A pilot study of the safety implications of Australian nurses' sleep and work hours

Chronobiol Int. 2006;23(6):1149-63. doi: 10.1080/07420520601059615.


The frequency and severity of adverse events in Australian healthcare is under increasing scrutiny. A recent state government report identified 31 events involving "death or serious [patient] harm" and 452 "very high risk" incidents. Australia-wide, a previous study identified 2,324 adverse medical events (AME) in a single year, with more than half considered preventable. Despite the recognized link between fatigue and error in other industries, to date, few studies of medical errors have assessed the fatigue of the healthcare professionals involved. Nurses work extended and unpredictable hours with a lack of regular breaks and are therefore likely to experience elevated fatigue. Currently, there is very little available information on Australian nurses' sleep or fatigue levels, nor is there any information about whether this affects their performance. This study therefore aims to examine work hours, sleep, fatigue and error occurrence in Australian nurses. Using logbooks, 23 full-time nurses in a metropolitan hospital completed daily recordings for one month (644 days, 377 shifts) of their scheduled and actual work hours, sleep length and quality, sleepiness, and fatigue levels. Frequency and type of nursing errors, near errors, and observed errors (made by others) were recorded. Nurses reported struggling to remain awake during 36% of shifts. Moderate to high levels of stress, physical exhaustion, and mental exhaustion were reported on 23%, 40%, and 36% of shifts, respectively. Extreme drowsiness while driving or cycling home was reported on 45 occasions (11.5%), with three reports of near accidents. Overall, 20 errors, 13 near errors, and 22 observed errors were reported. The perceived potential consequences for the majority of errors were minor; however, 11 errors were associated with moderate and four with potentially severe consequences. Nurses reported that they had trouble falling asleep on 26.8% of days, had frequent arousals on 34.0% of days, and that work-related concerns were either partially or fully responsible for their sleep disruption on 12.5% of occasions. Fourteen out of the 23 nurses reported using a sleep aid. The most commonly reported sleep aids were prescription medications (62.7%), followed by alcohol (26.9%). Total sleep duration was significantly shorter on workdays than days off (p < 0.01). In comparison to other workdays, sleep was significantly shorter on days when an error (p < 0.05) or a near error (p < 0.01) was recorded. In contrast, sleep was higher on workdays when someone else's error was recorded (p = 0.08). Logistic regression analysis indicated that sleep duration was a significant predictor of error occurrence (chi2 = 6.739, p = 0.009, e beta = 0.727). The findings of this pilot study suggest that Australian nurses experience sleepiness and related physical symptoms at work and during their trip home. Further, a measurable number of errors occur of various types and severity. Less sleep may lead to the increased likelihood of making an error, and importantly, the decreased likelihood of catching someone else's error. These pilot results suggest that further investigation into the effects of sleep loss in nursing may be necessary for patient safety from an individual nurse perspective and from a healthcare team perspective.

MeSH terms

  • Adult
  • Australia
  • Female
  • Humans
  • Male
  • Medical Errors
  • Middle Aged
  • Nurses*
  • Nursing Staff, Hospital
  • Occupational Health*
  • Personnel Staffing and Scheduling
  • Pilot Projects
  • Regression Analysis
  • Sleep
  • Sleep Disorders, Circadian Rhythm / diagnosis
  • Work
  • Work Schedule Tolerance*
  • Workload