Identifying Generic Predictors of Outcome in Patients Presenting to Primary Care With Nonspinal Musculoskeletal Pain

Objective To identify which generic prognostic factors, such as pain intensity, levels of disability, and psychological factors, are most strongly associated with outcome from musculoskeletal pain, regardless of the location of pain. We tested the hypothesis that pain location does not add predictive value to these generic prognostic models, and that such prognostic factors are equally important across different pain locations. Methods Data from a prospective observational cohort of primary care patients with acute (n = 413) and chronic (n = 414) nonspinal musculoskeletal pain were used to develop predictive models. The analysis was carried out in 3 steps: derivation of predictive models including generic factors only, investigation of the added predictive value of pain location, and investigation of effect modification by pain location. Results Generic factors predicted outcome over different time periods (3 months and 12 months) and for both acute and chronic musculoskeletal pain (area under the receiver operating characteristic curve 0.73-0.75). The most consistent predictors of poor outcome were having had the same complaint in the previous year (odds ratio range 2.03-3.46), a lower level of education, lower scores on the Short Form 36 vitality subscale, using pain medication at baseline, and being bothered by the complaint more often in the past 3 months. Pain location variables only slightly improved the predictive ability of the models over generic factors and were inconsistent across the models. Conclusion Generic factors appear to play an important role in the prognosis of acute and chronic nonspinal musculoskeletal pain, regardless of the location of pain. Copyright © 2012 by the American College of Rheumatology