|
Correction
to
Mitler et al.,
|
| |||||||||||||||||||||||||||||||||||||||||
The study has additional weaknesses that limit the conclusions that can be drawn. Volunteer drivers were neither randomly assigned to the four driving-schedule groups nor rotated through each group. Thus, the drivers were not necessarily comparable. For example, the mean ages of the two Canadian driver groups were significantly lower than those of the U.S. "steady-day" group, on the basis of 95 percent confidence intervals for the means,2 notwithstanding the authors' statement that "drivers were selected so that the average age of the drivers on each schedule was similar." Another shortcoming is that the observers who rated the video recordings of the drivers' faces for drowsiness were aware of the time of day and ambient light level for each segment. If the observers expected that the drivers would be drowsier at night, they may have been more likely to consider the drivers drowsy on video recordings made at night. Although the investigators report measures taken to ensure consistent scoring of polysomnographic data, including the rescoring of randomly selected records, they report no similar reevaluations of randomly selected video segments.
Elisa R. Braver, Ph.D.
Janella F. Pantula, Ph.D.
Insurance Institute for Highway Safety
Arlington, VA 22201
References
The acknowledgment does not mention that funds were also contributed by the American Trucking Associations' Trucking Research Institute. In 1992, the American Trucking Associations Foundation announced its intention "to raise $476,000 to fund its share of the project."1 The Federal Highway Administration's briefing package for the study stated that the Trucking Research Institute "is funding part of the data collection and analysis."2 Support from the Trucking Research Institute is noted in the Federal Highway Administration's report describing the same study.3
The Trucking Research Institute is a research affiliate of the American Trucking Associations Foundation, a nonprofit arm of the American Trucking Associations, which is a trade association for the trucking industry. Failure to acknowledge the support of the Trucking Research Institute is especially troubling because the Federal Highway Administration is in the process of revising its rules that limit driving hours for interstate truck drivers and has described this study as providing "a technically sound basis for evaluating the current hours-of-service requirements for commercial motor vehicle operators."4
The trucking industry has a financial interest in any changes in work-hour limits (some American Trucking Associations affiliates have called for lengthening the permissible driving period from 10 hours to 14 hours) and clearly has a financial interest in this study. Table 4 of the article shows that drivers on 13-hour schedules had more than twice as many video-recorded episodes of drowsiness (identified on the basis of drooping eyelids and a bobbing head) as drivers on 10-hour schedules.
Susan P. Baker, M.P.H.
Johns Hopkins School of Public Health
Baltimore, MD 21205-1996
References
Several lines of evidence point to inadequate sleep as a major culprit in truck crashes. A survey of tractor-trailers in four states indicated that 10 to 23 percent of drivers had fallen asleep while driving in the previous month.1 The National Highway Traffic Safety Administration (NHTSA) cited drowsiness and fatigue as factors in 15 percent of single fatal truck crashes and estimated that in 1990, 31 percent of 182 large-vehicle crashes that were fatal to the drivers were attributable to the drivers' fatigue.2
Large mass, very high mileage, and long operational life mean that tractor-trailer trucks, in particular, have the highest involvement in fatal crashes, despite an overall low risk of total crashes per miles traveled. When speed limits were raised from 90 to 100 km per hour (56 to 63 mph) in Israel, deaths from truck crashes increased by 60 percent, contributing to 40 percent of the additional road toll.3 A recent NHTSA study found that 67 percent of drivers with irregular schedules had been involved in fatigue-related crashes, as compared with 38 percent of drivers with regular schedules. The most important predictive factor in a fatal crash was the duration of the driver's most recent sleep period and the total hours slept.4
It is time to enforce reduced speed limits for trucks and mandatory rest periods for truck drivers, particularly at night. Eliminating incentive premiums and providing higher salaries and reasonable work conditions are essential steps in reducing fatigue-related fatal truck crashes.
Paul Barach, M.D., M.P.H.
Massachusetts General Hospital
Boston, MA 02114
Gerry Ben David, Ph.D.
Jerusalem College of Technology
Elihu Richter, M.D., M.P.H.
Hebrew University
Jerusalem 91120, Israel
References
David A. Benaron, M.D.
Stanford University School of Medicine
Stanford, CA 94304
Michael J. Green, M.D.
Pennsylvania State University College of Medicine
Hershey, PA 17033-0850
References
To the Editor: We thank Braver and Pantula for the opportunity to clarify the formula we used to compute episodes of drowsiness as evaluated on the video recordings of the drivers' faces. It appears that the numerator was based on 6-minute samples every half-hour because the following sentence was omitted: "When drowsiness was noted, the video was viewed from 30 minutes before to 30 minutes after the initially detected episode, and scored as 10 6-minute epochs." This procedure allowed us to investigate the duration of episodes of drowsiness and ultimately resulted in 1989 six-minute segments judged as showing drowsy drivers. The procedure also resulted in the viewing of most of the video recording, which is why the denominator is the universe of segments from which the 1989 observations of drowsiness were drawn. The resulting percentage was expected to be a slight underestimate, because it was unlikely that all the segments showing drowsiness were identified. With the overall unbiased estimator, based on the every-half-hour samples, the result was 8.1 percent (476/5862); the reported 6.8 percent (1989/29,310), which was based on the six-minute segments, was 1.3 percentage points lower. With this method, there were enough observations to permit the detailed breakdown of the data according to time of day and driving schedule, as shown in Table 4 of our article.
In reply to Baker: we did not acknowledge a number of important organizations that were helpful to our project, including the American Trucking Associations, Transport Canada, the Canadian Trucking Association, the Teamsters Union, and local drivers' organizations. We did not mean to obscure their involvement in the project. However, our direct financial support came from the cited contracts, including one with the Trucking Research Institute, which is, as Baker points out, affiliated with the American Trucking Associations. Through contract number DTFH61-90-C-053, which governed operations among the Federal Highway Administration, Trucking Research Institute, Essex Corporation, and Scripps Clinic, the Trucking Research Institute contributed about $240,000 under a cost-sharing agreement with the Federal Highway Administration. This accounted for about 5 percent of the total cost of the study, which was $4.45 million.1
With respect to Baker's observation that Table 4 of our article shows that drivers on 13-hour schedules had more than twice as many video-recorded episodes of drowsiness as drivers on 10-hour schedules, this table also shows (in the denominators) that drivers on 13-hour schedules had twice as much night driving as drivers on 10-hour schedules. It is not evident which is more important: time of day or time on the job. However, we addressed this question using additional statistical analyses (nonparametric methods, graphic analysis, mathematical modeling, and canonical correlation). We found that the prevalence of drowsiness was strongly related to the time of day but not significantly related to the number of hours spent on the job.1
We thank Barach and colleagues for their comments. There are indeed many sources of evidence that justify studying the extent of, and ways to reduce, fatigue in truck drivers.
Merrill M. Mitler, Ph.D.
Scripps Clinic and Research Foundation
La Jolla, CA 92037-1205
C. Dennis Wylie, B.A.
Wylie and Associates
Goleta, CA 93117
References
| |||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
