• patient safety
• doctors’ hours of work
• hours of sleep
• occupational stress

The long hours that doctors work, and the length and quality of their sleep, have long been viewed as influencing their health and the safety of care they give.¹ In 2000 it was agreed by the European Parliament that the Work-time Directive, which had limited working hours in general, should also apply to doctors in training. This will mean no more than 58 hours per week by August 2004 and eventually a maximum of 48. In Europe this has more to do with the risks towards the doctor while in the US efforts to reduce long hours have come more from their threat to patient safety. But is this focus on hours the right one?

There is certainly some evidence that a long working week will affect your health: a meta-analytical review of workers in general found small but significant positive correlations between overall health symptoms, physiological and psychological health symptoms, and hours of work.² In Japan, where working hours are particularly long, this has been implicated in cardiovascular disorders and diabetes mellitus.³ Nevertheless, most studies show surprisingly little evidence for the relationship between hours of work and psychological well being: even when hours are long, they have little or no correlation with levels of stress or depression.^4,5 On the other hand, a combination of the intensity of demands (rather than long hours per se) combined with low discretion over how one’s work is done continues to be an influential model of the causes of job related stress.⁶ Doctors have consistently shown unusually high levels of stress symptoms and consider overwork to be involved in this.¹ What is likely, however, is that the tiredness caused by long hours makes any potential job related stressors— such as lack of support, dealing with patient deaths, or the high cost of error in medicine—much more difficult to cope with and, in this way, stress symptoms may be a mediator between hours worked and psychological disorders such as depression or anxiety.⁷ Focusing only on reducing hours of work may not have the desired effect of reducing these symptom levels.

“... it may be better to aim for ensuring good support, leadership, and teamwork...”

Nevertheless, doctors certainly see overwork—both long hours and a lack of support—as contributing to incidents of poor care,⁸ and evidence from a military study suggests that individual errors increase with fatigue.⁹ However, the same study found that, over time, the team can compensate for these errors, and so those who worked together over a longer period had higher individual error scores but lower team error scores than those working together only briefly. It seems that the relationship of hours to error is complex,⁷ mediated by such factors as the support of colleagues, including leaders, as well as the levels of control one has over one’s job.⁶ Certainly for younger doctors, long hours of work, well managed,^10,11 may be a relatively insignificant factor compared with the positive effects of being part of a team and the enjoyment of practising medicine. Rather than continuing the pursuit of a shorter working week, it may be better to aim for ensuring good support, leadership, and teamwork since good teams have healthier staff and better outcomes.^12,13

Although there is inevitably a relationship between hours of work and hours of sleep, the negative effects of sleep loss on both mood and performance have been found much more consistently than those of long hours. Sleep deprivation necessarily varies in real life studies, but most definitions involve fewer than 4 hours sleep in 24, or frequently interrupted sleep. Most studies show that mood is lowered after a long on-call shift and young doctors report feeling more confused and less confident.¹⁴

These psychological effects are mirrored where the quality of care is considered. Errors increase with sleep loss, and data on shift work show a rise in industrial injuries on the night shift.¹⁵ In medicine, such data are rarely collected and instead a variety of cognitive tests are used as proxy measures of performance—assessing memory, concentration, alertness, and attention to detail after nights on call. Review studies¹⁶ and meta-analyses¹⁷ agree that sleep deprived people perform significantly less well than controls, particularly in terms of mood and cognitive tasks, whether proxy assessments or simulations are used. For example, surgical house officers following a weekend on call showed significant impairments in concentration and speed.¹⁴ Performance in general was highly related (0.5–0.6) to the number of hours of sleep: an impairment to vigilance estimated to be close to that caused by 0.7 g/kg alcohol which, as they point out, is near the legal UK limit. Simulations of laparoscopic surgery show that more errors occur with increasing sleep loss¹⁸ and after a night on call.¹⁹ Moreover, there is some evidence to suggest that there are deleterious effects even 2 days later.²⁰

In the short term, an educational approach to the dangers of sleep loss and ways to improve sleep quality may be useful, alongside encouraging leaders to ensure that those deprived of sleep are not put in potentially dangerous situations.¹⁰ The benefits of napping²¹ or rest periods²² in terms of subsequent performance have been reported and could be built into the working day officially rather than surreptitiously by individuals. However, it seems important that we integrate our approaches to improving working lives and patient care,¹ and resist focusing only on hours or any single factor. What is needed is a systems approach towards a cultural change that uses an evidence base to address the complex factors that contribute to staff working below par, and that treats healthy alert staff as a key element of patient safety.