Clinical course and prognostic factors across different musculoskeletal pain sites: A secondary analysis of individual patient data from randomised clinical trials.

ABSTRACT

BACKGROUND: Previous research has identified similar prognostic factors in patients with musculoskeletal (MSK) conditions regardless of pain presentation, generating opportunities for management based on prognosis rather than specific pain presentation. METHODS: Data from seven RCTs (2483 participants) evaluating a range of primary care interventions for different MSK pain conditions were used to investigate the course of symptoms and explore similarities and differences in predictors of outcome. The value of pain site for predicting changes in pain and function was investigated and compared with that of age, gender, social class, pain duration, widespread pain and level of anxiety/depression. RESULTS: Over the initial three months of follow-up, changes in mean pain intensity reflected an improvement, with little change occurring after this period. Participants with knee pain due to osteoarthritis (OA) showed poorer long-term outcome (mean difference in pain reduction at 12 months -1.85, 95% CI -2.12 to -1.57, compared to low back pain). Increasing age, manual work, longer pain duration, widespread pain and increasing anxiety/depression scores were significantly associated with poorer outcome regardless of pain site. Testing of interactions showed some variation between pain sites, particularly for knee OA, where age, manual work and pain duration were most strongly associated with outcome. CONCLUSIONS: Despite some differences in prognostic factors for trial participants with knee OA who were older and had more chronic conditions, similarity of outcome predictors across regional MSK pain sites provides evidence to support targeting of treatment based on prognostic factors rather than site of pain. SIGNIFICANCE: Individual patient data analysis of trials across different regional musculoskeletal pain sites was used to evaluate course and prognostic factors associated with pain and disability. Overall, similarity of outcome predictors across these different pain sites supports targeting of treatment based on prognostic factors rather than pain site alone.