Working harder -- working dangerously?

Fatigue and performance in hospitals

Leslie G Olson and Antonio Ambrogetti

MJA 1998; 168: 614-616
For editorial comment, see Holmes
See also Nocera & Khursandi
 

Patients in hospital don't stop requiring medical care at night. That means either overtime or shift work for somebody. "Somebody" is usually a junior doctor. The early postgraduate years are not easy -- the problems of professional responsibility, study and combining a career and a personal life are difficult enough, but to these challenges are added hours of work that society does not expect of any other occupational group. The routine 50- hour shifts of not so long ago have, we believe, disappeared, but hospital doctors still commonly work long hours and unreasonable shifts. What impact this has had in our hospitals -- how it contributes to mistakes, suicides, drop-outs and divorces -- is unknown. However, fatigue undoubtedly impairs professional performance, learning and quality of life. For legal, educational and humanitarian reasons, hospitals need to do whatever can be done to limit excessive hours of work.   A physiological, not an industrial, problem How many hours are "excessive"? We do not know. It depends on what we are trying to achieve, and the answer will be different if we want to maximise teaching and learning than if we just want to avoid disastrous errors. How many hours are doctors working? We do not know that either (Box 1). Most data derive from retrospective self-reports, which is obviously unsatisfactory, and concern average hours per week, which is not the issue. The issue is the existence of any work weeks that do not allow adequate recovery between shifts or which impose excessive periods of continuous work. The Association of American Medical Colleges has suggested that residents should not work more than 80 hours per week averaged over four weeks.1 This approach is irrational -- it is like saying you are fit to drive at any time if your blood alcohol level averaged over four weeks is less than 0.05%. We have practically no data on the frequency with which doctors work exceptionally long hours. Hospitals may have data on rostered hours, but this is only a portion of the total hours worked, and the information is not entirely accurate. Rosters change because swaps are unregulated, and unrostered overtime is usually not included in hospital estimates of hours worked. Time sheets would be a better source of data, but their use raises confidentiality problems, and they miss unpaid overtime. Unrostered overtime, whether obligatory or self-imposed, is unregulated and mostly unrecorded because the pressure to work and not complain is overwhelming. The only accurate source of data on total hours worked would be direct observation, and no such study has been done. It is critical that this is seen not as an industrial issue but as a physiological problem. An 80-hour work week including 40 hours of unpaid overtime and no days off, however unacceptable industrially, could be acceptable physiologically if the worker had eight hours' sleep between shifts. A 24-hour shift at double-time with paid meal breaks is not acceptable physiologically, however attractive it may be to a junior doctor with a mortgage. Imposing a 24-hour shift once a year because all the other medical staff are off sick may not be unfair, but it is unsafe.   The body clock keeps ticking at night Humans have a marked circadian preference for sleep at night and, even under optimal conditions, being awake at night is associated with impaired performance. When long hours and sleep deprivation are added to the circadian problem, the performance deficit is exacerbated. The effects of fatigue on performance are well defined.2 Concentration, data processing and short-term memory are impaired. The variability of performance increases, so that normal performance alternates with periods of poor work, and astute decisions are mixed with lapses of judgement. Performance declines sharply as the duration of a task increases, and fatigued workers sacrifice accuracy to speed. Fatigue causes less performance decrement in workers with more control over their work because they can schedule non- urgent tasks for periods when they are at their best. Doctors will thus cope better than staff with less job flexibility, such as nurses. Effort can compensate for fatigue, but as fatigue worsens the ability to summon an effort of concentration declines and the time for which it can be maintained shortens. Performance in crises, however, is preserved until fatigue is extreme. Mistakes caused by fatigue are most likely to occur during routine tasks and tasks which require sustained vigilance, especially when the factors that trigger an effort of concentration (such as an obviously ill patient) are absent. Fatigue mistakes characteristically involve failure to recognise the existence of a serious problem. Giving the wrong antibiotic to a patient recognised as having sepsis is not a typical fatigue mistake, but failing to recognise sepsis at all is. For this reason fatigue-related errors of judgement are difficult to prevent and often disastrous. Are doctors seriously impaired by fatigue? We are not sure. Some studies have found performance decrements attributable to fatigue,3 but others have not.4 The methodological issues that make most existing studies of limited use are twofold. Firstly, the psychometric tests usually used to assess fatigue have never been shown to predict real-life medical performance.2 Secondly, many studies use unrealistic definitions of fatigue. In one often-quoted study of junior doctors with work weeks of 100 hours, "fatigued" was defined as less than four hours' sleep in 24 hours, and "rested" as more than four hours' sleep in 24 hours.4 It is little wonder that no difference was discernible between the "fatigued" and "rested" doctors! Doctors and army officers are the only occupational groups reported to be unaffected by fatigue,2 but this is generally regarded as an artefact of methodology.2,5 A "healthy worker effect" is a possible explanation (people who need a lot of sleep don't last very long as surgical registrars or army cadets), but not needing a lot of sleep is hardly the