'Cannabis-Involved' Traffic Fatalities Fall in Colorado

The figure refers to crashes in which a driver exceeded the threshold at which state law presumes impairment.


The number of traffic deaths involving drivers who tested positive for marijuana rose by 11 percent last year in Colorado, while the number of "cannabis-involved fatalities" fell by 33 percent. The divergence between those two trends illustrates the difficulty of ascribing crashes to marijuana intoxication, which complicates the question of how legalization has affected road safety in Colorado, the state with the longest history of state-licensed recreational sales.


According to a new report from the Colorado Department of Transportation (CDOT), the number of fatal crashes in which a driver tested positive for cannabinoids rose from 75 in 2014, when legal recreational sales began, to 98 in 2015, 125 in 2016, and 139 last year. But as CDOT notes, "The presence of a cannabinoid does not necessarily indicate recent use of marijuana or impairment." In light of that limitation, CDOT since 2014 also has reported the number fatalities in crashes where a driver's THC blood concentration was five nanogams per milliliter or more, which is the level at which state law allows juries to presume impairment. That number, which CDOT calls "cannabis-involved fatalities," was 19 in 2014 and 2015, then rose to 52 in 2016 before falling to 35 last year.

While the narrower measure gets closer to the issue of concern, many regular cannabis consumers are perfectly capable of safely operating a motor vehicle at five nanograms. More generally, there is little scientific basis for defining marijuana impairment based on THC blood levels. "Whereas the impairment effects for various concentration levels of alcohol in the blood or breath are well understood," the National Highway Traffic Safety Administration notes, "there is little evidence available to link concentrations of other drugs to driver performance." Hence "specific drug concentration levels cannot be reliably equated with a specific degree of driver impairment." Or as the AAA Foundation for Traffic Safety put it in a 2016 report, "a quantitative threshold for per se laws for THC following cannabis use cannot be scientifically supported."

The upshot is that, just as "marijuana-related" crashes (where a driver has consumed cannabis at some point, possible days or even weeks earlier) may be unrelated to marijuana, "cannabis-involved" crashes (where a driver tests at or above five nanograms) may not involve cannabis in any causal sense. That uncertainty makes it hard to say whether legalization has led to a surge in dangerously stoned drivers.

Another complication is the potential substitution of marijuana for alcohol, which has a much more dramatic impact on driving ability. If an increase in stoned drivers were accompanied by a decrease in drunk drivers, the result could be a net decrease in traffic fatalities. So far that does not seem to be happening in Colorado, where traffic fatalities involving drivers who had blood alcohol concentrations of 0.08 percent or more (the legal cutoff for driving under the influence) rose from 160 in 2014 to 171 last year, a period when total fatalities rose from 488 to 648.

Can that overall increase be pinned on marijuana legalization? Probably not, since traffic deaths have been rising nationwide in recent years, and they began rising in Colorado before legalization took effect.

NEXT: Please Enjoy This Video of John McCain Doing 'The Robot'

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. traffic deaths have been rising nationwide in recent years,

    in accordance with the nationwide rise in “mobile device” abuse by drivers.

    1. in accordance with the nationwide rise in “mobile device” abuse by drivers

      You have the wrong moral panic, Rich. This week it is Drugged Driving. Texting While Driving will be back next week.

  2. ‘fraid not.

    Fact is, mobile devices have reached saturation in the US a decade ago. Everyone has one. Everyone has been ‘abusing’ them for a decade. The number of mobile devices in the hands of drivers is not increasing but the nember of fatalities is. No correlation.

    1. From here in Denver, it’s pretty obvious to me that the increase in shitty drivers is a combo of texting/mobiles and new-to-state.

      1. Having been here since 1970, it became readily apparent that drivers from different parts of the country had very different styles of movement with a car…:-)…JFree, well observed; well written. Let’s just all play nice on the road, and help each other out, rather than gladiating,(?behaving as a gladiator) yes? If you are in a hurry, where ya’ goin’? Think how long it would take to go from A to B walking, by horse, or bicycle; slow down, we’ll all get there sooner, yes?

    2. The number of mobile devices may not have gone up, but I’d bet usage has. Per Cisco, mobile data traffic grew 4000 times between 2006 and 2016.

      The Rise of Mobile

    3. And as a younger generation, whom have put so much of their thought power into their phone, become a majority of drivers? Agammamon; good point. I’m gonna have to peek at the age and number of phone attached drivers. Just thinking….peace

  3. If cannabis-impaired driving were really causal of traffic fatalities, we would expect increased fatalities if cannabis use is increasing. More years of data will help address this question, but the obvious implication is that cannabis impairment does not increase traffic fatalities.
    No one should drive impaired, but actual impairment should be measured, and the level of impairment from cannabis that is criminalized should be the same as the level of impairment for the blood alcohol limit. I have developed a new public health app that is a general measure of impairment from cannabis or any source–anything that impairs reaction time, hand-eye coordination, balance and the ability to perform divided attention tasks–it is called DRUID (an acronym for “DRiving Under the Influence of Drugs”) available now in the App Store and in Google Play. DRUID measures reaction time, decision making, hand-eye coordination, time estimation and balance, and then statistically integrates hundreds of data points into an overall impairment score. DRUID takes just 2 minutes.

    DRUID was recently featured on the PBS News Hour ( and in Wired magazine:
    Our website is

    After obtaining my Ph.D. at Harvard, I have been a professor of psychology at UMass/Boston for the past 40 years, specializing in research methods, measurement and statistics.

    1. Nice work Doc; just don’t let math always replace considered common sense. Bumblebees can’t fly, remember…peace

  4. Is there a legal argument for push back of a law based on completely arbitrary, unproven with science application? 5 nanograms? Anyone understand titration (tolerance) for that matter? Titration occurs with cannabis, though a very different biochemistry and neurological effect than other molecules. Just safe..#:-)

  5. This will all be obviated with the arrival of the driverless car.

Please to post comments

Comments are closed.