalties did not deter them from using cannabis before decriminalization.

All studies of the effects of cannabis policy changes on car crashes (our own included) require cautious interpretation for a number of reasons.

First, as the authors of a similar study, we are very aware of the limitations of state mortality data and the information used to code state- and city-level cannabis policies. There are major state differences in resources devoted to enforcing criminal penalties for cannabis use and marked differences between states in the liberality of their medical cannabis programs and the extent to which they permit retail cannabis dispensaries to operate. The statistical analyses of state-level mortality data also have a limited capacity to control for the effects of other variables that can affect cannabis use and car crash deaths.¹

Second, these studies examined fatalities for the good reason that they are more dependably identified in state data systems than are injuries. They cannot assess the possibility that an increased cannabis use may increase the number of nonfatal accidents (e.g., if more drivers are impaired by cannabis but fewer of the accidents are fatal because they occur at a slower speed). There are data from a New Zealand birth cohort, for example, indicating that cannabis users are likely to be involved in more nonfatal motor crashes than are their nonusing peers.⁶ An increase in motor vehicle injuries would be much less serious from a public health perspective than would an increase in motor vehicle deaths, but it could still have substantial economic impacts via increased hospitalizations and disability and the costs of motor vehicle repairs.

Third, the findings of the Cook et al. study are consistent with studies showing that cannabis legalization has had minimal short-term effects on motor vehicle deaths. The limited effects to date, however, may not predict what will happen if federal cannabis prohibition in the United States is repealed. Federal prohibition has placed major constraints on cannabis commercialization that have restricted cannabis supply.¹ Local government restrictions on the location of retail cannabis outlets also mean that legalization has so far also had a limited impact on the illicit cannabis market. Removal of these constraints would allow the full commercialization of cannabis markets, lowering prices, increasing potency, expanding access, and permitting widespread promotion of cannabis use in the mass media.¹

Fourth, other state cannabis policy changes could conceivably increase the rates of car crashes. For example, allowing the sale of cannabis-based beverages in bars could increase the combined use of alcohol and cannabis before driving. Allowing the consumption of cannabis in vaping lounges could increase the number of persons who drive after using cannabis.

To better understand the public safety risks of cannabis use, we need large representative cohorts of cannabis users among whom to study associations between cannabis use, driving while impaired, and car crashes, fatal and nonfatal. The impact of cannabis legalization on road safety is an important public health issue for states that have legalized cannabis. They should hypothecate a percentage of tax revenue from legal cannabis sales to fund research to inform governments about how to effectively enforce laws against driving while cannabis impaired and how best to deter cannabis users from putting the safety of others at risk by driving while intoxicated.