Margit Kooijman

Therapist effects | 107 Table 2 shows that patients with complete data were different with respect to age, whether it concerned a recurrent complaint and origination of the complaint compared to patients with missing severity of shoulder complaints. TABLE 2. Characteristics of patients with complete and missing data on severity of their shoulder omplaints Factors associated with change in severity of shoulder complaints Mean severity of shoulder complaints at the start of treatment was 6.5 (sd 1.7, 95% CI 6.4;6.6), mean severity at the end of treatment was 1.8 (sd 2.1, 95% CI 1.7;2.0). During treatment, the severity of shoulder complaints significantly decreased (p < 0.001) with 5.0 points (95% CI -5.4; -4.6), adjusted for the patient and therapist variables of model 1. The proportion of total variance explained (intraclass correlation coefficient) in change in severity was 0.88 at patient level and 0.12 at the physiotherapist level meaning that 12% of variance in change of severity of the shoulder complaint was explained by (characteristics of) physiotherapists. Table 3 presents the results of the univariate analysis of possible predictors of change in severity of shoulder complaints. Complete data (%) (n = 1013) Missing severity (%) (n = 1103) P value Age, mean (SD) 53 (16) 5 (17) 0.01 Gender 0.35 Male 44 42 Female 56 58 Referral 0.19 GP 53 56 Medical specialist 6 6 Direct access 41 38 Duration of complaint 0.41 < 1 month 35 33 1-3 months 31 30 > 3 months 34 37 Recurrent complaints <0.001 Yes 23 33 No 77 67 Origination 0.01 Acute 29 24 Gradual 71 76