Drowsy driving statistics vary sharply by measurement method. Police reports, surveys, naturalistic video, and meta-analyses capture different parts of the problem, so the figures below keep their periods, populations, geographies, and limitations explicit.
Key Drowsy Driving Statistics
The most broadly useful figures include:
- In the United States in 2017, 91,000 police-reported crashes involved drowsy drivers, according to a NHTSA estimate.
- In the United States in 2017, about 50,000 people were injured in estimated drowsy-driver crashes.
- In the United States in 2017, nearly 800 people died in estimated drowsy-driver crashes.
- During 2017–2021 in the United States, drowsy driving was reported in 1.8% of fatal crashes.
- In the United States in 2021, 684 people were killed in crashes involving a drowsy driver.
- In the United States in 2021, those 684 deaths represented 1.6% of all motor-vehicle traffic fatalities.
- In the CDC’s 2009–2010 survey of 19 states and the District of Columbia, 4.2% of 147,076 adults reported falling asleep while driving during the prior 30 days.
- In that 2009–2010 CDC analysis, men reported falling asleep while driving at 5.3%, compared with 3.2% of women.
- In the same analysis, adults aged 65 or older reported drowsy driving at 1.7%.
- In a New York State licensed-driver survey, 54.6% had driven while drowsy during the previous year.
- In the CDC’s 2011–2012 survey of 10 states and Puerto Rico, 9.1% of respondents sleeping five hours or less reported falling asleep while driving.
- In that 2011–2012 survey, 2.7% of respondents sleeping at least seven hours reported falling asleep while driving.
- In the same survey, snorers reported falling asleep while driving at 5.6%, compared with 2.9% of non-snorers.
- A 2017 international meta-analysis found a pooled odds ratio of 2.51 for motor-vehicle accidents associated with sleepiness at the wheel.
- A Fiji case-control study found an adjusted odds ratio of 5.7 for injury-involved crashes among sleepy or not-fully-alert drivers.
- A meta-analysis of adult working populations found an odds ratio of 2.36 between obstructive sleep apnea syndrome and car accidents.
- Among U.S. civilian workers in 2010, 30.0% reported sleeping six hours or less per day.
- In a U.S. naturalistic driving study, PERCLOS classified 8.8%–9.5% of examined crashes as involving drowsiness.
Contents
- Key Drowsy Driving Statistics
- Drowsy Driving Crash, Injury, and Death Statistics
- Drowsy Driving Prevalence in Driver Surveys
- Drowsy Driving Trends by Age, Sex, Race, Work, and State
- Sleep Duration and Drowsy Driving Data
- Drowsy Driving Crash Risk and Injury Outcomes
- Sleep Apnea, Shift Work, and Commercial-Driver Statistics
- International Evidence and Measurement Limits
Drowsy Driving Crash, Injury, and Death Statistics
NHTSA estimated 91,000 police-reported crashes, about 50,000 injured people, and nearly 800 deaths involving drowsy drivers in the United States in 2017. NHTSA cautions that police and hospital data undercount drowsiness because sleepiness is difficult to identify conclusively: NHTSA, “Drowsy Driving: Avoid Falling Asleep Behind the Wheel”.
More recent police-coded figures are lower than modeled estimates, illustrating why the measures should not be combined as though they were the same series.
| Measure | Period and geography | Result |
|---|---|---|
| Fatal crashes with reported drowsy driving | United States, 2017–2021 | 1.8% |
| Deaths in drowsy-driver crashes | United States, 2021 | 684 |
| Share of all traffic fatalities | United States, 2021 | 1.6% |
These observed figures come from NHTSA, “Countermeasures That Work: Drowsy Driving—Data/Surveillance”. Fatal-crash coding depends on reported evidence of drowsiness.
NHTSA’s research-plan range attributes 2%–20% of annual U.S. traffic deaths to driver drowsiness, a wide modeled range rather than one observed count. An AAA analysis of 1999–2013 estimated drowsiness contributed to 6% of all crashes and 21% of fatal crashes; applying that older modeled share to 39,508 fatal crashes in 2021 implies more than 8,300 deaths, but this is a derived illustration, not a NHTSA fatality count.
The same source is NHTSA, “Countermeasures That Work: Drowsy Driving—Data/Surveillance”.
Drowsy Driving Prevalence in Driver Surveys
The CDC’s optional sleep module covered 19 states and the District of Columbia in 2009–2010. Among 147,076 adults, 4.2% reported falling asleep while driving during the prior 30 days; the unadjusted prevalence was 4.1%, and the measure was self-reported rather than nationally representative.
The same CDC analysis found different reported rates by sex and age:
- Men reported falling asleep while driving at 5.3%, compared with 3.2% among women.
- Adults aged 18–44 reported drowsy driving at more than 4.9%.
- Adults aged 65 or older reported drowsy driving at 1.7%.
Source and scope: CDC, “Drowsy Driving — 19 States and the District of Columbia, 2009–2010”. These are age-adjusted survey associations and do not establish causes.
An older telephone survey of licensed drivers in New York State found that 54.6% had driven while drowsy during the previous year. In the same survey, 22.6% had ever fallen asleep at the wheel without a crash, 2.8% had ever crashed after falling asleep at the wheel, and 1.9% reported ever crashing while driving drowsy; the latter concepts may overlap.
The study is “The scope and nature of the drowsy driving problem in New York State”, and its abstract does not state the survey year.
