Community management of knee pain in older people: is knee pain the new back pain?

Over the last 10–15 yr there has been considerable academic interest in the community management of low back pain. This has been due, at least in part, to the high indirect costs, resulting largely from lost production, that have been attributed to spinal problems. Although questions still remain about how those working in primary and community care should manage back pain, the combination of new epidemiological studies 1, careful systematic reviews of existing data 2 from completed trials 3, 4 and ongoing trials 5, plus guideline development 6 and implementation 7, and new models of service delivery 8, are changing the conservative management of back pain in the community. This overall approach could be used as a model for addressing the community management of other painful musculoskeletal disorders for which this clarity of approach is not evident; for example, knee pain in older people. It is possible that the perceived importance of lost production due to spinal disorders has diverted attention away from tackling knee pain in older people, a group who are not usually economically active. This raises ethical questions about the equitable distribution of health-care resources. Age and potential contribution to the wider economy are not relevant reasons for the provision of unequal resources for research and treatment for two equally disabling conditions 9, 10. In an important study, reported in this issue of Rheumatology 11 and elsewhere 12, 13, Jinks et al. have highlighted the health impact of knee pain in older people. Because patients seek treatment for knee pain, not osteoarthritis, and there is an imprecise relationship between radiological change and symptoms, it is gratifying that they have studied the health impact of knee pain, not knee osteoarthritis. In a population survey of nearly 9000 people aged 50 yr or over, they found that one in four people had chronic knee pain (>3 months) and one in five had non-chronic knee pain (<3 months). Furthermore, half of these people had severe pain or severe difficulty with physical function. On the basis of 2001 census data 14, this means around four and a half million people aged 50 yr or over in the UK have severe problems from knee pain. This number is set to increase substantially. The population aged 50 or over is predicted to increase by a quarter, to 25 million, by the year 2020 15. Additionally, the proportion of the population who are obese is predicted to increase to 20–25% by 2010 (compared with 6% for men and 8% for women in 1980) 16. There are well-documented relationships between increasing weight and incident knee osteoarthritis 17 and between obesity and knee pain 12. Thus, over the next 10 or 20 yr we can expect a substantial increase in the number of older people with severe problems related to knee pain. The estimated annual need for knee replacements in England is around 56 000 18; 41 000 were done in 2000 19. These rates are too low to have a major impact on the overall prevalence of knee pain. This means that there is a clear need to both identify and implement the most effective conservative treatment approaches for knee pain. Recognizing the heterogeneity of duration, symptoms and consulting pattern, Jinks et al. have developed a simple typology that has the potential to allow us to target appropriate interventions at different population groups. Conceptually, this could be very important, as it will allow us to test our interventions in clinically rather than radiologically defined groups. Hopefully, the predictive value of this classification will be born out in their longitudinal study and insight into reasons for different consultation patterns obtained from their qualitative study. Searching the literature in July 1999, the authors of the EULAR (European League Against Rheumatism) recommendations for the management of knee osteoarthritis identified 592 randomized controlled trials of non-surgical interventions for knee osteoarthritis 20. Unsurprisingly, 365 (62%) of these were trials of non-steroidal anti-inflammatory drugs (NSAIDs). Other conservative treatments for which there is not a regulatory need to produce randomized controlled trial evidence for efficacy have been studied less intensively. Only for NSAIDs and patient education did they find category 1A evidence (meta-analysis of randomized controlled trials) for efficacy. Health economic data on non-drug interventions are sparse. As the authors of the EULAR recommendations point out, data derived from highly selected homogeneous groups of subjects with osteoarthritis recruited to randomized controlled trials may not be directly relevant in clinical practice 20. Clinical Evidence, reviewing the literature up to March 2002 21 on the management of osteoarthritis in any joint concluded that: (i) paracetamol, NSAIDs and topical agents all had short-term beneficial effects; (ii) exercise and physical aids were likely to be beneficial; (iii) NSAIDs were likely to be ineffective or harmful in older people; and (iv) the effectivenesses of education, glucosamine, chondroitin and glucosamine plus chondroitin were unknown. A subsequent, community-based trial of home-based exercise for knee pain has also found a significant reduction in pain 22. Apart from promising data on exercise regimens, robust randomized controlled data are not available to directly inform the long-term community management of knee pain in older people. A further problem with interpreting the existing data is that many of those with knee pain also have one or more other musculoskeletal pains 12, 23. This means that using a measure of outcome specific to knee pain may be too limited, and that a measure of overall pain and disability may also be required. Extrapolating from the available data, it is plausible that exercise and/or education regimens may be helpful and are unlikely to be harmful. However, it is not clear which approach or approaches are likely to be most effective. In 2001, 19 414 200 prescriptions for oral NSAIDs were dispensed in the community in England 24. In spite of the well-recognized side-effects of these drugs, older people with knee pain are likely to have taken a substantial proportion of these prescriptions. It is plausible that strategies to reduce NSAID use in this group will also be beneficial. Jinks et al. have identified eight groups of older people with knee pain that could benefit from such interventions, tailored to their needs. Should those with chronic non-severe knee pain who are accessing services be referred for lifestyle and exercise advice? For example, should there be programmes to identify those with chronic severe pain who are not accessing services in order to optimize their management? Sadly, the data to underpin the optimum management for people in any of Jinks' eight groups are not available. Thus, it may be premature to embark upon the development and implementation of primary care guidelines to tackle the epidemic of knee-pain-related disability in older people. Is there a way forward? First, the importance of the knee pain epidemic needs to be recognized and the same amount of attention and resources needs to be given to addressing it as that given to back pain in the 1990s 25. General practices are likely to be primarily responsible for managing knee pain in older people and for prescribing NSAIDs in the UK. It is of concern that in the new general practitioner contract proposed in 2003 there were no quality indicators for the management of musculoskeletal problems or the prescribing of NSAIDs 26. Also, the National Service Framework for Older People for the NHS, published in 2002 27, had no recommendations on the management of osteoarthritis. However, the NHS Health Technology Assessment programme has identified osteoarthritis management as a research priority 28. In 2002 it funded a study comparing oral and topical ibuprofen for knee pain and invited proposals for a long-term study of glucosamine, although sadly none of the proposals submitted were funded. Although the evidence supporting non-drug interventions is weak, the risk of harm to participants is low. If they are effective, they have the potential to reduce costs elsewhere in the NHS by reducing NSAID use. Primary care organizations may want to consider the development or funding of services targeted at one or more subgroups of older people with knee pain. However, if any such services are introduced it is important that they are well evaluated to inform future service development. Accepting that services may develop in the absence of robust data to underpin their approach does not obviate the need for new research to determine the best approach to older people with knee pain in each in each of Jinks' categories. It is difficult to decide priorities amongst the many candidate interventions for knee pain. However, promising approaches with comparatively few adverse effects that need further work include (i) educational interventions, (ii) exercise regimens, and (iii) glucosamine/chondroitin. Additionally, work is needed to find out whether any changes to models of service delivery for older people with knee pain could improve access to appropriate care and, consequentially, outcomes for older people with knee pain. For all of these interventions, long-term follow-up studies are needed with robust health economic analyses, which will inform appropriate management at different levels of disability. We will then be able produce appropriate guidelines and services to help those working in primary care to manage knee pain in older people. I am supported by an NHS R&D Primary Care Career Scientist Award. Conflict of interest The author has declared no conflicts of interest.