Results

The baseline survey identified 2715 adults free of low back pain, of whom 1412 were employed full or part-time (47% male, mean (SD) age 38 (13) years), and 1303 non-employed (42% male, mean (SD) age 53 (19) years). Follow up data were available for 1668 subjects: 847 in employment (45% male, mean (SD) age 39 (13) years) and 802 non-employed (43% male, mean age 55 (18) years).

The cumulative incidence of new low back pain episodes recalled by the employed and non-employed groups were: consulting 4.4% and 3.6% respectively; non-consulting 32% in both groups.

For both perceived adequacy of income and degree of satisfaction with employment status, and with respect to both consulting and non-consulting new episodes, the non-employed showed similar risks of low back pain to the employed (table 1). Regardless of employment status, perception of income as inadequate is associated with a threefold risk of consulting because of back pain during the follow up year while dissatisfaction with their situation doubled the risk of study subjects reporting a new episode of low back pain.

A previous history of back pain and psychological distress have been shown to be risks for developing a new back pain episode,9 ,10 and lower social class is associated with consulting for back pain.11 Neither stratifying by a past history of back pain, nor adjusting for psychological distress or social class, made a significant difference to these associations. Furthermore, although the employed group were younger, restricting the non-employed to those under retirement age did not change the reported results.

Discussion

Most previous investigations of psychosocial risks associated with low back pain in workers have been workplace-based and, with few exceptions,2 ,3 cross sectional. In this study potential psychosocial risks for low back pain were identified in a back pain free population cohort, clarifying the temporal relation between cause and effect both in workers and non-workers. Previous studies have used similar questions to ascertain information from employees on their degree of satisfaction with work, and have reported similar associations with low back pain. However, few have elicited comparable information by referring specifically to those not in paid employment, such as the retired, unemployed, etc, as does the instrument used here.7 The question concerning the perceived adequacy of income was identical for both groups.

In common with other population studies, non-responders to the initial survey were younger and more likely to be male. Their overall consultation rate during the following year was marginally lower (6%) than that of responders (7%). The study design enabled data from the initial survey to be compared for responders and non-responders to the 12 month follow up survey. Non-responders to this were younger, and significantly more likely to smoke and have higher GHQ scores than responders. However, the validity of the results would only be effected if any systematic bias, should it exist, effected the association between the risk factors and low back pain in employed persons differently from that in the non-employed.

The difference between the one year cumulative episode incidence of low back pain reported in the follow up study (32%), and the cumulative consultation rate (approximately 4%) clearly illustrates that most people with a new back pain episode do not consult their general practitioner because of this—only about one in seven choose to do so. This confirms the results of previous studies12 and suggests that the association between low back pain and primary care consultation is not direct, but mediated by other factors.13

It could be argued that workers with lower incomes are more likely to consult their general practitioner to obtain the sickness certification necessary for state benefits. However, similar risks in the non-employed make this explanation unlikely. Perception of income as inadequate might not reflect actual income, but be a more subtle indicator of a propensity to consult, possibly mediated by somatising behaviour.14 Dissatisfaction with your situation might be more directly related to back symptoms, regardless of employment status. This is in line with Waddell’s comment that low back pain in workers is confounded by the high prevalence of this symptom in the general population,4 and suggests that the dominance of psychosocial risks of low back pain,2 ,3 ,5 compared with risks relating to physical activities at work,6 might be a function of subjects’ dissatisfaction with life in general.

Most reports of back pain in workers refer to the economic impact of the exponential rise in disability payments and absenteeism caused by low back pain.4 Our results suggest that, while psychosocial intervention in the workplace might reduce the economic costs to society of back pain, these effects would be limited if no account is taken of the influence of broader non-work related aspects. Moreover, for the individual, such interventions are unlikely to impact on the overall experience of back pain and consultation behaviour.