Search

Purpose: To validate the 23-item Work Instability Scale for Rheumatoid Arthritis (RA-WIS) for use in osteoarthritis (OA) using both classical test theory and item response theory approaches. Methods: Baseline and 12-month follow-up data were collected from workers with OA recruited from community and clinical settings ( n = 130). Fit of RA-WIS data to the Rasch model was evaluated by item- and person-fit statistics (size of residual, chi-sq), assessments of differential item functioning, and tests of unidimensionality and local independence. Internal consistency was assessed by KR-20. Convergent construct validity (Spearman r, known-groups) was evaluated against theoretical constructs that assess impact of health on work. Responsiveness to global indicators of change was assessed by standardized response means (SRM) and area under the receiver operating characteristic curves. Results: Data structure of the RA-WIS showed adequate fit to the Rasch model ( chi- sq = 83.2, P = 0.03) after addressing local dependency in three item pairs by creating testlets. High internal consistency (KR-20 = 0.93) and convergent validity with work-oriented constructs (| r| = 0.55–0.77) were evident. The RA-WIS correlated most strongly with the concept of illness intrusiveness ( r = 0.77) and was highly responsive to changes (SRM = 1.05 [deterioration]; −0.78 [improvement]). Conclusions: Although developed for RA, the RA-WIS is psychometrically sound for OA and demonstrates interval-level property. [ABSTRACT FROM AUTHOR]