The Nissan NV200 Is the Only Vehicle Where Drugs Outrank Alcohol in Fatal Crashes

According to the toxicology reports, America has a national drug-to-alcohol ratio in fatal crashes of 0.57. For every driver who tests positive for drugs, 1.75 test positive for alcohol.^[1] That ratio holds across ten years and 307 vehicle models in the FARS database. Split the dataset by vehicle type, though, and the national average disintegrates into two completely different countries.

The Nissan NV200 posts a drug-to-alcohol ratio of 1.08 across 212 fatally-crashed drivers. Drugs literally outrank alcohol: 12.7% tested drug-positive versus 11.8% alcohol-positive.^[1] No other vehicle in the entire FARS dataset crosses the 1.0 line. Amazon bought thousands of these vans for last-mile delivery before switching to the Rivian EDV, FedEx and UPS filled their fleets with them, and independent contractors and gig couriers drove them into the ground. And the toxicology data for the people who crashed them fatally looks nothing like the rest of the American fleet.

The NV200 is not a statistical fluke, because four of the ten most drug-shifted vehicles are commercial or cargo vans: NV200 (1.08), GMC Savana (0.82), Chevrolet Uplander (0.78), Ford Transit Connect (0.77).^[1] These are the gig economy's workhorses, contractor rigs, catering trucks, vehicles whose operators work schedules that punish sleep and reward stimulants, and whose employers screen for alcohol with breathalyzers while ignoring everything else.^[3]

Flip the table and family vans tell the opposite story: Ford Windstar sits at 0.30 and Mazda5 at 0.31, carrying the same FARS vehicle class code and the same "Van" designation in the database but completely inverted impairment profiles. Commercial vans are drug-shifted because their drivers face occupational pressures that select for stimulant use; family vans are alcohol-shifted because their crash demographics skew toward off-duty suburban driving where alcohol dominates. The van class is not one population but two populations wearing the same label, and any fleet safety policy that treats them identically is operating on a fiction.

The luxury segment tells its own story, and it involves a lot of wine. Audi A3 drivers tested alcohol-positive at 22.7%, the highest rate of any vehicle with 200 or more drivers in the FARS dataset, fully 50% above the national average of 15.1%.^[1] Audi Q5 (0.36), Q7 (0.33), and Tesla Model S (0.33 ratio, 19.1% alcohol rate) cluster in the same alcohol-heavy zone. Entry-level European performance vehicles attract a demographic that still drinks and drives at rates most of the American fleet left behind a decade ago. Their drug-to-alcohol ratios hover around 0.33 to 0.36, roughly half the national average, not because their drivers avoid drugs but because their drivers are so reliably drunk that drugs can barely register as a proportion.

What fleet managers should do

If you operate NV200s, Savanas, or Transit Connects, your pre-shift screening protocol should include oral fluid drug testing alongside breathalyzers. SAMHSA-certified oral fluid tests run $25 to $40 per test and detect methamphetamine, cannabis, and opioids that a breathalyzer will never catch.^[2] For individual buyers shopping used delivery vans as work trucks: these vehicles' impairment profiles reflect their prior operators, not some inherent property of the van. Check VINs for open recalls at nhtsa.gov/recalls, and understand that the NV200's elevated drug-positive rate is a workforce story, not a vehicle engineering story.

Limitations

FARS toxicology testing is not uniform across jurisdictions or years. Some states test for drugs in fewer than half of fatal crashes, and drug testing expanded substantially during the 2014 to 2023 window, which inflates drug-positive rates for vehicles whose crashes are concentrated in more recent years.^[1] "Drug-positive" collapses prescription medications and illicit substances into one bucket. The NV200 driver on prescribed blood pressure medication and the NV200 driver on methamphetamine are statistically identical in this dataset. Sample sizes for individual models are modest (NV200 n=212; minimum threshold of 200 applied) and subgroup analysis is not possible at this granularity.

Strongest counterargument

The NV200's drug-to-alcohol ratio of 1.08 may be an artifact of temporal testing bias rather than a genuine workforce signal. NV200 production ran from 2013 to 2021, concentrating its FARS crash data in more recent calendar years when jurisdictional drug testing rates were substantially higher. An older vehicle like the Ford Windstar, whose fatal crash data is weighted toward earlier FARS years when drug testing was uncommon, would naturally appear more alcohol-dominant even if its drivers' actual substance use was identical. This cross-tabulation cannot separate the vehicle-type signal from the testing-era signal, and honest analysis must acknowledge that some fraction of the observed spread reflects measurement expansion, not behavioral difference.