the longest hours (Committee of Interns and Residents, 2002a; Owens et al., 2001; Silberger et al., 1988). Despite recommendations that work shifts not exceed 24 consecutive hours, many interns and residents remain subject to call schedules requiring duty periods of up to 36 consecutive hours or longer on weekends (Czeisler et al., 2002; Leonard et al., 1998; Owens et al., 2001). Other residents opt to work 60–84 consecutive hours (Friday or Saturday morning through Monday afternoon) in a single “power weekend” each month (Czeisler et al., 2002).

The work hours of resident physicians have been the subject of research and frequent debate over the past 20–25 years. Although errors made by a sleep-deprived resident in a New York City hospital are believed to have caused a patient’s death, few studies have shown a direct link between fatigue and patient safety (Asken and Raham, 1983; Friedman et al., 1971; Parker, 1987; Poulton et al., 1978). The findings of Smith-Coggins and colleagues are typical. Emergency room physicians working at night reported feeling significantly more sluggish, less motivated, and less clear-thinking than when working days. Although, they were able to maintain their accuracy in interpreting 12-lead electrocardiograms (ECGs) and rhythm strips, their reactions times were slower and they took longer to intubate a mannequin when working the night shift (Smith-Coggins et al., 1997).

Only a few studies have demonstrated that clinical performance is adversely affected by sleep deprivation. Unlike earlier studies, recent studies have been tightly controlled. Earlier methodological flaws (e.g., tests that were too short or tested factual knowledge, which is relatively insensitive to sleep deprivation; included performance incentives; or, most significantly, failed to control for the residents’ actual sleep schedules prior to and during the studies) (Weigner and Ancoli-Israel, 2002) have been corrected. Researchers no longer expect to find differences between “rested” residents—e.g., those who had more than 4 hours of sleep (Bartle et al., 1988; Deaconson et al., 1988; Light et al., 1989), more than 5 hours of sleep, (Hawkins et al., 1985; Reznick and Folse, 1985), or “regular” sleep (Denisco et al., 1987; Storer et al., 1989), or were not on call (Orton and Gruzelier, 1989)—and “fatigued” residents. They assume all residents have a significant sleep deficit, even those tested when not on call (Weigner and Ancoli-Israel, 2002).

Several studies have shown impaired performance on measures of alertness and concentration, standardized tests of creative thought processes, and cognitive performance on a standardized computerized test battery after on-call periods ranging from 24 hours to an entire weekend (Leonard et al., 1998; Nelson et al., 1995; Wesnes et al., 1997). In studies using virtual-reality simulations, surgical residents made more errors and were slower to complete electrocoagulation of bleeding tissue as sleep loss increased



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement