1 in 5 Impaired Drivers Who Die in Crashes Are on Alcohol AND Drugs. The Breathalyzer Catches Half.

According to the toxicology reports, and there are 490,736 of them in FARS, 18,336 drivers who died in fatal crashes between 2014 and 2023 tested positive for both alcohol and at least one other drug. That is 18.6% of all impaired fatally-injured drivers. Not alcohol alone. Not drugs alone. Both, simultaneously, in the same bloodstream, at the moment of impact.^[1]

A peer-reviewed study published in Accident Analysis & Prevention tracked the trajectory: polysubstance prevalence in fatally-injured drivers climbed from 14.9% in 2018 to 18.8% in 2022, a 38% increase in odds over four years. Cannabis was the most common co-intoxicant at 25%, followed by stimulants at 19.2% and narcotics at 8.3%. Nighttime polysubstance drivers were three times more likely than daytime.^[2]

Run those numbers against the enforcement infrastructure. Approximately 10,000 Drug Recognition Experts work across the entire country, officers with specialized training to identify drug impairment through a 12-step behavioral evaluation. Washington, D.C. has six. Nationwide, the DRE program evaluates roughly 50,000 drivers per year. FARS logged 98,348 impaired drivers in fatal crashes over the same decade, or about 9,835 per year. Five-to-one throughput sounds adequate until you consider that fatal crashes represent roughly 1% of all impaired-driving incidents.^[3]

Meanwhile, the standard enforcement tool, the roadside breathalyzer, detects exactly one substance: ethanol. It cannot detect cannabis. It cannot detect methamphetamine. It cannot detect fentanyl. When a polydrug driver blows a 0.06 (under the legal limit) but is also on 30 ng/mL of THC (six to fifteen times the legal limit in states that set one), the breathalyzer says "pass" and the officer has to decide whether to call in a DRE or wave the driver through.

I cross-tabulated FARS toxicology against vehicle models with at least 150 drivers in the dataset and 20 or more impaired. Polysubstance rates vary by a factor of eight depending on what the dead driver was sitting in. Pontiac Bonneville: 40.5%. Acura TLX: 31.5%. Mercedes GLK-Class: 30.8%. At the bottom: Mercury Sable at 5.0%, Tesla Model S at 6.1%, Fiat 500 at 7.1%.^[1]

Vehicle class barely moves the needle. Sports cars lead at 4.2% polysubstance share, pickups and sedans cluster around 3.7-3.8%, vans trail at 3.4%. A 0.8-percentage-point class spread versus an 8x model spread tells you that the vehicle badge is a proxy for driver demographics, not a causal factor. Bonneville owners skew older, lower-income, Midwestern, and male. Fiat 500 owners don't. Same crash database, same decade, radically different chemical profiles at autopsy.

A Wright State University study of 246 deceased drivers in Montgomery County, Ohio found 41.9% tested positive for THC, with an average concentration of 30.7 ng/mL. Ohio legalized recreational cannabis in December 2023. Pre- and post-legalization prevalence was 42.1% versus 45.2%, a difference that didn't reach statistical significance.^[4] Cannabis prevalence in crash fatalities stayed flat through legalization. What grew was the combination: alcohol plus cannabis, alcohol plus stimulants, alcohol plus opioids.

One counterargument deserves acknowledgment: FARS toxicology measures substance presence, not impairment level. Someone who used cannabis five days before the crash could still test positive. That peer-reviewed study addressed this by requiring BAC at or above 0.08 for its alcohol threshold, but the raw FARS data counts any alcohol-positive result. This means some fraction of the 18,336 polysubstance drivers may have had clinically insignificant levels of one or both substances. Fair enough. But even if you discount the count by half, you're left with 9,000 confirmed polydrug fatalities over a decade, growing at 38% in odds, against an enforcement system built around a single-substance detector.

NTSB has recommended lowering the BAC limit from 0.08 to 0.05, a change that would catch more alcohol-impaired drivers at the margin. It would not detect a single additional milligram of THC, methamphetamine, or fentanyl. Policy debate is still calibrating the breathalyzer while toxicology reports document a multisubstance problem that a breathalyzer is structurally unable to address.^[5]

What to do about it: If you drive, understand that the legal BAC limit is a floor, not a ceiling of safety. Combining alcohol with any other substance, even legal cannabis or prescribed medication, compounds impairment nonlinearly. If you're prescribed opioids or benzodiazepines, your physician should have told you not to drive; if they didn't, ask. For policymakers: roadside oral fluid testing can detect cannabis, amphetamines, opioids, and benzodiazepines in under ten minutes. Several states are piloting it. technology exists. Deployment doesn't.