Classical test theory and Rasch analysis validation of the Recent-Onset Arthritis Disability questionnaire in rheumatoid arthritis patients

Abstract

Disability has been identified as a core outcome measure in rheumatoid arthritis (RA). The aim of this study was to perform a comprehensive psychometric analysis of the Recent-Onset Arthritis Disability (ROAD) questionnaire in patients with RA. The questionnaire was completed by 583 patients with RA: 196 subjects participating in the NEW INDICES study and 387 subjects who were taking part in a long-term observational study. At confirmatory factor analysis for categorical data, data fit for a three-factor model was adequate to good (non-normed fit index = 0.98, comparative fit index = 0.99, root mean square error of approximation = 0.079, standardized root mean square residual = 0.047), with standardized item-to-factor loadings ranging from 0.60 to 0.90 and a cumulative explained variance of 83 %. The bifactor model of ROAD presented a clean independent cluster structure. The loadings in the unidimensional model were very similar to those on the general factor in the bifactor model. Rasch analysis showed a correct functioning of rating categories, a good fit of the data to the model for all three subscales, and satisfactory separation indexes and respective reliability (for both persons and items). This study, using both classical test theory and Rasch analysis methods, provides psychometric evidence of the reliability and internal and structural validity of ROAD in RA patients. Our results support the use of separate subscores for upper limb function, lower limb function, and activities of daily living/work, and the appropriateness of reporting an overall score (i.e., the mean of the three subscales).