Perceptualand Motor Skills, 1990, 70, 723-730. O Percepmal and Motor Skills 1990

SLEEP REDUCTION I N NIGHT-SHIFT WORKERS: IS IT SLEEP DEPRIVATION O R A SLEEP DISTURBANCE DISORDER? ' ROBERT P. MAHAN

The University of Georgia ANTONIO B. CARVALHAIS AND STEPHEN E. QUEEN

University of Connecticut Summary.-Two explanations of the reported reduction in night-shift workers' sleep length have emerged. One argues that these reductions in sleep are related to sleep dismrbances which may require medical treatment as disorders. The other argues that sleep reduction is mainly due to a sleep deficit caused by the voluntary choice to delay sleep onset and reduce the sleep duration which requires a systems approach aimed at increasing sleep length and reducing the effect of a sleep deficit at the workplace. Using survey data from 1078 industrial shift workers, this smdy demonstrates that permanent night workers reporting sleep dilficulties, as well as those who do not, report sleep reductions. This finding is interpreted as supporting, in part, the position that night-shift workers suffer primarily from a self-induced sleep deficit.

The human difficulties associated with night-shift work are complex and multidimensional in nature. There is a general consensus among shift work researchers that workday sleep length is related to time-of-day worked. Workers on the night shift, on the average, have significantly reduced workday sleep length (Tepas, Armstrong, Carlson, Duchon, Gersten, & Lezotte, 1985). Polysomnographic studies have demonstrated that these reductions in the sleep length of permanent night shift and rotating workers can be associated with decrements in performance (Tepas, Walsh, Moss, & Armstrong, 1981; Tilley, Wilkinson, Warren, Watson, & Drud, 1982). The patterns of sleep staging recorded in these studies are similar to those seen in research subjects undergoing chronic sleep-length reduction (Wilkinson, 1968). Finally, the workday sleep reduction found among night workers is not a transient phenomenon but appears to persist as long as workers are on night duty (Tepas, 1982; Gersten, 1987). Two explanations have generally been proposed for these reductions. One focuses on the temporal placement of the night worker's sleep period during daytime hours being out of phase with the 24-hour cyclic behavior of various physiological functions (Czeisler, Weitzman, Moore-Ede, Zimmerman, & Knauer, 1980). This c h r o n ~ b i o l o ~ i c approach al to night-shift work suggests that workers suffer mainly from insomnia-like symptoms w h c h re'Requests for reprints should be sent to Robert P Mahan, Department of Psychology, The University of Georgia, Athens, GA 30602. The authors thank Dr. Donald Tepas for helpful comments on an earlier draft of the manuscript. We also express appreciation for the expert comments of the editors and anonymous reviewers.

724

R.

P. MAHAN. E T A L

sult when there is a major phase change in the normal sleeplwake cycle (Weitzman, Kirpke, Goldmacher, McGregor, & Nogeire, 1970; Weitzman, 1976). Sleep reductions among night workers are viewed as a physiological response to the inversion of the sleeplwake cycle. This response occurs when night workers attempt to sleep at times contrary to a biological constitution which dictates they should be awake and active. Likely primary symptoms of such a disturbance may be sleep-onset insomnia, frequent waking during the sleep period, and difficulty staying asleep. A second approach which is associated with the sleep-deficit model suggests the sleep reduction among night workers is a multidimensional phenomenon (Tepas & Mahan, 1989). Whereas the chron~biolo~ical explanation emphasizes the actual time-of-day one sleeps, the sleep deficit approach proposes that sleep length is affected by a variety of variables. However, social factors are viewed as a primary determinant of reduced sleep length in permanent night-shift workers. Night workers who sleep during daytime hours are compelled to modify their sleep behavior in an effort to maintain social obligations which are often constrained by the demands of a diurnally oriented society. Night workers are viewed as voluntarily reducing their sleep period to engage in the many activities traditionally bounded by the 9:00 AM to 5:00 PM social daytime schedule. These activities often include (a) spending time with family, (b) participating in day-to-day responsibilities, such as food shopping, food preparation, banking, seeking medical care, and (c) engaging in leisure activities with friends. Environmental factors such as noise, light, and heat may also contribute to reductions in workers' sleep time. If social factors are a significant determinant of sleep reductions among night workers, the symptoms would be mainly those of sleep deprivation, such as reduced sleep latency and increased napping. The sleep-deficit explanation also suggests that the reduction in sleep length associated with nightshift work is not related to the incidence of sleep disturbance. This study examines the sleep length and sleep latency by permanent day- and night-shift workers who either report they frequently experience difficulties staying asleep, falling asleep, or both staying and falling asleep, or report they do not experience these sleep difficulties. Comparisons in sleep length and sleep latency are made between those workers reporting sleep difficulties, and those who do not report these difficulties. If night-shift workers demonstrate a reduced sleep length regardless of the presence or absence of these reported difficulties, it becomes hard to conclude that chronobiological variables are a sole source of this reduction.

