Objectives: To review the efficacy of common interventions for shoulder pain.
Search strategy: We searched the Cochrane Musculoskeletal Group trials register, Cochrane Controlled Trials Register, Medline, Embase, Cinahl, and Science Citation Index) up to May 1998, and hand searched major textbooks, bibliographies of relevant literature, the fugitive literature, and the subject indices of relevant journals including: American College of Rheumatology;British College of Rheumatologists; the Biennial Conference of the Manipulative Physiotherapy Association of Australia;International Federation of Manual Therapists conference proceedings; British Orthopaedic Association;and American Orthopaedic Association.
Selection criteria: Each identified study was assessed for possible inclusion by two independent reviewers based on the blinded methods sections. The determinants for inclusion were that the trial include an intervention of interest (non-steroidal anti-inflammatory drugs, intra-articular or subacromial glucocorticosteroid injection, oral glucocorticosteroid treatment, physiotherapy, manipulation under anaesthesia, hydrodilatation, or surgery); that treatment allocation was randomized; and that the outcome assessment was blinded.
Data collection and analysis: Methodological quality was assessed by two independent, blinded reviewers. Data relating to selection criteria, outcome measurement and treatment effect was extracted from the blinded trials. Range of motion scores were entered as degrees of restriction to movement, and all pain and overall effect scores were transformed to 100 point scales. For continuous outcome measures, where standard deviation was not reported it was either calculated from the raw data or converted from standard error of the mean. If neither of these were reported, authors were contacted in an effort to obtain the missing values. Effect sizes were calculated and combined in a pooled analysis if study population, endpoint and intervention were comparable.
Main results: Thirty one trials met inclusion criteria. Mean methodological quality score was 16.8 (9.5 - 22) out of possible score of 40. Selection criteria varied widely even for the same diagnostic label. There was no uniformity in outcome measures used and their measurement properties were rarely reported. Effect sizes for individual trials were small (-1.4 to 3. 0). The results of only three studies investigating rotator cuff tendonitis could be pooled. Benefit of subacromial steroid injection over placebo for improving range of abduction (weighted difference between means (WMD) 35 degrees, 95% CI 14 to 55) was the only positive finding.
Reviewer's conclusions: There is little evidence to support or refute the efficacy of common interventions for shoulder pain. As well as, the need for further well designed clinical trials, more research is needed to establish a uniform method of defining shoulder disorders and developing outcome measures which are valid, reliable and responsive in these study populations.