Intervention trials on upper body pain among computer operators

Short abstract

Commentary on the paper by Rempel et al (see page 300)

Computer work is now one of the most widespread work tasks in the world. In some western countries more than half of the workforce use a computer station for more than half of their working day. Adverse effects of the computer could thus have a large impact on public health, even if effect sizes are small. Most concern has been expressed about pain and disorders of the musculoskeletal system in those using the keyboard for data entry and other keying tasks, and in the use of the computer mouse and other input devices.¹

Concern about non‐specific neck and arm pain is not new but has existed for centuries, various terms being used at different times according to the suspected causal exposures and affected populations (for example, writers' cramp, telegraphists' cramp).² At the present time, however, no well established and clinically accepted diseases of the musculoskeletal system have been ascribed with certainty to computer use. On the contrary, several large epidemiological studies encompassing clinical assessments have not found the prevalence or incidence of specific musculoskeletal disorders to be higher than in the general population.³,⁴,⁵,⁶,⁷,⁸,⁹ Nevertheless, many workers and researchers regard upper body pain as a work related problem of computer users, and interest in the media has been huge. Belief in a pain syndrome ascribed to computer use is widely shared.

The modern solution is ergonomics—the classical approach of adjusting the workstation to the worker. Nowadays this is a large and expensive industry, but without much scientific evidence to define the “correct” or optimal workstation. Intervention studies with several different focuses have been performed over the past 10–15 years to develop an evidence base. In this issue of the journal, Rempel et al¹⁰ report on an intervention study among 182 customer service operators in a one year randomised controlled study with four experimental arms. The intervention consisted of training and the introduction of trackballs and forearm support. The outcomes were weekly pain scores, and diagnoses of incident musculoskeletal disorders in the neck and upper extremities. The main finding regarding pain severity was that the armboard intervention was associated with a significant mean reduction in pain of 0.48 points on a 0–10 point scale.

Could such a reduction be described in a qualitative way, such as a reduction from “some pain” to “minor pain”, or from “severe pain” to “some pain”? I think not: it is difficult to give some meaningful expression to this small change, from an overall mean pain level of around 2–3 to 2 or a little less. Despite this reservation, a significant proportion of participants also rated their experience of pain as decreased. However, I would have preferred the outcome of interest to be a well defined and clinically important decrease in pain score (for example, a decrease of at least 2 points). The clinical assessment also revealed a protective effect for the armboards, with a reduction of the hazard rate of incident neck‐shoulder disorders to 0.49, which means a reduction by approximately half. Surprisingly, the authors found as many as 22 new cases of shoulder tendonitis, corresponding to an incidence of more than 10%. In two other studies, the incidence of this outcome was much lower: 1.3%⁴ and less than 0.1%⁷ respectively. In all three studies, diagnosis required the symptom of shoulder pain and the semi‐objective examination sign of pain on resisted movements, but the example of these three studies illustrates one of the major problems in epidemiological surveys of musculoskeletal pain: all too often the diagnostic criteria lack a solid foundation in terms of validity or reliability. Shoulder tendonitis is but one example. When it comes to the diagnosis of somatic pain syndrome and thoracic outlet syndrome, the confusion seems even more impressive. In general, much more work should be encouraged to improve case definitions for epidemiological purposes.

Recently, another randomised controlled trial of postural interventions for prevention of musculoskeletal symptoms among computer users was performed and the findings published in this journal.¹¹ Gerr et al found no differences in risk of musculoskeletal symptoms among 376 participants randomly assigned to two workstation and postural interventions in comparison to no workstation or postural intervention. Meanwhile, in another intervention study, Aarås et al found a reduction in shoulder pain in parallel with a reduction in trapezius load in a small group of female data dialogue workers after instituting a training programme and providing more ergonomic information.¹² However, the results are difficult to interpret due to the small sample size and lack of information on study eligibility.

More research is needed. In the meantime, what should health and safety practitioners do? Given the limitations of our current knowledge I find it difficult to make recommendations regarding postural or other specific adjustments of the workstation among computer users. Maybe the only recommendation should be that computer users should be satisfied with their workstation. Every reasonable effort should be made to give them the set‐up they want. They should have the opportunity to influence their own work and how they perform it, including the right to make their workstation more comfortable or to use an armboard if they wish.