Intended for healthcare professionals


Working long hours and health

BMJ 1994; 308 doi: (Published 18 June 1994) Cite this as: BMJ 1994;308:1581
  1. J M Harrington

    Whether working long hours adversely affects health has been debated for many years. A recent European Council directive on working time (93/104/EC) has heightened the controversy. Its proposals include a minimum daily rest period of 11 consecutive hours in each 24 hour period, at least one rest day a week, four weeks' annual leave, and a restriction on night work to a maximum of eight hours on average. With some exceptions (including doctors in training), employees would, under the directive, have the legal right to refuse to work more than 48 hours a week. The directive would be implemented under article 118A of the Treaty of Rome, which requires the directive to be based on health and safety considerations and not general employment conditions. The British government disputes that this is a health and safety measure and is challenging its legal basis before the European Court of Justice.

    The topic has attracted further attention recently with the death of a junior hospital doctor after an extended period on duty, and again in the aftermath of the untimely death of the leader of the Labour party. So what is the evidence linking working hours and health? The explanatory memorandum that accompanies the directive should help, and there is certainly no shortage of scientific papers.

    Unfortunately, neither of these sources provides an unequivocal answer. The explanatory memorandum is particularly unhelpful. Most of the references are German: this is not a reflection of the world literature; many are obscure, and some are untraceable. It is not possible to examine, let alone support, assertions on health and working hours from this standpoint. A general literature search uncovers hundreds of papers since my review of the subject in 1978.1 Three recent reviews are particularly good.*RF 2-4* In addition, over 20 reviews have concentrated on specific concerns such as sleep, fatigue, performance, and the scheduling of hours of work.

    At least half of the individual papers are concerned with attitudinal surveys, devoted to employees' (occasionally employers’) opinions. Most of the rest are still rooted in studies of shift work, with few attempting to investigate the growing fashion for 10 hour or 12 hour working days. Virtually no work has been done on the influence of rest periods on health, and the early pioneering research on fatigue in munitions workers during the first world war remains unparalleled.5 Indeed, no substantive research has looked at holiday periods or work on Sundays.

    Working hours outside “normal” office hours, either because of extended work schedules or because of shifts entailing night work, disrupts circadian rhythms. This may cause fatigue and can certainly disrupt social life. Few employees like such work, but the need of industry and commerce carries great economic weight. Little work has been undertaken to study the influence of gender on work schedules.

    There is general agreement that working abnormal hours leads to loss of quantity and quality of sleep.6 Sleep disturbance is greatest after a night shift and may lead to napping at work. Not surprisingly, fatigue is a frequent - if immeasurable - complaint of shift workers. Stress is equally difficult to quantify, but some good studies have provided limited evidence for such work schedules being linked to anxiety or depression.7 But shift workers are a highly selected group, and distinguishing cause from effect is difficult.

    Although the case for linking gastrointestinal disease - particularly peptic ulceration - with shift work is reasonably good, no recent studies have tested this longstanding association. By contrast, the case for linking cardiovascular mortality and morbidity and unsocial hours has strengthened considerably recently, although most of the studies are Scandinavian.3 The phenomenon of sudden death linked to overwork (“karoshi”) is mainly derived from Japan8 and mainly published as case reports. Sound epidemiological studies are lacking to support or refute this assertion.

    New working patterns

    Many good quality studies now exist to support the notion that work performance and output are poorer at night.9 Some evidence exists to suggest that safety records are poorer at night as well. Overall, it seems that, although working abnormal hours is difficult to avoid in some industries, work scheduling, such as rapid, forward rotating shifts (spells of three or four days on each shift with a morning/afternoon/evening sequence), is less disruptive than most other options.4 The compressed working week of three or four 10 to 12 hour shifts is gaining vogue. The trade off of a longer block of rest days seems to offset the increased fatigue of the longer shifts, but few data exist on long term risks to health and safety.

    With regard to the directive, there is no unequivocal scientific evidence to support a 48 hour week; nor is there any overwhelming evidence against it. One study suggests that working more than 56 hours a week carries serious health and safety risks.10 After a 12 hour break seems reasonable.

    Yet the directive rolls on and carries in its wake the requirement for “health checks” on night workers. The widespread introduction of unfocused health surveillance measures is unjustifiable on grounds of preventive health. Astonishingly, 100 years after the first experiments on shorter working hours and performance11 we are still unable to decide whether working long hours is bad for health. Seeing how the European Court of Justice handles it will be interesting.