pediatric housecalls Robert R. Jarrett M.D. M.B.A. FAAP

Legal Marijuana: Makes Roads Less Safe

An article with the above title appeared in a medical news mashup site that I read frequently and gave some interesting facts and insight.

I’ve been mulling them over in my mind since; so thought that I’d take another opportunity to talk about the difference between “correlation” and “causation” as it relates to scientific, medical research – using their findings as discussion points.
Colorado experiencing major hike in fatalities since medical marijuana legalized
It seems obvious but correlation relates to the fact that some “result” is statistically shown to accompany some prior “action,” either coincidentally or relatedly; while, causation holds a higher standard and requires that the prior “action” be shown to actually be the trigger which produces some result.

At least it’s obvious to most of us who’ve even partially listened to any of our science teachers since about 6th grade as they’ve taught us how not to be so gullible as to be fooled by superstition and advertisers. To journalists, it seems… not so much.

Medical Marijuana and Driving

Correlation v. Causation

For example, let’s talk about “correlation.” Colorado legalized the use of “medical marijuana” and it became commercially available almost instantly in mid-2009. This is actually a good opportunity to do a study because there is a clean start-point for a “before and after,” something rare as far as population studies go.

So, if you were setting out to show a statistical “correlation” between legal marijuana use and, let’s say, automobile fatality rates you could search out rates before and rates after. You should probably also control for other unsuspected variables by using a “control” – perhaps like other surrounding states’ fatality rates who haven’t legalized marijuana use.

On the other hand, to prove “causation” you would need to do something like finding a whole lot of people with the same “conditions” which allow them to purchase medical marijuana, make half of them smoke marijuana and half of them smoke something else (innocuous) before every time they drove and monitor their driving meticulously to see how many of them either died or killed someone.

Obviously, “correlation” is a lot easier and cheaper to prove – and produces a whole lot less uproar by the public. Sometimes, all you’ve got at the moment is “correlation” and it will have to do; but, you should never forget that you are acting without formal causative proof so need to be open to further information.

Colorado “Medical Marijuana” Study

Back to the real world, the federally funded research report and data was published in the journal Drug and Alcohol Dependence back in April of this year (2015) and caused a bit of stir, especially in Colorado.

Bottom line: “There was an obvious increase in the number of drivers involved in fatal motor vehicle crashes who tested positive for marijuana.” Truly, are you really surprised?

The investigators simply and easily analyzed data from the 1994 to 2011 Fatality Analysis Reporting System, which tracks fatal car crashes in all of the United States. Here were their findings:

“Medical Marijuana” vs. Fatalities

  • Marijuana use for “medical purposes” became legalized in Colorado in 2009.
  • Huge numbers of “suppliers” made it commercially available in mid-2009.
  • There are no controls over quality or strength of the product, either federal or state.
  • The proportion of drivers in a fatal motor vehicle crash testing positive for marijuana had actually decreased in Colorado until mid-2009.
  • There was a huge spike in Colorado marijuana-positive drivers in fatality crashes beginning mid-2009.
  • Proportion of Colorado marijuana-positive drivers in fatality crashes has been increasing since mid-2009.
  • (For you math types the change in trend was 2.16 with the “P” value < .0001)
  • There was no significant change in 34 states that had no medical marijuana laws.
  • No similar change for other drug impairments such as alcohol in Colorado during same time period.
  • No similar change for other drug impairments such as alcohol in any non-marijuana-legal state.

Some times, statistics just speak for themselves and really shouldn’t need explanation. There has been a substantial and statistically significant correlation between the use (even legal) of marijuana and the appearance of such drivers as being involved in fatal crashes – either being killed or killing someone else!

Lessons Learned

To my way of thinking, the legislation legalizing medical marijuana use in Colorado sits right up there in “Amateur Status” with Obama’s many bailouts’ not thinking far enough ahead to prevent company CEOs from garnering huge personal financial windfalls all the while they “restructured” thousands out of a job and into pay cuts!

This law should never have been passed in Colorado without first answering how monitoring for quality/dosage and individual usage would be accomplished, who would do it and how often. Testing and monitoring for marijuana use is nowhere near as established or available as in other types of impairment like alcohol and the propensity for abuse is phenomenal.

This should be heeded as a warning by any other state who thinks they should legislate medical issues. I offer the position statement of the American Psychiatric Association (APA) who can speak for most other groups with intelligence:

“Medical treatment should be evidence-based and determined by professional standards of care; it should not be authorized by ballot initiatives. Marijuana that is dispensed under a state-authorized program is not a specific product with controlled dosages. The buyer has no way of knowing the strength or purity of the product, as cannabis lacks the quality control of FDA-approved medicines.

… [even] physicians who recommend use of smoked marijuana for ‘medical’ purposes should be fully aware of the risks and liabilities inherent in doing so”


[Drug Alcohol Depend. Published online April 23, 2014. Abstract]


Share on Google+Tweet about this on TwitterShare on LinkedInPin on PinterestShare on FacebookEmail this to someone
Please share...