CNN claims "a new report found" that "driving under the influence of drugs was deadlier in 2015 than driving while drunk." That is not what the new report found, and CNN's confused, alarmist story is a good illustration of how not to cover this subject.
The report, an updated version of a guide originally published by the Governors Highway Safety Association in 2015, summarizes data from the U.S. Department of Transportation's Fatality Analysis Reporting System (FARS). In 2015, the introduction notes, FARS "reported that drugs were present in 43% of the fatally-injured drivers with a known test result, more frequently than alcohol was present." That comparison was the basis for CNN's assertion that drugs are now "deadlier" than alcohol and for its headline, which says "drugged driving surpasses drunken driving among drivers killed in crashes." But neither claim is supported by the FARS data, because saying "drugs were present" is not the same as saying they impaired the driver or played any role in the crash.
CNN reporter Robert Jimison would have realized that if he had read the report a little more carefully. "FARS records only drug presence, not drug concentrations analogous to BAC [blood alcohol concentration] levels for alcohol," notes the author, traffic safety researcher James Hedlund. In fact, the goal of establishing "drug concentrations analogous to BAC levels for alcohol" has proved persistently elusive. On that point, Hedlund quotes the "generally accepted" conclusion of the National Highway Traffic Safety Administration (where he used to work): "At the current time, specific drug concentration levels cannot be reliably equated with a specific degree of driver impairment." A 2015 report from the Government Accountability Office likewise noted that "identifying a link between impairment and drug concentrations in the body, similar to the 0.08 BAC threshold established for alcohol, is complex and, according to officials from the Society of Forensic Toxicologists, possibly infeasible."
In short, FARS does not collect information about drug concentrations, which in any case cannot be reliably related to impairment. "The relations between a drug's presence in the body, its concentration, measured in blood, breath, saliva or urine, and its impairing effects are complex and not understood well," Hedlund writes. "A drug may be present at low levels without any impairing effects. Some drugs or metabolites may remain in the body for days or weeks, long after any impairment has disappeared. In particular, marijuana metabolites can be detected in the body for weeks after use." That last point is crucial to understanding the data cited by CNN, since marijuana is by far the most commonly detected illegal drug, and "the FARS marijuana codes do not distinguish clearly between the active impairing component THC and various inactive and non-impairing metabolites." Even when active THC is detected, the level may be too low to affect driving ability.
Despite all those caveats, Jimison repeatedly conflates drug "presence" with impairment, calling drivers who test positive "drugged" and saying they were "under the influence of drugs." Although that mistake is common in reporting on "drugged driving," Jimison really should have known better. Not only did Hedlund's report clearly explain why the premise of CNN's story is wrong; sources Jimison interviewed (including Hedlund, who observes that "drug impairment is a complicated topic") tried to warn him away from claiming that illegal drugs pose a bigger highway hazard than alcohol does.
"Some safety experts caution that drunken driving remains a bigger problem," Jimison writes in the second sentence of his story, right after claiming that "driving under the influence of drugs was deadlier in 2015 than driving while drunk." Unless something very unusual was happening in 2015, those statements are blatantly inconsistent. Later in the story, Russ Rader, senior vice president for communications at the Insurance Institute for Highway Safety, tells Jimison "there's no question that alcohol remains our biggest highway safety problem." That can't possibly be true if, as Jimison informs us, other drugs account for more highway fatalities.
If CNN's interpretation is wrong, what conclusions can we safely draw from the FARS numbers? The share of fatally injured drivers who tested positive for drugs other than alcohol rose from 28 percent in 2005 to 43 percent in 2015, which suggests (consistent with survey data) that drug use rose during that period, although we don't know the extent to which that increase translated into a larger number of dangerously impaired drivers. Marijuana (including inactive metabolites and nonimpairing THC levels) was detected in 36 percent of the drug-positive drivers, amphetamine in 9 percent. The largest category was "other," a hodgepodge of 78 substances, including prescription and over-the-counter medicines as well as illegal intoxicants, that together were detected in 55 percent of the drug-positive drivers.
Assuming the number of marijuana-impaired drivers has increased during the last decade or so, how big a threat do they represent? Hedlund summarizes the evidence on that point, including laboratory studies that show marijuana undermines driving ability, although not as dramatically as alcohol does. Measuring the impact of such impairment in the real world is tricky. A landmark NHTSA study published in 2015 found no statistically significant difference in crash risk between THC-positive and THC-negative drivers once the data were adjusted for confounding variables (age and sex in particular). Other estimates of the increase in crash risk attributable to marijuana intoxication range from 22 percent to 92 percent, either of which qualifies as "a slightly increased risk," as Hedlund notes. By comparison, a BAC of 0.10 percent (which used to be the cutoff for driving under the influence in the U.S.) is associated with a 400 percent increase in crash risk.
Marijuana's impact on traffic safety is obviously relevant in evaluating the effects of legalization, assuming that legalization increases consumption and some of that consumption occurs within an hour or two of driving. If stoned drivers are simply added to drunk drivers, traffic fatalities could rise, especially if there is an increase in the number of drivers who are simultaneously stoned and drunk—a dangerous combination, since the two drugs have a synergistic effect on driving ability. But if stoned drivers replace drunk drivers, the net result could be fewer traffic fatalities, since alcohol impairs driving ability much more than marijuana does.
These issues are worth discussing and investigating. Sloppy coverage like CNN's obscures distinctions that are essential to that task.