Intended for healthcare professionals


Fatigue and risk: are train drivers safer than doctors?

BMJ 2017; 359 doi: (Published 13 November 2017) Cite this as: BMJ 2017;359:j5107
  1. Paul Greig, honorary senior clinical research fellow,
  2. Rosamund Snow, postdoctoral research assistant
  1. University of Oxford—Nuffield Department of Clinical Neurosciences, John Radcliffe Hospital, Oxford OX2 6HG, UK
  2. Correspondence to: paul.greig{at}

As evidence about the effects of fatigue grows, Paul Greig and Rosamund Snow argue that approaches from other industries where safety is critical should be translated into medical culture

Key messages

  • Fatigue is a major risk factor for mistakes and poor decision making in the workplace

  • Individual workers are very poor at assessing their own fatigue risk

  • Attitudes to breaks and long hours among healthcare workers would be considered unprofessional and illegal in other workplaces where safety is critical

  • No evidence shows that clinical workers are better able to withstand fatigue than those in other industries

  • You can assess and reduce your risk profile and support colleagues to do so

The working hours of clinicians spark much debate around the world, with regulators seeking to balance the risks of staff becoming fatigued against the need to maintain throughput and keep care costs down. A recent national survey of anaesthetic trainees found that fatigue continues to be a hazard,1 and the topic was debated this year in The BMJ.2

Good evidence shows that tired workers are more likely to make errors of judgment, react slowly, misinterpret data, omit key information, and fail to question things that are unsafe.345 Doctors have historically worked long hours, and in some countries shift lengths are still lightly regulated.6 European Working Time Regulations perhaps go the furthest towards managing fatigue risk, but they focus on individual shift length and averaged total working hours. Doctors can opt out of elements of the regulations, and the guidelines allow practitioners to discount certain kinds of work, such as training, private practice, and commuting time.7 Other industries where safety is critical place much more importance on the cumulative effects of fatigue, considering working patterns and careful management of breaks when assessing risk.

This places doctors in a more dangerous position than personnel in other industries where safety is critical such as aviation, public transport, or haulage. Doctors have undergone different training and have different experiences to pilots and bus drivers, but they share the same basic underlying physiology. The question is not whether evidence shows that industrial data can be adapted to medicine, but rather, where is the evidence that medical training makes doctors differentially susceptible?

Effects of fatigue

Fatigue is a multisystem disorder with both physical and cognitive effects. It has been widely investigated in observational trials in both workplace and laboratory settings. Cognitive effects of fatigue include poor concentration, motor incoordination, slowed response times, memory impairment, and more frequent lapses in attention.34 Fatigued people make more errors, are injured more frequently at work,8 and can be involved in (or cause) major safety incidents.3 Fatigue increases risk taking behaviour.9 Uninterrupted breaks of 15 minutes every few hours can overcome reductions in performance due to fatigue; people may need more frequent breaks during night work.4

Lessons from industry

Industrial data (largely from mining and manufacturing industries) show that working a 10 hour shift is associated with a 13% higher risk of incident or injury, compared with eight hour shifts, and working a 12 hour shift increases this risk by 27%.4 The degree of cognitive impairment after 17 hours awake is equivalent to that associated with a blood alcohol concentration above the drink driving limit for most of Europe.4 Risk is cumulative: it rises with each subsequent night shift, up to sixfold when comparing the fourth consecutive night with the first.8 Age is an important factor in determining resilience to fatigue; objective performance measures, such as reaction time, decline more steeply from the age of 35 upwards.8 Although people reliably recognise that they are tired, self assessment of impaired performance is poor.10

Fatigue is the largest identifiable and preventable cause of safety incidents in all modes of transportation.11 Crew members’ fatigue is routinely considered when investigating incidents; clinical investigators have said that “if the same analysis were applied to accidents involving the care of patients in teaching hospitals, fatigue on the part of clinicians would almost always be cited as a contributing factor.”12

Fatigue in medicine

Observational studies of junior doctors have shown that fatigue has measurable effects on performance measures including response time13 and diagnostic ability.14 Fatigued trainees are more likely to sustain needlestick injuries.15 In a survey of 1366 doctors working ≥40 hours a week, more than two in five junior doctors reported making errors at work related to fatigue.16 A randomised study in two intensive care units (each having 10 beds) found a significant decrease in errors after reducing shift length and permitted working hours per week.5 The effects of acute sleep deprivation seem to vary between people,15 and data directly linking fatigue to major adverse outcomes such as mortality are lacking; some studies report improved safety with restricted duty hours,517 whereas others show little or no benefit.18 The risks of fatigue extend beyond the hospital, and several studies have found more road traffic incidents among doctors at the end of duty shifts.119 Data on burnout are conflicting, and we don’t know whether staff retention improves when fatigue is managed better.

