← The Crash Report
Sobriety Report

The Cocktail Effect: 1 in 5 Impaired Fatal Crash Drivers Were on Alcohol AND Drugs

Dale Impactor III • Clautomotive Desk
March 28, 2026

FARS doesn't just track whether a fatally-crashed driver was drunk or high. It tracks both. Decompose 490,736 driver toxicology records from 2014 to 2023 into alcohol-only, drug-only, and both-at-once, and a pattern surfaces that standard impairment reporting completely buries.

18.6%
of impaired fatal crash drivers tested positive for both alcohol AND drugs

Nearly one in five impaired drivers in fatal crashes weren't just drunk. Weren't just high. They were both. Simultaneously. That's 18,200 poly-impaired drivers over the decade, derived from the overlap between the 15.0% alcohol-positive rate and 8.6% drug-positive rate within the 20.0% overall impairment figure.[1]

The math: alcohol-positive + drug-positive - any-impaired = poly-impaired. Run that equation across 307 vehicle models, and you get something NHTSA doesn't publish: a per-vehicle poly-impairment breakdown.

The Cocktail Cars

The BMW M5 leads the field at 10.1% poly-impairment among its 109 fatally-crashed drivers. One in ten M5 drivers who died in a crash had both alcohol and drugs in their system. The Porsche Macan follows at 7.5%. Mercedes GLK at 7.0%. Jeep Commander at 7.0%. BMW 7 Series at 6.5%.

But sandwiched between the German metal: the Buick Park Avenue at 9.3% and the Pontiac Bonneville at 8.9%. The Oldsmobile Alero at 7.8%.

Two demographics, same cocktail. Wealthy buyers mixing substances in performance cars they can afford new, and lower-income drivers doing the same in depreciated former luxury sedans they bought for $3,000 off Craigslist. The GM B-body and the BMW M car inhabit opposite ends of the showroom and identical ends of the toxicology report.[2]

The Constant

Across vehicle classes, the poly-impairment share among impaired drivers barely moves. Sports cars: 18.5%. Sedans: 18.7%. Pickups: 18.4%. SUVs: 18.7%. Vans: 18.7%. The National Institute on Drug Abuse flags that "people tend to mix various substances" and that the crash risk from combinations "appears to be greater than that for each drug by itself."[3] FARS confirms: whatever you drive, if you're impaired, there's roughly a 1-in-5.4 chance you're impaired on multiple fronts.

The overall rate does vary by class. Sports cars lead at 4.2% total poly-impairment, vans trail at 3.4%. But the ratio holds. The vehicle changes the likelihood you show up impaired in a fatal crash. It does not change the cocktail probability once you do.

The Luxury Shield

Luxury brand drivers show 21.0% total impairment vs. 20.0% for mass-market. They're more impaired, not less. But their VMT fatality rate averages 0.57 deaths per 100M miles, versus 1.05 for mass-market. Five-star crash structures, pre-collision braking, and 4,500-pound curb weights working as designed.

The M5 driver and the Park Avenue driver share the same cocktail. They do not share the same survival odds.

Sources & References

  1. NHTSA, Fatality Analysis Reporting System (FARS), 2014–2023. Toxicology data for 490,736 drivers in fatal crashes, cross-tabulated by vehicle make and model. nhtsa.gov
  2. NHTSA FARS Query Tool, used to derive alcohol-positive, drug-positive, and any-impaired rates by vehicle. Poly-impairment calculated as: alcohol + drug − any. cdan.dot.gov
  3. National Institute on Drug Abuse, “Drugged Driving DrugFacts.” Notes that poly-substance crash risk exceeds risk from each substance alone. nida.nih.gov
  4. CDC, “Impaired Driving: Get the Facts.” Reports ~44% of fatally crashed drivers tested positive for drugs; 13.5M drove under alcohol influence, 11.7M under drug influence in 2021. cdc.gov

Methodology

Poly-impairment is calculated from FARS toxicology fields: poly = alcohol_positive + drug_positive − any_impaired. This captures the union overlap. Only vehicle models with ≥100 driver toxicology records are included in per-model rankings. Class-level aggregates include models with ≥50 records (490,736 total drivers). The 18.6% figure represents 18,200 poly-impaired drivers divided by 97,800 total impaired drivers. All rates are driver-level, not crash-level.

Limitations

FARS toxicology reporting is not universal. Testing protocols vary by state and jurisdiction. Some states test only for alcohol if BAC is above the legal limit, suppressing drug-positive counts. Drug-positive includes prescription medications, not just illegal substances. THC can persist in blood for weeks after use, so a positive result doesn't confirm impairment at the time of the crash. Small sample sizes for individual models (BMW M5 n=109, Porsche Macan n=107) mean per-model rates carry wide confidence intervals.

Counterargument

The vehicle-level pattern may reflect demographics, not vehicle characteristics. Wealthy buyers of M5s and Macans may have higher substance access and social drinking cultures, while Park Avenue and Bonneville buyers may overlap with populations facing substance abuse challenges. The vehicle is a proxy for the driver, not a cause. Additionally, jurisdictions that test more thoroughly may disproportionately cover areas where certain vehicles are more common, introducing geographic bias into per-model rates.

Source: NHTSA FARS 2014–2023. Poly-impairment derived from union overlap of alcohol and drug toxicology fields. Per-model rankings limited to vehicles with ≥100 driver records. See methodology for caveats.