Pain in the back and lower limbs due to lumbar disorders has a multifactorial etiology, and there is no clear relationship between the morbid pathological change observed and the symptoms experienced. Of the precipitating factors, unaccustomed heavy work, postural fatigue, and injury are common. "Back injury," however, embodies a variety of phenomena, few of which can readily be distinguished either in national statistical data or in previous epidemiologic studies. Thus the causal relationship between back pain and work is far from clear, a difficulty compounded by the prevalence of back symptoms in all groups of the population. There are few epidemiologically established methods for identifying people who are susceptible to a first attack of back pain, though, once back pain has been reported, recurrence may be predicted. Selection of the preventive approach depends on reliable information about the prevalence of back pain/discomfort, the back injury rate, sickness absence, etc; about accidents, work stoppages, job turnover, or any other pointers to the design of the workplace or work practices; and about the resulting losses of productivity and the costs of the remedy. Though the decisions are up to management, the quality of the information required is mainly an occupational health responsibility.