best basis for choosing doctors. Willingness to work when fatigued is widely seen as "professional". This view goes back to acts of genuine heroism in wars and epidemics and to some of the noblest traditions of medicine. But the diurnal rhythm of alertness and the drive to sleep are basic physiological processes, and commitment does not affect physiology. Heroic workloads are out of place in the routine organisation of the urban teaching hospital. Many doctors, junior and senior, voluntarily undertake long hours of work for reasons of money, professional advancement or altruism. Neither money nor seniority is an antidote to fatigue, however, and fatigue is no less damaging to performance when it is incurred voluntarily. It is just as inappropriate for a consultant surgeon to perform emergency operations all night and a routine list the next morning as it is for a junior doctor to be forced to stay at work to assist. Hospitals must not only stop forcing staff to work dangerous hours, they must also prevent them choosing to do so.   Hours ain't hours Given that night work is inevitable, and that it inevitably impairs performance and quality of life, it seems obvious to insist on the use of minimally damaging patterns of shift work. The reality in hospitals is different: a number of practices known to cause particularly severe impairment are common. The most obvious suspicion about shifts is true: long ones are worse than short ones.6 Data from a number of occupational groups suggest that eight-hour night shifts are optimal and that 12-hour night shifts are acceptable only if workloads are light. Night shifts longer than 12 hours, and daytime shifts longer than 16 hours, have consistently been found to be associated with reduced productivity and more accidents.6 It is also important to note that the later the night shift ends, the less sleep is obtained that day.7 Even when the night shift has officially ended, casual extension of night shifts into the morning for hand-over rounds and morning report is common. In addition, libraries and medical administration offices rarely have opening hours that allow night workers access without intruding on their sleep time. Physiological adaptation to night work is largely a myth8 and there is no reason to extend periods of night work in the hope that adaptation will occur. For intellectually demanding tasks, short periods of night work (one or two shifts) are better tolerated than longer periods because the accumulated sleep deficit is less. It is easy to demonstrate that the progressive sleep loss of a seven-night roster causes a progressive rise in accidents and a fall in productivity.9 Sleep deficits cannot be repaired immediately. Night-shift workers seldom sleep more than five or six hours in each 24- hour period, so that after seven nights the accumulated sleep deficit is 15 to 20 hours. At least 48 hours off duty are usually needed to recover this deficit,7,10 and rosters that require workers to go from night shifts to day or afternoon shifts with no break are dangerous. The pattern of long work days followed by nights interrupted by calls is regarded as normal by the medical profession. These calls are, at best, simple telephone calls. At worst, they may entail returning to the hospital several times a night. Few other workers do this, and data on its long-term effects are entirely lacking. Being on-call impairs sleep even when there are no calls,11 and as five or six hours' uninterrupted sleep is needed to avoid performance decrement10 perpetual or frequent on-call work is probably damaging.   Solving the problem We have very little of the knowledge that we need to understand fatigue in medical work. Studies are required to determine whether there is a serious problem of fatigue-related impairment of performance in some or all Australian hospitals. Such studies need to focus on how learning and quality of life, as well as professional performance, are affected by fatigue. Junior doctors are not the only ones affected by fatigue, and senior staff also need to be studied. We need controlled trials to determine optimal work schedules and to test countermeasures to fatigue (eg, benzodiazepines to assist daytime sleep, and bright light and caffeine for stimulation). We do know enough, however, to eliminate the worst abuses of the human sleep-wake cycle (Box 2), and we need to see a shift by both hospital employers and the medical profession towards addressing this issue.   References Executive Council, Association of American Medical Colleges. Resident supervision and hours: recommendations of the Association of American Medical Colleges. J Med Educ 1988; 63: 417-426. Dinges DF, Kribbs NB. Performing while sleepy: effects of experimentally induced sleepiness. In: Monk TH, editor. Sleep, sleepiness and performance. Chichester: Wiley, 1991: 97-128. Rubin R, Orris P, Lau SL, et al. Neurobehavioural effects of the on-call experience in housestaff physicians. J Occup Med 1991; 33: 13-18. Deaconson TF, O'Hair DP, Levy NW, et al. Sleep deprivation and resident performance. JAMA 1989; 260: 1721-1727. Leung L, Becker CE. Sleep deprivation and house staff performance. J Occup Med 1992; 34: 1153-1160. Rosa RR. Extended work shifts and excessive fatigue. J Sleep Res 1995; 4 (Suppl 2): 51-56. Kecklund G, kerstedt T. Effects of timing of shifts on sleepiness and sleep duration. J Sleep Res 1995; 4 (Suppl 2): 47-50. Dahlgren K. Long-term adjustment of circadian rhythms to a rotating shiftwork schedule. Scand J Work Environ Health 1981; 7: 141-151. Knauth P. Speed and direction of shift rotation. J Sleep Res 1995; 4 (Suppl 2): 41-46. Carskadon M, Roth T. Sleep restriction. In: Monk TH, editor. Sleep, sleepiness and performance. Chichester Wiley, 1991: 155-167. Torsvall L, Akerstedt T. Disturbed sleep while being on call. An EEG study of apprehension in ships engineers. Sleep 1988; 11: 35-38.

Authors' details

Department of Medicine, University of Newcastle, Newcastle, NSW.
Antonio Ambrogetti, MD, FRACP, Sleep and Respiratory Physician.

Reprints will not be available from the authors.
Correspondence: Dr L G Olson, Department of Medicine, University of Newcastle, John Hunter Hospital, Locked Bag 1, Hunter Region Mail Centre, NSW 2310.