Study objectives: To quantify the prevalence of self-reported near-miss sleepy driving accidents and their association with self-reported actual driving accidents.
Design: A prospective cross-sectional internet-linked survey on driving behaviors.
Setting: Dateline NBC News website.
Results: Results are given on 35,217 (88% of sample) individuals with a mean age of 37.2 +/- 13 years, 54.8% women, and 87% white. The risk of at least one accident increased monotonically from 23.2% if there were no near-miss sleepy accidents to 44.5% if there were > or = 4 near-miss sleepy accidents (P < 0.0001). After covariate adjustments, subjects who reported at least one near-miss sleepy accident were 1.13 (95% CI, 1.10 to 1.16) times as likely to have reported at least one actual accident as subjects reporting no near-miss sleepy accidents (P < 0.0001). The odds of reporting at least one actual accident in those reporting > or = 4 near-miss sleepy accidents as compared to those reporting no near-miss sleepy accidents was 1.87 (95% CI, 1.64 to 2.14). Furthermore, after adjustments, the summary Epworth Sleepiness Scale (ESS) score had an independent association with having a near-miss or actual accident. An increase of 1 unit of ESS was associated with a covariate adjusted 4.4% increase of having at least one accident (P < 0.0001).
Conclusion: A statistically significant dose-response was seen between the numbers of self-reported sleepy near-miss accidents and an actual accident. These findings suggest that sleepy near-misses may be dangerous precursors to an actual accident.