Drowsy Driving Trends by Age, Sex, Race, Work, and State
The CDC’s 2009–2010 BRFSS results show variation across demographic and employment categories, but not a time trend or exposure-adjusted risk ranking.
| CDC subgroup | Reported drowsy driving |
|---|---|
| Non-Hispanic Black respondents | 6.1% |
| Hispanic respondents | 5.9% |
| Other race/ethnicity | 6.0% |
| Non-Hispanic White respondents | 3.2% |
| Employed respondents | 5.1% |
| Unable to work | 6.1% |
| Retired respondents | 1.0% |
| Students or homemakers | 2.1% |
| Unemployed respondents | 3.1% |
The race and employment categories are age-adjusted self-reports; “other race/ethnicity” is heterogeneous, and small subgroups may have wider uncertainty. The source is CDC, “Drowsy Driving — 19 States and the District of Columbia, 2009–2010”.
State prevalence in the participating jurisdictions ranged from 2.5% in Oregon to 6.1% in Texas. These state results are not a national estimate and should not be read as a change over time.
Sleep Duration and Drowsy Driving Data
Sleep duration is a related exposure, not an interchangeable definition of drowsy driving. In the CDC’s 2011–2012 survey of 10 states and Puerto Rico, reported falling asleep while driving varied by usual sleep duration.
| Reported sleep duration | Reported falling asleep while driving |
|---|---|
| Five hours or less | 9.1% |
| Six hours | 5.2% |
| At least seven hours | 2.7% |
The seven-hour category is a comparison category, not a claim that one duration is optimal for every person. Source: CDC, “Drowsy Driving and Risk Behaviors — 10 States and Puerto Rico, 2011–2012”.
The CDC also reported that respondents with insufficient sleep on all 30 days had higher drowsy-driving prevalence than those reporting no insufficient-sleep days. A short-sleep threshold of six hours or less was independently related to drowsy driving in the 2009–2010 analysis, while a CDC-linked 2011–2012 analysis found increased modeled probability for five-hour-or-less and six-hour sleepers versus seven-hour sleepers when sleep was always perceived insufficient.
Snoring was another symptom proxy: 5.6% of snorers reported falling asleep while driving, compared with 2.9% of non-snorers. Snoring is not a diagnosis of obstructive sleep apnea, and all of these survey measures are observational and self-reported.
Drowsy Driving Crash Risk and Injury Outcomes
A 2017 systematic review and meta-analysis pooled the association between sleepiness at the wheel and motor-vehicle accidents at OR 2.51, with a 95% confidence interval of 1.87–3.39. The review included 10 cross-sectional studies, six case-control studies, and one cohort study, with I-squared heterogeneity of 83%, so the pooled estimate summarizes varied observational evidence rather than an absolute probability.
A population-based case-control study of four-wheel vehicles in Fiji reported an adjusted OR 5.7 for injury-involved crashes among sleepy or not-fully-alert drivers, with a 95% confidence interval of 2.7–12.3. Reporting fewer than six hours of sleep in the prior 24 hours was associated with OR 5.9, with a 95% confidence interval of 1.7–20.9.
The Fiji study estimated population-attributable risks of 34% for crashes associated with sleepy or not-fully-alert driving and 9% for crashes associated with less than six hours of prior-24-hour sleep. Those estimates depend on local exposure prevalence and assumptions and are not directly transferable to U.S. drivers.
Sleep Apnea, Shift Work, and Commercial-Driver Statistics
A meta-analysis of adult working populations found obstructive sleep apnea syndrome associated with OR 2.36 for car accidents, with a 95% confidence interval of 1.92–2.91. The pooled accident odds ratio was 2.80 among commercial motor-vehicle drivers with OSA and 2.32 among noncommercial drivers; sleep-disordered breathing more broadly was associated with OR 1.81.
Short sleep was common among U.S. workers in 2010, but it is not itself a crash measure:
- 30.0% of civilian employed U.S. adults reported sleeping six hours or less per day, representing an estimated 40.6 million workers.
- Short-sleep prevalence ranged from 24.1% to 41.6% across industries.
- Night-shift workers reported short sleep at 44.0%, compared with 28.8% among day-shift workers.
- The estimated number of night-shift workers with short sleep was about 2.2 million, compared with about 28.3 million day-shift workers.
Source: CDC, “Short Sleep Duration Among Workers — United States, 2010”. These figures describe sleep exposure, not the number of drowsy drivers.
International Evidence and Measurement Limits
A large-scale AAA naturalistic study monitored more than 3,500 U.S. drivers for several months and examined 701 crashes with pre-crash video data. Using PERCLOS eye-closure classification, it estimated that drowsiness contributed to 8.8%–9.5% of examined crashes and 10.6%–10.8% of crashes involving airbag deployment, significant property damage, or injury.
In the study’s crash table, 19.0% of 153 crashes occurring in darkness involved a PERCLOS-classified drowsy driver, compared with 6.1% of 408 daylight crashes. These are sample-specific subgroup estimates, not population night-driving rates.
The measures across this evidence base are not directly interchangeable: police-coded crashes, modeled estimates, self-reported survey prevalence, PERCLOS video classifications, and clinical sleep-disorder associations describe different constructs. Odds ratios are not absolute crash probabilities, short sleep and snoring are risk indicators rather than drowsy-driving definitions, and the CDC surveys are jurisdiction-limited optional modules rather than national prevalence studies.