Reliability of the Modified Rankin Scale

Abstract

Background and Purpose— A perceived weakness of the modified Rankin Scale is potential for interobserver variability. We undertook a systematic review of modified Rankin Scale reliability studies. Methods— Two researchers independently reviewed the literature. Crossdisciplinary electronic databases were interrogated using the following key words: Stroke*; Cerebrovasc*; Modified Rankin*; Rankin Scale*; Oxford Handicap*; Observer variation*. Data were extracted according to prespecified criteria with decisions on inclusion by consensus. Results— From 3461 titles, 10 studies (587 patients) were included. Reliability of modified Rankin Scale varied from weighted κ=0.95 to κ=0.25. Overall reliability of mRS was κ=0.46; weighted κ=0.90 (traditional modified Rankin Scale) and κ=0.62; weighted κ=0.87 (structured interview). Conclusion— There remains uncertainty regarding modified Rankin Scale reliability. Interobserver studies closest in design to large-scale clinical trials demonstrate potentially significant interobserver variability.