Health-interested internet readers were shocked by the headline from popular medical website WebMD, which declared, “Fatal Car Crashes Involving Pot Use Have Tripled.” The lead paragraph warns how marijuana legalization is becoming popular, and to be wary of its “dark side.”  “Fatal crashes involving marijuana use tripled during the previous decade,” writes WebMD, citing a new study from Columbia University's Mailman School of Public Health.

The researchers took a look at crash statistics from six states -- California, Hawaii, Illinois, New Hampshire, Rhode Island and West Virginia -- including over 23,500 drivers who died within one hour of a crash between 1999 and 2010. In both 1999 and 2010 they found alcohol contributed to about 40% of all fatality crashes. But in 1999, drugs contributed to 16% of fatality crashes, compared to 28% in 2010. Worse, marijuana rose from contributing to 4% of all fatality crashes in 1999 to 12% in 2010 -- that’s the “triple” they are referring to.

Sounds pretty dramatic, doesn’t it? Triple the dead drivers from pot wrecks in a decade? If you go by WebMD’s article you’d think we’d be overwhelmed by stoned mayhem on the freeways by now. Yet, it seems to me that driving now is a lot like driving a decade ago.  Could WebMD be deceiving us?

The first hint you’re being deceived is to notice the weasel words. Those are words you can use to weasel out of being pinned down to facts when you report on science. Notice it is “Crashes involving marijuana use” and not “marijuana causing fatality crashes.” That’s because all the data is telling us is that marijuana is detected in the blood of drivers who died within an hour of a crash. Notice no mention of whether crashes overall have risen, either.

If one bothers to Google the actual study, “Trends in Alcohol and Other Drugs Detected in Fatally Injured Drivers in the United States, 1999 - 2010”, one finds the researchers issuing some very important caveats that WebMD’s editors seemed uninterested in including: “First, the study is based on data from only six states where this information is captured,” the authors wrote, “Secondly, the effects of drugs on driving performance and crash risk vary by drug type, dosage, and the driver’s physiological response and tolerance level.” In other words, detecting marijuana in drivers’ blood, particularly in California, may just be netting you medical marijuana patients who always have it in their blood and have developed quite a tolerance that allows them to drive unimpaired.

The final caveat WebMD chose to omit, “Also, it is possible for a driver to test positive for marijuana in the blood up to one week after use,” is the one that deflates the scary statistic. These marijuana-positive dead drivers may just as well be innocent unimpaired drivers who smoked a joint a couple of days ago and got hit by some drunk driver.

All that a tripling of detection of marijuana in the blood of dead drivers tells us is that more people are smoking marijuana more often, not that marijuana is causing them to crash. We can verify this by consulting the National Survey on Drug Use and Health. In 1999, there were an estimated 8.7 million adults 18 and older who had smoked pot in the prior month. By 2012, there were 17.2 million, almost double the number of monthly tokers. If we break it down to people who toke 21 days or more per month, we’ve gone from 2.1 million to 6 million, almost triple. Among daily tokers, we’ve increased from 1.1 million to 3.5 million, more than triple.

So, with all this tripling of marijuana use, what has happened to traffic fatalities? They’re down, nationally and in almost every state, to their lowest recorded levels. The Fatality Analysis Reporting System (FARS) of the National Highway Traffic Safety Administration (NHTSA) shows us that in 1999, there were 37,140 fatal crashes, while in 2012, there were 29,757 - the first year there have been fewer than 30,000 traffic deaths in America. Measured in deaths per hundred million miles traveled (1.55 to 1.10), deaths per 100,000 population (15.30 to 10.39), deaths per 100,000 vehicles (19.61 to 12.57), and deaths per 100,000 licensed drivers (22.29 to 15.28) are at the lowest levels ever.

As for those six states studied by Columbia? Five of the six have fewer fatalities -- less than one per 100,000 miles -- than the national average of 1.1. West Virginia far exceeds the national average at 1.78, but even that has declined 14.4% since 1999; the rest of the states’ fatality rates declined more than that percentage.

The lead author, Dr. Li, added, “it is important to interpret the prevalence of non-alcohol drugs reported in this study as an indicator of drug use but not necessarily as a measurement of drug impairment.” Dr. Li suggests the expansion of drug testing for drivers, but even NHTSA says to this day, “It is difficult to establish a relationship between a person's THC blood or plasma concentration and performance impairing effects.  It is inadvisable to try and predict effects based on blood THC concentrations alone.”

"Radical" Russ Belville is the host of The Russ Belville Show, which airs live at 3pm Pacific.