Impact of Work Schedules on Sleep Duration of Critical Care Nurses
A. J. Hirsch Allen, MPH, Julie E. Park, MSc, Nassim Adhami, RN, Demetrios Sirounis, MD, Harriet Tholin, RN, BSN, Peter Dodek, MD, Ann E. Rogers, RN, PhD, Najib Ayas, MD
Disclosures Am J Crit Care. 2014;23(4):290-295.
Abstract and Introduction
Abstract
Background. Sleep deprivation leads to reduced vigilance and potentially impairs work performance. Nurses may work long shifts that may contribute to sleep deprivation.
Objective. To assess how nurses' sleep patterns are affected by work schedules and other factors.
Methods. Between October 2009 and June 2010, a total of 20 critical care nurses completed daily sleep and activity logs and a demographic survey and wore an actigraph to objectively measure sleep time for 14 days.
Results. In a multivariate model with controls for repeated measures, mean sleep time between consecutive work shifts was short: 6.79 hours between 2 day shifts and 5.68 hours between 2 night shifts (P = .01). Sleep time was much greater between days when no shifts were worked (8.53 hours), consistent with catch-up sleep during these times. Every minute of 1-way commuting time was associated with a reduction of sleep time by 0.84 minutes.
Conclusion. Critical care nurses obtain reduced amounts of sleep between consecutive work shifts, particularly between consecutive night shifts. Whether this degree of sleep deprivation adversely affects patients' safety needs further study.
Introduction
Acute and chronic sleep deprivation leads to reduced vigilance, motor discoordination, and impaired cognitive function, thereby increasing the risk of errors in the workplace. Specifically, psychomotor performance after 24 hours of continuous wakefulness is similar to that of persons with a blood alcohol concentration of 0.10%, a level consistent with legal intoxication in many jurisdictions.[1] Increased sleepiness and poorer neurobehavioral performance are associated with less than 5 to 6 hours of time in bed, even for a single night.2 Increasingly, evidence3 indicates that sleep deprivation in health care professionals adversely affects performance and may have adverse effects on patient and occupational safety
Nurses' work shifts in hospitals are usually at least 12 hours long. Frequent work shifts of this length, especially on multiple consecutive days, most likely leave little time for adequate sleep between shifts, especially when time required for commuting and transfer of patient information at shift changes is considered. Additionally, time required for responsibilities at home may further limit opportunities for sleep.[2]Sleep duration may be even less between consecutive night shifts, because the internal biological (circadian) clock reduces sleep drive during the daytime.[4] Although registered nurses are the largest group of health care providers in the United States, the impact of work schedules on sleep patterns of nurses has not been well studied. We think that this topic is an important issue because inadequate sleep duration in nurses could potentially affect their performance and the safety of patients.
The primary purpose of this study was to assess how sleep patterns are affected by work schedules. Specifically, we hypothesized that nurses in the critical care unit obtain inadequate sleep between 2 consecutive work shifts and that this problem would be accentuated between 2 consecutive night shifts. Second, we hypothesized that sleep duration would be reduced by other factors, such as responsibilities at home and commuting time.
Methods
Participants
The study sample consisted of nurses who worked in multidisciplinary intensive care units (ICUs) in 3 different urban hospitals in Vancouver, British Columbia: St Paul's Hospital, Vancouver General Hospital, and Mt St Josephs Hospital. All nurses who worked regularly in these 3 hospitals were eligible to participate in the study unless they were nurse managers or educators. Day shifts start at 7:30 AM, and night shifts start at 7:30 PM. Shifts are 12.5 hours long, because 0.5 hours is needed for handover of patient information. Participants in the study were recruited by means of advertisements posted in the ICUs at these hospitals.
Nurses who participated were asked to complete detailed daily sleep and activity logs plus a 1-time demographic survey; they also wore an actigraph to provide an objective measure of sleep duration. An actigraph is a noninvasive, portable device similar in size to a wristwatch and is worn on the nondominant wrist. The device contains a motion sensor that allows it to differentiate sleep from wakefulness on the basis of the wearer's movement. Actigraphy is a reliable and valid method for measuring sleep in healthy adults[5] and has been used to study sleepwake patterns for more than 30 years.[6] The actigraph was worn continually for 14 days; it was removed only if the device would be exposed to water (eg, showering). Each nurse in the study was given an honorarium of Can$50/wk. The research ethics board of Providence Health Care and the University of British Columbia approved the study.
