Managing painful musculoskeletal disorders is a major part of general practice. Estimates for the proportion of the population consulting annually for musculoskeletal disorders, derived from general practice consultation databases, range from 6.6 to 20.7%.1 Fortunately, many patients improve independent of any treatments we may advise. However, a minority develop chronic pain and disability which has a substantial health and social impact. Predicting which patients are more likely to have a poor outcome from their musculoskeletal pain may help us to make better use of resources. The pain that presents most commonly for treatment — and which is perceived to have the highest economic cost — is low back pain. Consequently, it has attracted considerable research and policy interest in recent years. There is increasing recognition that low back pain commonly coexists with other musculoskeletal pain; and that the features of different non-specific pain syndromes may also have elements in common and frequently occur together.2–4
In this issue of the Journal, Mallen et al report a systematic review of prospective cohort studies for a range of musculoskeletal pain.5 They raised concerns about the quality of some studies and noted that few studies were of non-spinal pain; however, a range of factors predictive of poor outcome was identified. The most common were baseline pain severity, duration, and pain at multiple sites. Similar factors were identified as predictors of outcome for pain in different body regions. This supports the notion that there are more similarities than differences between low back pain and other non-specific pain syndromes and suggests that approaches used for low back pain may be applied more generically.
Questions remain unanswered. Is pain in certain areas more disabling than in others? Is low back pain the most disabling of these pain? Should we, as Mallen et al suggest, carry out more research in the area of peripheral joint pain rather than axial pain? Interestingly, very few studies they identified focused on prognostic indicators for recovery from, or adaptation to, musculoskeletal pain.
Much existing research is body site-specific. However, as others have observed, site-specific chronic pain is less common than multi-site or widespread chronic pain.2,6–8 It is undoubtedly easier to compartmentalise pain for research, economic evaluation, treatment, and management; but by doing so we may be providing a distorted image of the distribution and nature of pain. Consequently, we may be mismanaging patients, allocating funding and other resources inappropriately, and disregarding some of the more important factors that predict a poor overall patient outcome.
Considering musculoskeletal pain in isolation may still be too narrow a focus. Patients with chronic widespread pain may be more likely to consult their GP about other non-specific disorders.9 A generic whole-patient approach used by the Expert Patient Programme to address such disorders is an attractive option. Unfortunately, although this approach seems to have a modest effect on self-efficacy, it does not appear to have a significant impact on clinically relevant outcomes when evaluated in randomised trials in the UK.10
Some factors, such as duration of pain and possibly number of sites affected, are not alterable. Conversely, factors such as somatic perceptions, coping, distress, anxiety, and depression are potentially modifiable. Knowledge of these predictors will improve GPs' ability to identify patients with poor prognoses; their clinical judgement is already nearly as good as the more complex prediction rules with which they have been compared.11
Even for low back pain, there is a lack of evidence as to whether targeting more intensive or specific treatments to particular patients will affect clinically relevant outcomes.12 New work is needed to find out whether identifying such factors will enable us to identify which patients are likely to gain the greatest benefit from particular treatments.
Future investigators should consider that Malen et al were unable to pool data from different studies of musculoskeletal pain because of marked heterogeneity in the populations studied, data collected, and analytical approaches used. The MMICS collaboration has started to address this issue, proposing a core set of factors to be collected in prospective cohort studies of acute back pain.13 This approach could also inform the design of future cohort studies in the wider spectrum of musculoskeletal pain.
Observational studies may also need to consider the impact of health and social care systems on outcome in addition to individual patient factors. The cost of supporting a large population of patients with either local or widespread chronic pain includes healthcare costs, incapacity benefit, and the loss of productive working time. Addressing indicators that predict poor prognosis and exploring those that predict recovery or improve outcomes and response to treatment, should help to inform the delivery of effective and appropriate care, thereby lessening the overall burden for individuals and society.
- © British Journal of General Practice, 2007.