Psychometric testing of the modified Care Dependency Scale for Rehabilitation

Objective: To determine the psychometric properties of the modified Care Dependency Scale for Rehabilitation. Design: All patients were assessed independently by a different set of two randomly selected nurses. Setting and subjects: A convenience sample consisting of 1197 patients and 42 nurses was recruited in an Austrian rehabilitation centre specialized on the treatment and care of neurological and orthopaedic patients. Measures: The Care Dependency Scale for Rehabilitation and the Barthel Index. Data analysis: To estimate aspects of reliability, Cronbach's alpha, Cohen's kappa, percentage agreement, intraclass correlation coefficient (ICC) and standard error of measurement (SEM) were calculated. To evaluate the construct validity, a factor analysis was conducted. Criterion-related validity was examined by comparing data of the modified Care Dependency Scale for Rehabilitation with the Barthel Index and Pearson calculation. Results: The calculation of Cronbach's alpha yielded values of 0.97. Kappa values were between 0.89 and 0.96, values of percentage agreement between 96% and 99%, ICC values between 0.94 and 0.99, and SEM values between 0.08 and 0.17. Pearson (r) was 0.88. The factor analysis revealed a one-factor solution. Conclusion: The modified Care Dependency Scale for Rehabilitation shows a high internal consistency, as well as a high level of equivalence. Furthermore, criterion-related validity and construct validity have shown satisfying results.