Few studies of healthcare workers compare shifts shorter than 12 hours’ duration, unlike those of transport workers, and almost none look at the effects of taking breaks. This includes recent major publications, such as the FIRST trial.20 This trial randomised US surgical training programmes to restricted or more flexible duty hours. It showed that more flexible hours led to non-inferior patient outcomes. Restricted shifts were 16 (postgraduate year 1) or 28 (year 2 and above) hours per shift, with 8-10 hours between shifts, and not more than 80 hours a week. The flexible group were allowed to deviate from all these limits. Notably, the restricted shifts in this trial exceed the European Working Time Regulations. If risk is already increased at 12 hours, then studies comparing shifts of 16 and 28 hours merely compare fatigued staff with more fatigued staff.

Shorter shifts require more frequent handovers, and some evidence indicates that this may increase error rates, but participants’ baseline level of training on handover has not been considered in this research.21 Reassuringly, the FIRST trial did not find a safety disadvantage associated with the increased handovers required of restricted duty hours.

Not all medical work is the same. Anaesthetists, for example, often have the highest workloads at the start and end of a case, but maintaining the anaesthesia can be less physically and mentally demanding. Despite this they must stay alert and vigilant throughout. Such periods of low tempo are likely to impair their ability to remain alert.22 The surgeon involved in the same case will be much more dependent on psychomotor skills, which, although affected by sleep limitation,1723 may be more resistant to fatigue than cognitive skills. The nature of the task is also relevant. Maintaining vigilance is easier in crisis situations than in routine procedures,22 but much medical work relies on safe working outside emergency situations.

An inherent tension exists between managing fatigue and ensuring that clinicians receive adequate training. But learning is an active cognitive process that also deteriorates with fatigue. Studies have considered the effects of reducing working hours on workplace educational opportunities and found either no difference or an improvement in procedural skills exposure.5

Laws and culture

Tight regulation is applied to workers in other industries where safety is critical: airline staff, maritime crew, anyone in road transport, and operators of public service vehicles cannot opt out of these rules. In parts of the transport industry, drivers are not legally permitted to be behind the wheel for more than four and a half hours without a 45 minute uninterrupted break. Devices such as tachographs can be used to monitor compliance, and non-compliant drivers face losing employment.24 In most countries pilots have restricted duty hours, with as few as seven hours at a time spent flying the aircraft; breaks of 30 minutes may be enforced at least every six hours.25 Commuting time to a job where safety is critical is often included in calculations of shift length if workers are driving themselves; being on-call in the workplace and training days may also be considered part of working time.47

By contrast, doctors’ work time is relatively unregulated worldwide. Europe imposes some of the tightest regulations on doctors’ hours in the world, restricting staff to an average working week of no longer than 48 hours and a maximum shift duration of 13 hours, with uninterrupted breaks of 20 minutes every six hours. Less clear is how strictly these rules are enforced in healthcare; staff can opt out of elements of the regulations, and in some cases doctors tacitly accept, or are actively encouraged to, work during days off or after night shifts.26 Relatively few papers have looked at rest periods in work, but anecdotal evidence indicates that many doctors either do not take breaks at all or use them to catch up on administration tasks. This is very different to other industries, where adequate break periods are recognised to improve performance and safety; in many industries break periods are not left to staff discretion but are strictly enforced.24 In addition, the regulation that does exist in the clinical environment tends to apply only to junior staff, whereas the working hours of senior clinicians are much less regulated or unregulated.2728

Acting on the evidence

An assumption seems to exist, not based on evidence, that medical training qualifies a person to better overcome fatigue or make their own judgments about risk. As Michael Farquhar said, commenting on the survey of anaesthetic trainees cited above, it is “terrifying . . . how little organisational awareness of, and response to, the magnitude of the problem there is.”29

Cultural norms clearly have a strong role in the assumptions around fatigue in medicine, which affect not only individual behaviour and national or organisational policy but also the kind of research that is undertaken in the field—and therefore what evidence is available. Clinicians who trained under traditional systems where 24 hour shifts and more than 80 hours a week were (or are) the norm may find it difficult to accept that even 13 hour shifts would be considered unprofessional or even illegal in other workplaces.

René Amalberti and colleagues said that “becoming ultrasafe may require healthcare to abandon traditions and autonomy that some professionals erroneously believe are necessary to make their work effective, profitable, and pleasant.”30 Another important obstacle to changing this cultural landscape is doctors’ internalised assumption that they should be able to overcome fatigue in a way that other human beings are not expected to.

One argument against further restricting hours is that care becomes more fragmented. Shorter shifts might increase handover errors in the short term, but the logical response is to improve the quality of handover—safer handover can be specifically taught, but fatigue is not amenable to training interventions.

