24-item Roland-Morris Disability Questionnaire was preferred out of six functional status questionnaires for post-lumbar disc surgery

R.W.J.G. Ostelo, H.C.W. de Vet, D.L. Knol, P.A. van den Brandt

Research output: Contribution to journalArticleAcademicpeer-review

83 Citations (Scopus)
277 Downloads (Pure)

Abstract

OBJECTIVE: Measurement properties of questionnaires should be based on samples of populations on whom these measurements will be used. The purpose of this study is to establish an evidence based recommendation regarding the use of functional status questionnaires in patients following a lumbar disc surgery by a direct comparison of the reproducibility and responsiveness. STUDY DESIGN AND SETTING: The measurement properties of six functional status questionnaires were assessed: 1) Roland-Morris Disability Questionnaire (RDQ-24), 2) Modified Roland-Morris Disability Questionnaire (MRDQ), 3) short Roland-Morris Disability Questionnaire (RM-18), 4) Physical Functioning scale, 5) Role Limitations-Physical scale of the SF-36, and 6) The Main Complaint (MC). Subjects (n=97) that still suffered residual complaints 6 weeks following a lumbar disc surgery completed the questionnaires before and 3 months after treatment. In a direct comparison the A) The test-retest reproducibility (Intraclass Correlation Coefficients [ICC] and the Standard Error of Measurement [SEM]) and B) 3 parameters of responsiveness (Minimal Detectable Change [MDC], Standardised Response Mean [SRM], and the Area Under the receiver operator characteristic Curve [AUC]) were assessed. RESULTS: This study suggests the superiority of the 3 versions of the RDQ compared to the 3 other questionnaires. Comparing the 3 versions of the RDQ reveals no substantial differences thereby indicating that the 2 modified version of the RDQ hold no better measurement properties in this specific population. CONCLUSION: The use of the RDQ-24 for this specific post-surgery population is suggested. The optimal cut-off point of the RDQ-24 that minimizes the overall classification error was found to be 3.5 with a sensitivity of 94.6% and a specificity of 88.2%.
Original languageEnglish
Pages (from-to)268-276
JournalJournal of Clinical Epidemiology
Volume57
Issue number3
DOIs
Publication statusPublished - 1 Jan 2004

Cite this