The importance of data presentation and peer review in medical reporting

Pippa Hawley is the Head of the Division of Palliative Care at the University of British Columbia

 

The recent headlines about legalization of cannabis in Canada - a subject already fraught with bias - has illustrated the need for care when reporting on statistical observations. An example of things going badly wrong is a recent Vancouver Sun column headline “Fatal car crashes triple among drivers high on marijuana after legalization in Colorado; double in Washington state”. This was based on an article in the October issue of the BCMJ. The headline would seem to indicate that there has been a dramatic increase in fatal car crashes in those two states caused by people driving high on newly-legal cannabis.

This would be very important information and a pretty persuasive argument against legalization of cannabis, if it were true.

The facts actually show the opposite. Traffic fatalities have been falling in both Colorado and Washington since legalization of cannabis, and in fact were falling even prior to legalization.

Traffic crash-related death rates in Colorado and Washington are amongst the lowest in the country. Washington’s death rate in 2014 was only 6.5/100,00 people, or 0.8 deaths per 100 million miles driven, as compared with a national average of 10.2/100,000, or 1.08/100 million miles. The state of Mississippi has the horrific rate of 25.7 deaths per 100,000, still nearly twice as high as Washington’s when factoring in the increased number of miles driven. The only states that had a lower rate than Washington were the District of Columbia (3.5/100,000 people and 0.65/100 million miles), Rhode Island (4.9 and 0.68 respectively), and New York (5.3, 0.8). And funny how the District of Columbia also has legalized access to recreational cannabis! This information is freely available on the internet from the Insurance Institute for Highway Safety, Highway Loss Data Institute.

So how did this message get so twisted?

Let’s look at the original statements which led to the misleading headline. The primary source of data included a report produced by the American Automobile Association’s Foundation for Traffic Safety in May of 2016. They looked at all the people who had had blood tests drawn following a road traffic accident for three years prior to (2010-12), and two years (2013-14) after introduction of legalized access to cannabis for recreational purposes. They stated that finding a cannabis metabolite in the blood of these people was substantially more common after legalization than before, in the just under half that actually got tested. Of those who tested positive, two thirds also had alcohol and/or other drugs in their blood, which were probable contributors to driving impairment. There is no mention of the prevalence of positive cannabinoid blood tests in people who were not in traffic accidents, nor any mention of whether the people tested were the causes of the crashes or were victims. There was also no mention of the actual number of traffic crashes or fatalities occurring before and after legalization.

Although there is no reason to believe this group were biased, it seems odd that these glaring omissions were allowed to pass through all levels of review leading up to publication.  It seems unlikely that the wording of the report was not intended to alarm.

This is where the chain started, but unfortunately not where it ended. The report was used in an article that was published without peer review in the BCMJ as evidence to support the author's concern that driving whilst under the influence of cannabis can cause impairment in driving capability.

The next step in the “telephone game” was when a well-meaning medical reporter read the BCMJ article, trusted the reputation of the BCMJ for authenticity, and made it into a news story. The story created a vision of legions of newly pot-addled drivers creating carnage on the roads just across our border to the South.

Just because something is published in a journal that is held to be reputable, if it has not been peer-reviewed or edited it should be considered gossip until thoroughly checked out. The medical media has no lesser responsibility than the mainstream media in this regard, disclaimer or no disclaimer.

In the rhetoric we are going to face as our governments consider legalization of cannabis for recreational purposes, we as scientists and health care professionals have a responsibility to present the best evidence possible that can help inform the law-makers. We should present this in an unbiased and understandable form so that the media can convey it correctly.

Driving whilst under the influence of cannabis may not be a good idea, but it seems to be at least safer than driving drunk.

Leave a Reply

Your email address will not be published. Required fields are marked *