BACKGROUND: The aim of this study was to evaluate reliability, agreement, and validity of the 1988 version of the Rowe score in patients with superior glenoid labrum lesions (SLAP) or recurrent anterior dislocations. METHODS: Thirty-eight patients with SLAP lesions and 33 with anterior recurrent dislocations were independently tested twice by 2 experienced clinicians with the 1988 version of the Rowe score. In addition, the patients completed disease specific and generic questionnaires--the Western Ontario Shoulders Instability Index (WOSI), Oxford Instability Shoulder Score (OISS), and EuroQol (EQ-5D and EQ-VAS)--at both test days. RESULTS: No significant differences between clinicians and test days were detected for the Rowe score. Limits of agreement of the total score varied from -17.9 to 20.5. For the categorical score (total units), there were significant differences between the 2 examiners (P < .001). The ICC 2.1 was acceptable (>.70) for the total score. No floor or ceiling effects were observed for the total score, but considerable floor and/or ceiling effects were detected for some of the domains. Discriminant validity was acceptable, but content, construct, or convergent validity was not acceptable. CONCLUSION: Results using the 1988 version of the Rowe score should be critically interpreted.
BACKGROUND: The aim of this study was to evaluate reliability, agreement, and validity of the 1988 version of the Rowe score in patients with superior glenoid labrum lesions (SLAP) or recurrent anterior dislocations. METHODS: Thirty-eight patients with SLAP lesions and 33 with anterior recurrent dislocations were independently tested twice by 2 experienced clinicians with the 1988 version of the Rowe score. In addition, the patients completed disease specific and generic questionnaires--the Western Ontario Shoulders Instability Index (WOSI), Oxford Instability Shoulder Score (OISS), and EuroQol (EQ-5D and EQ-VAS)--at both test days. RESULTS: No significant differences between clinicians and test days were detected for the Rowe score. Limits of agreement of the total score varied from -17.9 to 20.5. For the categorical score (total units), there were significant differences between the 2 examiners (P < .001). The ICC 2.1 was acceptable (>.70) for the total score. No floor or ceiling effects were observed for the total score, but considerable floor and/or ceiling effects were detected for some of the domains. Discriminant validity was acceptable, but content, construct, or convergent validity was not acceptable. CONCLUSION: Results using the 1988 version of the Rowe score should be critically interpreted.