Individual sleep periods were determined from sleep onset to sleep offset as defined by standard actigraph measures. The data obtained from the actigraphs were also reviewed manually to ensure validity. The individual reviewers of the data were unaware of the nurses' work schedules. Total sleep duration was the sum, in hours, of all sleep periods between shifts, excluding naps.
Survey
The survey elicited information about the nurses' daily activities, such as time spent commuting or attending to family responsibilities, work shifts, work breaks, and mood. In addition, nurses were expected to complete daily logs that elicited information about how many hours they worked at the hospital and what kind of work was done, how physically and mentally tired they were, how stressed they were, how they slept, and when they slept. Participants were also asked, "On a typical day, how many minutes does it take you to go from your home to the hospital where you work?" to assess the impact of commuting time on sleep.
Statistical Analysis
The number of hours each nurse slept was recorded continually for a period of 14 days. Therefore, between consecutive days, 8 different between-shift types were possible:
• day-day (eg, a day shift on Tuesday followed by another day shift on Wednesday)
• day-night (eg, a day shift on Tuesday followed by a night shift starting Wednesday night)
• night-night (eg, a night shift on Tuesday night followed by a night shift on Wednesday)
• day–no shift (eg, a day shift on Tuesday with no night or day shift on Wednesday)
• night–no shift (eg, a night shift starting Tuesday night with no day or night shift on Wednesday)
• no shift–no shift
• no shift–day shift
• no shift–night shift
During the 14 days of the study, 13 between-shift types were derived on the basis of the working schedules recorded by the nurses.
The major purpose of the study was to determine sleep duration between consecutive day and night shifts. First, a univariate model was fitted to estimate the effect of shift type on sleep duration while correcting for repeated measures but ignoring the effect of potential confounders. Linear mixed effect models were used to account for correlation between repeated measures in a single person.
The secondary purpose was to assess the impact of a variety of demographic variables, including age, presence of children at home, commuting time, living arrangement (living alone vs not), and sex, on sleeping time. A univariate regression analysis was performed first to identify a preliminary list of potential confounders (P < .10); thereafter, a stepwise regression approach was taken to build a multivariate model, which provided an estimate of the effect on sleep duration.
The value P < .05 was considered statistically significant except as noted earlier. All statistical analyses were done by using SAS, version 9.2, software (SAS Institute Inc).
Results
Characteristics of Nurses
A total of 20 nurses participated in the study between October 2009 and June 2010. Characteristics of the sample are shown inTable 1 . In general, the nurses were young, and most were female. Median commuting time from home to work (ie, 1-way commuting time) was 27.4 minutes. Of the 20 nurses, 18 worked full-time (≥35 hours per week).
Sleep Duration Between Shift Types
A total of 260 between-shift periods were possible (20 nurses × 13 between-shift periods). After data cleaning, 247 between-shift periods were analyzed. A total of 9 between-shifts periods that were missing were excluded. Also, on the basis of visual inspection of the diagnostic plots, 4 outliers were identified and sensitivity analyses were performed. Because these outliers were not influential points (ie, did not alter the direction or significance of results), they were excluded from the analysis to obtain stable statistical inferences. The largest proportion of shifts (99 of 247; 40%) were between 2 days when no shifts were worked.
Table 2 shows the mean sleep duration by shift type after accounting for repeated measures within each participant. Mean sleep duration between 2 night shifts (5.68 hours) was significantly less than the mean sleep duration between 2 day shifts (6.79 hours). Nurses slept much more between nonworkdays (8.53 hours) and during the night between a day shift and a day off (8.93 hours). Mean sleep duration between night shift and no shift was only 4.84 hours; however, this short sleep duration was attributed to the fact that sleep duration was calculated only until midnight the day after a shift. Nurses who had the following day off, most likely came home and took a nap (rather than a full sleep), knowing they could get a full night's sleep later that evening.
Impact of Other Variables
A set of univariate analyses was done to assess the potential impact of nonshift factors on