|
Article Excerpt
INTRODUCTION
I'm beyond the stage where I need a bed, or even to lie down.... My mind clicks on and off, as though attached to an electric switch with which some outside force is tampering. I try letting one eyelid close at a time while I prop the other open with my will. But the effort's too much. Sleep is winning. My whole body argues dully that nothing, nothing life can attain, is quite so desirable as sleep. My mind is losing resolution and control.
--Charles Lindbergh, 9 hr into his historic flight (Lindbergh, 1953, p. 233)
Concern with the negative effects of fatigue is paramount in aviation (Rosekind, Gander, et al., 1994), nuclear power (Baker, Olson, & Morisseau, 1994), mining (Duchon, Keran, & Smith, 1994), the military (Neville, Bisson, French, Boll, & Storm, 1994), health care (Veasey, Rosen, Barzansky, Rosen, & Owens, 2002), and other settings in which personnel must perform over extended periods. Sleep loss stemming from extended operations can lead to impaired performance, and fatigue has been implicated in accidents such as the grounding of the Exxon Valdez (National Transportation Safety Board, 1989), the 1995 New York City subway train collision (National Transportation Safety Board, 1996), and the American International Airways DC-8 aircraft accident at Guantanamo Bay, Cuba (National Transportation Safety Board, 1994). In industries in which the tolerance for error is low and the consequences for error are high, the problem of fatigue is of significant importance.
Because of the scope of this problem, a considerable amount of research has examined the effectiveness of various interventions that might reduce fatigue. These interventions, referred to as fatigue countermeasures, include naps (e.g., Bonnet & Arand, 1995), bright lights (e.g., Thessing, Anch, Muehlbach, Schweitzer, & Walsh, 1994), caffeine (e.g., National Research Council, 2001), and activity breaks (Neri et al., 2002). Although there is general agreement that naps may be a useful fatigue countermeasure, there is less consensus on how naps should be managed as an effective operational strategy. In a pioneering study on the effects of in-flight naps on aviation flight crews, Rosekind, Graeber, et al. (1994) found that naps benefited performance but had little impact on subjective ratings of alertness. Moreover, they raised questions they were not able to answer conclusively as to whether naps of longer or shorter duration may be more effective, over what postnap period the performance effects remain, and whether sleep inertia (the grogginess experienced after waking) is a significant concern. Several narrative reviews of the efficacy of naps as a fatigue countermeasure have been published (e.g., Bonnet, 1990; Caldwell, 1997; Dinges & Broughton, 1989; Gillberg, 1985; Rosekind et al., 1995). However, it is a difficult task, at the narrative level, to integrate the results of disparate empirical studies and derive specific guidelines for operational use. To address these questions, we conducted a meta-analysis of the efficacy of naps as a fatigue countermeasure.
THE EFFICACY OF NAPS AS A FATIGUE COUNTERMEASURE
A meta-analytic integration of the research literature on the efficacy of naps as a fatigue countermeasure can accomplish two primary objectives. First, it can provide a very specific and precise summary of the overall effects within this research domain. The collective wisdom seems to be that naps can reduce but not reverse the effects of sleep loss (e.g., Bonnet, 1990; Dinges & Broughton, 1989; Gillberg, 1985; Rosekind et al., 1995). However, to date, there has been no precise empirical specification of whether naps exert a strong, moderate, or weak effect in reducing the effects of fatigue. Therefore, one goal of the present effort was to specify the precise significance and magnitude of the efficacy of naps as a fatigue countermeasure.
A second goal of this analysis was to examine the extent to which the efficacy of naps increases or decreases as a function of certain theoretically relevant and practically important moderators. Researchers have noted that naps may be more or less effective depending on how they are implemented (e.g., Bonnet, 1991; Gillberg, 1985; Rosekind et al., 1995). Several specific questions regarding the efficacy of naps as a fatigue countermeasure can be derived from the narrative reviews cited previously, questions that can be subjected to empirical scrutiny with meta-analysis. This capacity for meta-analysis to account for systematic variability in effect sizes, and to render precise tests of the effects of theoretically relevant and practically important moderators, can be one of its greatest contributions to the understanding of a phenomenon (Mullen, Driskell, & Salas, 1998). These questions are addressed in the following.
The effects of naps on performance versus fatigue. Is there a difference between the effects of naps on performance and the effects of naps on feelings of fatigue? Johnson, Freeman, Spinweber, and Gomez (1988) reported that the largest correlation between measures of performance and measures of feelings of fatigue was a modest r = .18. This suggests that naps may exert an impact on one outcome measure (e.g., performance) but not necessarily exert an equivalent impact on another outcome measure (e.g., fatigue). Gillberg (1985, p. 86) went so far as to speculate that naps would exert a noticeable effect on fatigue even though performance might remain relatively unaffected. Thus another goal of this study was to examine the differential effects of naps on performance and on fatigue.
