Driving Drowsy

On a hot, humid July morning in 1986, five 17-year-olds took a 3-hour drive from the suburbs of Cincinnati into southern Kentucky to go caving. They had stayed up for most of the prior night, making plans for the day ahead, putting batteries in flashlights, filling canteens with water for their underground hike, and watching movies. As the sun came up, they all felt tired, but were not about to let a little problem like lack of sleep interfere with their adventure. They got on the road, and after a couple of hours driving, they found themselves traveling along a deserted, 55-mile-per-hour two-lane county highway, among bluegrass covered limestone hills. Coming around a curve in the road, the driver, bleary-eyed from the long night and the long drive, failed to notice an old brown Chevrolet that was stopped dead in the road, preparing to make a left turn. Slamming on the brakes too late, he smashed into the back of the Chevrolet at high speed, locking his right knee, dislocating his patella, and tearing ligaments on impact. The front passenger, who had taken off his shoulder belt for a moment to reach down and retrieve an audiocassette from the floor, put his face through the windshield. The three remaining passengers escaped the crumpled wreck of the vehicle with a sprained wrist, a sprained ankle, and somehow, no injury at all; the driver of the Chevrolet, the only person in the other car, was also unhurt. I witnessed this scene unfold from an outstanding, unforgettable, terrifying perspective, with the sound of screaming metal and splintering glass loud in my ears. I was the passenger with the sprained ankle.

Drowsy driving is a problem of staggering, widely underrecognized scope. In approximately 56,000 crashes per year and 1,550 fatalities, fatigue is cited by police officers as the cause of the crash,¹ but these statistics likely underestimate the problem. Data on sleep are not routinely gathered at crash scenes in many states, and even in states that do gather these data, drowsiness is typically deemed the cause only when all other causes have been excluded. The recently completed 100-Car Naturalistic Driving Study, in which 100 vehicles were equipped for a year with multiple cameras and sophisticated sensors, found that drowsiness was a contributing cause in 22% of all motor vehicle crashes (MVCs) and 16% of all near-crashes, a number far in excess of that suggested by police reports.² Making a very rough extrapolation from this figure to the annual number of MVC fatalities, as many as 8,000 fatal crashes might occur each year in the U.S. due to drowsiness. Importantly, teenagers and young adults—especially young males³—are disproportionately represented in drowsy driving crashes.

In this issue of the Journal of Clinical Sleep Medicine, Danner and Phillips⁴ address the relationship between drowsy driving and a sleep-promoting intervention that has received considerable attention with respect to academic performance: the delay of school start times. Prior research has demonstrated that early school start times are associated with decreased sleep,⁵ and consequently there has been ongoing debate about whether delaying school start times could lead to improvements in adolescents' school performance. The current investigation expands upon this discussion by considering not just school performance, but the safety of adolescents under different schedules.

Coincident with the implementation of a delay in school start times in one county in Kentucky, the authors found that teen motor vehicle crash rates fell 16.5%, while crash rates concurrently increased in the rest of the state. Students' self-reports of sleep were gathered for one year before and one year after the change, demonstrating increases of 12–30 minutes per night across grade levels. The data on sleep extension provides a plausible biologic mechanism that might explain the decrease in crash rates. That said, the study has important limitations, including (a) the reliance on non-validated self-report to determine sleep times; and (b) the absence of data directly linking individuals' motor vehicle crashes with sleep patterns, as the data were gathered from disparate sources without crosslinking, longitudinal identifiers. Consequently, causality cannot be established. Nevertheless, the finding is very provocative, suggesting that further research into the effects of school start time are merited, with a focus on safety as well as academic outcomes. As motor vehicle crashes are the leading cause of death for adolescents, and as drowsiness is emerging as a leading cause of MVCs, any intervention that appears to significantly decrease drowsy driving could have important public health implications.

Ultimately, development of effective, evidence-based policy regarding drowsy driving and interventions to mitigate its effect will require the collection of high-quality longitudinal data. Research is needed to further explore the role of school start times in drowsy driving crashes across multiple counties and states, that addresses the principal question raised by Danner and Phillips' study: can a causal chain be definitively established between school hours, sleep hours, and MVC risk? If so, then policy makers will need to decide how best to address this risk. Should local and state governments prioritize delays in school start times as a means of addressing drowsy driving crashes, independent of the questions regarding start time and academic performance? In part, the answer to this question will depend upon a consideration of how school start time interventions compare with alternative approaches to reducing drowsy driving crashes, such as increasing the driving age, implementing educational programs, prohibiting night driving by teenagers, or enforcing sanctions for drowsy driving. To date, we still know little about the relative costs and consequences of these various approaches. It is becoming clear, however, that drowsy driving is a problem of sufficient magnitude to deserve an investment of research and policy efforts to reach definitive conclusions.

Twenty-two years ago, my friends and I were lucky. Oblivious to the dangers of driving drowsy, we nearly lost our lives, remarkably close to the place where Danner and Phillips later conducted their thought-provoking study. Fortunately, we all recovered from that crash, with the help of splints, casts, and an enormous number of sutures in the case of my friend who went through the windshield. It could have been far worse. Danner and Phillips' study is an early step towards understanding and addressing the problem of drowsy driving, and reducing the chance that the next group of teenagers does not suffer a more tragic outcome.

DISCLOSURE STATEMENT

Dr. Landrigan has indicated no financial conflicts of interest.