Subjects

The data were collected from 1078 hourly workers on fixed, 8-hr. per-

NIGHTSHIFT WORK, SLEEP REDUCTION

725

manent day or night shifts. These data were generated from workers given a confidential Work/Sleep Survey to complete. The over-all response rate for this survey was 90.4%. The Work/Sleep Survey, the population studied, and the methods utilized in collecting these data have been described in previous reports (Tepas, et al., 1985; Gordon, McG111, & Maltese, 1981). The age of the respondents ranged from 18 to 66 yr., with a mean of 34.8 yr. Men represented 66.4% of the respondents. These workers were employed on a discontinuous shift system. They were classified as either day or night workers if they worked at least seven consecutive hours within the following time windows: 0600-1600 for the day shift and 2200-0700 for the night shift. These definitions and criteria have been summarized by Tepas and Monk (1987). The time periods refer to intervals that define permanent fixed-shift day and night workers, i.e., 7 of the 8 hours of a shift must fall within the intervals for classification. This reflects some variance in start and stop times.

Procedure Workers were divided into four groups on the basis of whether or not they reported having sleep difficulties: (a) day-shift workers reporting sleep difficulties (DaylYes), (b) day-shift workers reporting no sleep difficulties (Day/No), (c) night-shift workers reporting sleep difficulties (NightlYes), and (d) night-shift workers reporting no sleep difficulties (Night/No). A worker was considered to have sleep difficulties if he responded 'yes' to each of the three following questions typically used for patient screening in sleep-disorder clinics: (a) "Do you have difficulty falling asleep?", (b) "Do you have difficulty staying asleep?", and (c) "Do you have difficulty falling and staying asleep?" All subjects were evaluated on three additional questions: (a) "About how often do you feel tired or sleepy at work?" which was measured on a six-point scale where 1 represented "Never" and 6 represented "About every day," (b) the yes/no question "Do you often take naps during the work week?" Sleep length was computed from the survey question ashng workers to report the time at which they go to bed and the time they get up during the work week. Tepas, Walsh, and Armstrong (1981) found this method of estimating subjective sleep length correlates 0.83 with laboratory polysomnographic measurements of total sleep time. Sleep latency was determined by the survey question asking workers to estimate how long it took them to fall asleep. Subjective reports of increased sleep latency as an index of sleep disturbance have been partly validated by the finding that workers who report having sleep dfficulties appear to have longer sleep latencies than those who report no sleep difficulties (Stock, 1779). Sleep length and sleep latencies from the four groups were compared using a univariate analysis of covariance, controlling for age and sex on the continuous items, and using nonparametric tests for the categorical items. Age and

726

R. l? MAHAN, ETAL.

sex have been related to sleep length in research on shiftwork. Because sleep length and latency were not correlated, multivariate procedures were not applied to the present data.

The results of the analysis of covariance with the between-subjects independent variable represented by difficulty/no difficulty groups and sleep length as the dependent variable, with sex and age as covariates, indicates that for these workers there are statistically significant differences for sleep length (F,,,,,, = 10.84, p < ,001). An identical analysis of covariance for the dependent variable sleep latency was conducted using the same four-level independent variable, with sex and age as covariates. The analysis indicated statistically significant differences between worker groups on sleep latency (F,,,,,, = 74.55, p < .001). Means and standard deviations for these measures are presented in Table 1. TABLE 1

MEANS AND STANDARD DEVLA~ONS FORSLEEPLENGTH AND SLEEPLATENCY MEASURES AMONG WORKERS'GROUPS Group

n

Sleep Length M SD

Sleep Latency M SD

Scheffe' post hoc tests on workers' sleep length indicated that groups Day/Yes and Day/No had significantly longer sleep length than groups Night/Yes and Night/No ( p < .01). No statistically significant differences on sleep length were seen between groups Day/Yes and Day/No, or between groups Night/Yes and Night/No. ScheffC post hoc tests on workers' sleep latency showed that Day/Yes and Night/Yes workers had significantly longer sleep latencies than Day/No and Night/No workers ( p < .01). Again, no statistically significant differences were found on sleep latency between groups Day/Yes and Night/Yes or between groups Day/No and Night/No. Nonparametric Kruskal-Wallis and Mann-Whitney analyses were computed for categorical dependent variables. Statistical significance among the group differences was found in response to the question "About how often do you feel tired or sleepy at work?" ( X , 2 = 89.70, p

Sleep reduction in night-shift workers: is it sleep deprivation or a sleep disturbance disorder?

Two explanations of the reported reduction in night-shift workers' sleep length have emerged. One argues that these reductions in sleep are related to...
298KB Sizes 0 Downloads 0 Views