Measuring functional disability in early rheumatoid arthritis: the validity, reliability and responsiveness of the Recent-Onset Arthritis Disability (ROAD) index

Disability has been identified as a core outcome measure in rheumatoid arthritis (RA). The aim of this study was to test the Recent-Onset Arthritis Disability (ROAD) questionnaire for validity, reliability and responsiveness in Italian patients with early RA.

The psychometric properties of ROAD were tested in 159 patients with early RA, mean age 54.7 (+/- 8.8), 74.3% women, mean disease duration 14.5 months (+/- 1.9 months). All completed the ROAD, the Medical Outcomes Study SF-36 Health Survey (SF-36), the Health Assessment Questionnaire (HAQ) and the patient global assessment (PGA) of functional disability twice, in order to test for validity and responsiveness. Of the 159 patients who completed the health status instruments on two occasions, 121 were included in the responsiveness analyses. The test-retest reliability of the ROAD questionnaire was calculated using intraclass correlation coefficients (ICCs) and the Bland and Altman method on 77 patients who completed the questionnaire twice over an interval of one week. Construct validity was assessed using Spearman's correlations, while responsiveness was evaluated by 3 different methods: (1) effect size (the mean difference between the baseline scores and thefollow-up scores divided by the standard deviation of the baseline scores); (2) standardized response mean (the mean change in scores divided by the standard deviation of the change in scores); (3) receiver operating characteristics (ROC) curve analysis.

ROAD fulfilled the established criteria for validity, reliability and responsiveness. In comparison with the SF-36, the expected correlations were found when comparing items measuring similar constructs, thus supporting the convergent construct validity. Significant correlations were seen between ROAD scores and HAQ scores (rho = 0.372), SF-36 physical component summary (PCS) (rho = -0.413), PGA functional disability (rho = 0.417), pain (rho = 0.639), Ritchie index (rho = 0.357), number of swollen joints (rho = 0.387), patient and physician assessment of disease activity (rho = 0.467 and 0.323, respectively), and Disease Activity Score (rho = 0.476). Test-retest reliability was satisfactory, with ICCs of 0.927 (upper extremity function), 0.892 (lower extremity function), and 0.851 (activity of daily living/work). Bland-Altman plots confirmed this finding. The results of responsiveness analysis indicate that the ROAD subscales were slightly more sensitive to perceived change in functional disability than those of HAQ, SF-36 PCS, and PGA offunctional disability.

Our data suggest that the ROAD index is a reliable, valid and responsive tool for measuring physical functioning in patients with early RA, and is suitable for use in clinical trials and daily clinical practice. Its generalizability and utility for assessing aggressive treatment and functional outcomes must now be evaluated in broader settings.