The effects of nap duration. Is there a significant effect of the duration of naps on the efficacy of naps? There is some controversy in the research literature regarding the effects of nap duration on nap efficacy. On one hand, some scholars (e.g., Gillberg, 1985, p. 85) have argued that the duration of a nap does not seem to be critical. On the other hand, some scholars (e.g., Bonnet, 199 l, p. 313) have argued that the beneficial effects of naps vary as a direct linear function of nap duration. Thus another goal of this study was to examine the effects of nap duration on the efficacy of naps as a fatigue countermeasure.
The effect of postnap interval. Is there a significant effect of the postnap interval on the efficacy of naps? There seems to be a consensus that the postnap interval is important; however, the precise nature of the effect of the postnap interval remains open to speculation. Some scholars (e.g., Naitoh, Englund, & Ryman, 1983; Rosekind et al., 1995) have argued that the beneficial effects of a nap dissipate as the postnap interval increases, whereas other scholars (e.g., Bonnet, 1991) have argued that the beneficial effects of a nap can continue to provide benefits perhaps for as long as a 54-hr postnap interval. Thus another goal of this study was to examine the effects of postnap interval on the efficacy of naps as a fatigue countermeasure.
The effect of circadian rhythm. Is there a significant effect of circadian rhythm on the effectiveness of naps? As several authors have suggested (e.g., Bonnet, 1990; Gillberg, 1985), variation across studies in the effectiveness of naps on performance might be a function of natural variation in performance efficiency as a function of the time of day, reflecting the effects of circadian rhythm. Thus another goal of this study was to examine the effects of circadian rhythms on the efficacy of naps as a countermeasure.
The possibility of sleep inertia. Rosekind et al. (1995) and others have voiced concerns regarding conditions under which naps may compromise operator safety. Sleep inertia refers to a period of disorientation and performance decrement that may occur immediately upon waking from a nap. Although Rosekind, Graeber, et al. (1994) found no evidence of sleep inertia in their study of planned cockpit rests in flight operations, others have provided estimates of the effects of sleep inertia ranging from 15 min (Bertelson, 1979) to 20 min (Taub, 1979) to 35 min (Dinges, Orne, Evans, & Orne, 1981). Thus one goal of this study was to investigate the possibility of detrimental effects of sleep inertia.
A META-ANALYTIC INTEGRATION
Procedure
In general terms, every available study testing the effects of naps as a fatigue countermeasure was obtained, and the specific test or tests of the effects of naps were extracted. Using all of the standard literature search techniques, an exhaustive search was conducted for studies examining the effects of naps. Specifically, on-line computer searches (using MEDLINE, PsycINFO, and the Defense Technical Information Center's STINET) were conducted using the keywords nap(-s, -ping), rest, sleep, and fatigue. This computer search was supplemented by ancestry approach searches (examining the reference lists of retrieved studies), descendancy approach searches (locating subsequent studies that cited retrieved studies in Science Citation Index and Social Science Citation Index), and scanning the past 25 years of leading biomedical and behavioral research journals. (See Mullen, 1989, for a discussion of literature search techniques.) Studies that were available as of May 2004 were eligible for inclusion in this integration.
In order to be included in the present effort, a study had to provide one or more clear and unequivocal tests of the effects of naps as a fatigue countermeasure in a normal, nonclinical adult population. Specifically, participants had to be described as actually sleeping during the prescribed nap period, not merely resting. Studies had to provide sufficient information for the accurate reconstruction of a precise test of the effects of a nap of a specific duration on a measure of performance or fatigue at some specified time after the participants awoke from the nap. Typically, as accurately characterized by previous reviewers of this literature (Bonnet, 1990; Dinges, Orne, Whitehouse, & Orne, 1987), the performance or fatigue of individuals was compared with their own baseline performance prior to the extended operation. In other words, participants were tested at baseline, kept awake for a period of time, then allowed a nap, and then tested again. Thus all included studies employed within-subjects pretest-posttest designs.
A number of studies that have been cited in support of the efficacy of naps as a fatigue countermeasure were not included in the meta-analytic database, for the following reasons. Several studies provided insufficient data to allow the extraction of statistical tests of the effects of naps (e.g., Daiss, Bertelson, & Benjamin, 1986; DeValk, DeGroot, & Cluydts, 2003; Dinges et al., 1987; Gillberg, 1984; Hartley, 1974; Moses, Lubin, Naitoh, & Johnson, 1978; Naitoh & Angus, 1987; Schweitzer, Muehlbach, & Walsh, 1992; Webb, 1987). Thus any studies with reports that did not allow the reconstruction of a precise statistical test could not be included in the present effort. It should be noted that in an effort to obtain the most complete meta-analytic database, attempts were made via letter to contact the authors of those studies that provided incomplete or insufficient data for extraction of statistical tests. We requested information regarding 29 studies and were able to obtain additional statistical information on 3 additional studies, which were incorporated into the meta-analytic database.
Some studies often cited in...
|
