LIVE Global accident monitoring — updating every 15 minutes
LIVEACCIDENTTRACKER
Home / Articles / Statistics
Statistics

How Many People Die in Drug-Involved Crashes Each Year?

Published 5 min read By Live Accident Tracker Editorial

Quick answer: 10,140 people were killed in drug-involved traffic crashes in the United States in 2024, down 11.8% from 2023, according to Fatality Analysis Reporting System (FARS) annual report file data presented by the University of California's SafeTREC centre. "Drug-involved" in this count means law enforcement reported a driver or motorcyclist to have been under the influence of drugs, which is narrower than a positive toxicology test. Among fatally injured drivers with known drug tests, 46.6% tested positive for legal or illegal drugs.

What drug-involved means, and what it does not

The definition matters, because three different measures are often confused. The 10,140 figure counts crashes where law enforcement reported a driver or motorcyclist to have been under the influence of drugs, as stated on the SafeTREC Traffic Safety Facts page. That is a judgement recorded at the crash, not a laboratory result.

A positive drug test is a different measure. It shows that a substance was present, not that it impaired the driver or caused the crash, and different drugs remain detectable for very different lengths of time. The 46.6% figure applies only to fatally injured drivers whose drug test result was known. It must not be read as 46.6% of all drivers, or as proof that drugs caused those deaths.

SafeTREC notes that driving can be affected by a variety of legal and illegal drugs, including over-the-counter medicines, by slowing reaction time, decreasing coordination, increasing aggressive and reckless driving, impairing cognitive function and causing drowsiness.

The testing gap behind the count

Every official drug-involved total is a floor, not a ceiling, because testing is incomplete and inconsistent. The National Highway Traffic Safety Administration (NHTSA) warned in its March 2022 report on drug testing and traffic safety that national estimates of drug prevalence and positivity built from FARS rested on a dataset where over 60% of the data were missing, which can produce inaccurate national estimates.

Testing practice varies sharply by state. In a report to Congress in April 2024, NHTSA found, citing 2019 FARS data, that the completeness of reporting of drug testing of fatally injured drivers ranged from more than 90% in some states to less than 10% in others, averaging 60% across all states, as summarised by the Governors Highway Safety Association (GHSA). Policies on who is tested, and in what circumstances, vary across states, and results are not always passed to the FARS analyst. NHTSA added new FARS data fields in 2022 for drug specimen, drug test result and drug test status, but the historical counts still carry the gaps of the system that produced them.

The practical consequence is simple: where drivers are rarely tested, the official count records few drug-involved crashes, whether or not drugs were present. Comparisons between states, or over time, must therefore be made with the testing gap in view.

What trauma centres found

A separate line of evidence comes from hospital toxicology screening rather than police crash records. More than half (56%) of people injured or killed in crashes on United States roads tested positive for alcohol and/or drugs in toxicology screening during a study of trauma centres between 2019 and 2021, according to GHSA's summary of NHTSA research. People were most likely to test positive for cannabis (25%), followed by alcohol (23%), stimulants (11%) and opioids (9%). One in five (20%) had two or more categories of drugs in their system.

That study window covers the COVID-19 public health emergency period, and it covers people injured or killed who were screened at trauma centres. It is not an annual death toll, and a positive screen is not proof of impairment at the time of the crash. Its value is different: it shows how often more than one substance is present, and how common cannabis presence had become in the screened group.

Public perception has not kept up. A national survey reported by GHSA found that 93% of people said driving while over the legal blood alcohol concentration limit is very or extremely dangerous, but only 71% believed it is dangerous to drive within an hour of consuming cannabis. GHSA and the AAA Foundation for Traffic Safety note that cannabis affects attention, decision-making, coordination and reaction time, all of which are critical for safe driving.

Keeping the tallies apart

Drug-involved deaths, alcohol-impaired deaths and drug-positive test results must be kept in separate, labelled paragraphs, because they count different things. Adding them together would double-count the same crashes, since a crash can involve both alcohol and drugs, and a single driver can appear in more than one measure.

California illustrates why state figures need their own label. The same SafeTREC page records 693 people killed in drug-involved traffic crashes in California in 2024, down 38.7% from 2023. That state section draws on FARS annual report file data for 2024 plus provisional California SWITRS data for 2024, and it covers all victims, including drivers, passengers, bicyclists and pedestrians fatally or seriously injured in drug-involved crashes in the state. It is not directly comparable with the national definition without noting the extra source.

How these numbers were counted

The national total of 10,140 killed in 2024, the 11.8% fall from 2023 and the 46.6% positive-test share come from FARS annual report file data for 2024, as presented by SafeTREC in its 2026 Traffic Safety Facts edition on drug-involved driving. Drug-involved crashes are defined there as crashes where law enforcement reported a driver or motorcyclist to have been under the influence of drugs. The testing-completeness figures come from NHTSA's 2022 drug-testing report and its April 2024 report to Congress citing 2019 FARS data, as summarised by GHSA. The trauma-centre figures come from NHTSA research covering 2019 to 2021, summarised by GHSA in material last updated in June 2026. FARS annual report file data can be revised, and a positive test result is not by itself proof of impairment or causation.

For related reporting, see United States road data, today's accident reports and drink-driving deaths by state.

In this story

More from United States

All United States articles →

Keep Exploring

More from the tracker

Watch accidents unfold in real time on the live world map, check the 2026 statistics, or browse all articles.