Margreth Grotle, Philip Wilkens, Andrew M. Garratt, Inger Scheel, Kjersti Storheim
FORMI (Communication Unit for Musculoskeletal Disorders), Clinic for Surgery and Neurology (C1), Oslo University Hospital, Oslo University Hospital, Kirkeveien 166, NO-0407 Oslo, Norway. E-mail: margreth.grotle@medisin.uio.no
DOI: 10.2340/16501977-1166

Background: The Roland-Morris Disability Questionnaire (RDQ) is one of the most frequently used and recommended outcome measures for patients with low back pain.
Objective: To examine the fit of data from 4 different versions of the RDQ to a Rasch model in a Norwegian sample of patients with chronic low back pain and degenerative lumbar osteoarthritis.
Methods: Patients with chronic low back pain and degenerative lumbar osteoarthritis completed the RDQ prior to treatment in a secondary healthcare clinic. Data were analysed using a dichotomous Rasch model.
Results: Of 250 included patients, 243 patients with a mean age of 48. 5 years completed all 24 items of the RDQ. None of the 4 RDQ versions (the original 24-item, the 18-item versions of Williams and Stratford, and the 11-item of Stroud) were a unidimensional measure of disability due to low back pain. Items 3 and 23 were redundant and items 13 and 18 did not fit the Rasch model. Several items showed differential item functioning, indicating that the items performed differently in subgroups of the sample.
Conclusion: In the absence of consistent findings across studies that have evaluated the RDQ by Rasch analysis, caution should be exercised in the development and application of alternative versions of the RDQ.