Those managing staff cover, budgets, training, and rosters might also have fears about planning for a service that no longer tolerates high levels of fatigue in its staff. But hospitals must already plan for a certain amount of allowable time off and illness and must accept that a certain amount of sleep is imperative for even the most specialist staff. The principles are no different, merely the level of fatigue that is accepted. This cultural change has already taken place in aviation, another area where training takes many years. Airlines now routinely schedule aircrew on standby, ready to fly if a crew member is unable to undertake their duties. In industries where rules for breaks and shift limits are enforced, those in charge of staffing a service routinely include these factors in workforce planning. Eventually this becomes normalised, and the system adapts.

Shorter shifts may require more staff to cover them, with associated cost implications, but simply stating that safe shifts are impossible to achieve and that doctors must try harder is unreasonable. Rather, the costs should be included in budget and funding discussions, as they are in other industries.

Individual doctors working within externally imposed shift patterns can still take steps to tackle unsafe cultural norms. Box 1 shows the warning signs that a working pattern needs review and risk management in the rail industry.4 Most of these apply in most hospitals, most of the time. Using this list, doctors can identify their most vulnerable times when breaks are most needed; they can begin to support each other to take breaks. If you would not let a colleague work under the influence of alcohol, the same professionalism should apply to fatigue. If it’s illegal to drive a bus all day without a break, why should it be acceptable to treat patients under similar conditions?

Box 1: Working patterns considered risky in the rail industry

  • Planned day shifts of longer than 12 hours or night shifts longer than 10 hours

  • Planned early shifts of greater than eight hours (starting between 2000 and 0500)

  • More than four consecutive night shifts of any duration

  • More than three consecutive nights of more than eight hours

  • More than 12 consecutive day shifts

  • More than seven consecutive shifts of more than eight hours

  • More than four consecutive shifts of more than 12 hours

  • Working more than 55 hours in any single week

  • Any rest period of less than eight consecutive hours (unbroken by on-call or emergency working)

Healthcare regulation has historically focused on the total number of hours worked by juniors. We think it should include all levels of seniority, how working hours are clustered, whether breaks are taken, the quality of breaks, and how much fatigue staff might be accumulating. Doctors of all ages who ignore these risks are not showing professionalism but a lack of awareness of their own fallibility.

Useful opportunities for research exist, such as determining what training (if any) rota writers should have in fatigue management. Simple steps can probably be taken with existing staffing levels to improve fatigue management. Such steps might include giving rota writers training in industrial practices in shift patterning. Lengthening one’s day is easier than shortening it, so the Health and Safety Executive for Great Britain consider patterns that move “clockwise”—from day shifts, through late shifts, to night shifts, followed by rest days —to be best practice. Research on the risks associated with fatigue should not only collect data on outcomes such as mortality, but also near misses. If a system is functioning, most errors should be “trapped” before reaching the patient; so, finding no improvement in survival does not necessarily mean that risks are being managed to a high standard.

Improvements could be made now, if rota writers received training on best practices in other industries. This would be an interim solution and alone would not be sufficient to tackle fatigue across the whole healthcare system. Professional bodies such as the royal colleges will need to provide input. The Association of Anaesthetists of Great Britain and Ireland has recently established a task group to examine some of the problems associated with fatigue. We call on other professional organisations in medicine to follow and to consider the effects on their own workforce. Allied professionals will also need to consider these problems and how they affect their workforce.

The uncertain political future faced by the UK over the next few years may be relevant to this debate. Working time regulations are currently generated by the European Union. The government’s plan to transfer European legislation into UK law is important. We strongly recommend against loosening these directives in the long term.


  • We thank our advisers from the rail, bus, and aviation industries: Paul Carpenter, Darragh Kelly, and Peter Stevenson; and Helen Higham and Rosie Warren, senior clinicians with many years’ experience in medicine and nursing.

  • Contributors and sources: PG is an honorary senior clinical research fellow in anaesthesia with a special interest in human factors and safe decision making. He is an active member of faculty in the OxSTaR simulation centre in the University of Oxford and has undertaken several research projects in the Nuffield Department of Clinical Neurosciences, focusing on situational awareness and attentional focus under pressure. He is guarantor. RS had experience in the transport industry, particularly training and communications for drivers of public service vehicles. As well as working for the University of Oxford medical school and the Nuffield Department of Clinical Neurosciences and Primary Health Care Health Sciences, she was also the patient editor for The BMJ.

  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following interests: none.

  • Patient involvement: One of the authors has a long term condition and has had multiple experiences as a hospital inpatient across night and day shift handovers. The scope of the article was discussed at planning stage with non-medical advisers from the transport industry, who have frontline experience operating public service vehicles and managing the allocation of safety critical shifts. They also advised on the final version, offering valuable insights into the similarities between types of work done in healthcare and transport.

  • Provenance and peer review: Not commissioned; externally peer reviewed

  • This paper is a tribute to Rosamund Snow, who was sadly unable to see the work completed before her untimely death


View Abstract