Shoulder pain in primary care--part 2: predictors of clinical outcome to 12 months.

OBJECTIVE Identify predictor variables and models for clinical outcomes for primary care shoulder pain patients to 12 months follow-up. DESIGN A non-randomized audit with measures of pain and disability at 3 weeks, 3, 6 and 12 months. PATIENTS Of 208 patients, 161 agreed to participate with 96.9, 98.1, 87.0 and 83.9% follow-up at 3 weeks, 3, 6 and 12 months respectively. Treatment consisted of exercise and manual therapy-based physiotherapy and corticosteroid injection under specified selection criteria. METHODS Potentially useful baseline variables were evaluated in univariate logistic regressions with the dependent variables determined by SPADI Questionnaire at 3 weeks, 3, 6 and 12 months. Variables associated (p-value ≤ 0.2) were retained for potential inclusion within multiple logistic regression analyses. RESULTS Pain not improved by rest, intermittent pain, lower pain intensity with physical tests and absence of subacromial bursa pathology on ultrasound at the 3-week follow-up, constant pain and lower pain intensity with physical tests are predictors of excellent outcomes at the 3-month follow- up. Worse baseline pain and disability, no history of asthma, pain better with rest, better physical functioning, greater fear avoidance, male gender, no history of pain in the opposite shoulder, pain referred below the elbow, sleep disturbed by pain, smaller waist circumference, lower pain intensity with physical tests are factors predictive of excellent outcomes at the 12-month follow-up. Only higher pain intensity with physical tests was associated with a poor clinical outcome. CONCLUSION Predictive models for clinical outcomes in primary-care patients with shoulder pain were achieved for excellent clinical outcomes, successfully classifying 70-90% of cases.