VALIDATION DATA

Test‐retest reliability (intraclass correlation coefficients) over a 2‐week period in the original study was 0.84 for the impairment/footwear subscale and 0.96 for the activity limitation/participation restriction subscale.

Candidate questionnaire items were derived from qualitative interviews with patients with rheumatoid arthritis and foot problems and then tested by patients with rheumatoid arthritis in the presence of an interviewer, ensuring content validity.

Selection of the final subscale items was performed by fitting them to the Rasch unidimensional measurement model. The 21 items in the final impairment/footwear subscale displayed a good fit to the Rasch model and no significant deviation from model expectation across the trait (overall mean item of −0.25 [SD 1.07], mean person of −0.34 [SD 0.71], and item‐trait interaction χ2 statistic of 45.07; P = 0.35 [42 degrees of freedom]). Similarly, the 30 items in the activity limitation/participation restriction subscale had a good fit to the Rasch model, with no significant deviation from model expectation (overall mean item of −0.38 [SD 0.93], mean person of −0.29 [SD 0.80], and item‐trait interaction χ2 statistic of 64.4; P = 0.18 [60 degrees of freedom]). Neither scale showed differential item function for age, sex,or disease duration.