Predicting poor functional outcome in community-dwelling older adults with knee pain: prognostic value of generic...

In longitudinal studies across a range of regional musculoskeletal pain syndromes, certain prognostic factors consistently emerge. They are "generic" in the sense that they appear to apply regardless of the particular anatomical site or underlying cause of the pain.

To investigate the value of generic indicators of poor functional outcome for knee pain and osteoarthritis in the community.

We conducted a population-based cohort study of adults aged >or=50 years with knee pain as part of the Clinical Assessment Study (Knee) (CAS(K)). At baseline, participants completed a postal questionnaire and attended a research clinic where they completed a further questionnaire and underwent structured physical examination and x rays. The 18-month follow-up was via a self-completed questionnaire. Risk ratios were calculated using Cox regression with a fixed time period assigned to each participant.

In total, 60% of participants experienced a poor outcome at 18 months. Twelve univariate associations were associated with poor outcome, with four variables remaining in the multivariate model (older age, being overweight or obese, having possible or probable anxiety, and more severe pain). Using a simple unweighted additive risk score (1 point each for age >or=60 years, body mass index >or=25 kg/m(2), possible or probable anxiety, Chronic Pain Grade II-IV), 90% of participants with all four generic indicators were correctly classified.

This study has demonstrated that generic prognostic indicators can be used to determine the prognosis of older people in the community with knee pain.