Interrater reliability of motor severity scales for hemifacial spasm

• Authors: Ha Yeon Lee, University of California, San Diego; Ingyun Park, University of California, San Diego; Minnie P. Luu, University of California, San Diego; Jerry Zhao, University of California, San Diego; Jeanne P. Vu, University of California, San Diego; Elizabeth Cisneros, University of California, San Diego; Brian D. Berman, Virginia Commonwealth University; Hyder A Jinnah, Emory University; Han-Joon Kim, Seoul National University; Catherine Y. Liu, University of California, San Diego; Joel S. Perlmutter, Washington University; Sarah Pirio Richardson, University of New Mexico; Anne Weissbach, University of Luebeck; Glenn T. Stebbins, Rush University; David A. Peterson, University of California, San Diego
• Journal Title: Journal of Neural Transmission
• Volume: Volume 130, Number 10
• Publisher: Springer Nature | 2023-07-19, Pages 1269-1279
• Type of Work: Article | Final Publisher PDF
• Publisher DOI: http://doi.org/10.1007/s00702-023-02667-w
• Subjects: Health Sciences, General
• Keywords: Severity rating scale; Hemifacial spasm; Interrrater reliability

Abstract

To compare the inter-rater reliability (IRR) of five clinical rating scales for video-based assessment of hemifacial spasm (HFS) motor severity. We evaluated the video recordings of 45 HFS participants recruited through the Dystonia Coalition. In Round 1, six clinicians with expertise in HFS assessed the participants’ motor severity with five scales used to measure motor severity of HFS: the Jankovic rating scale (JRS), Hemifacial Spasm Grading Scale (HSGS), Samsung Medical Center (SMC) grading system for severity of HFS spasms (Lee’s scale), clinical grading of spasm intensity (Chen’s scale), and a modified version of the Abnormal Involuntary Movement Scale (Tunc’s scale). In Round 2, clinicians rated the same cohort with simplified scale wording after consensus training. For each round, we evaluated the IRR using the intraclass correlation coefficient [ICC (2,1) single-rater, absolute-agreement, 2-way random model]. The scales exhibited IRR that ranged from “poor” to “moderate”; the mean ICCs were 0.41, 0.43, 0.47, 0.43, and 0.65 for the JRS, HSGS, Lee’s, Chen’s, and Tunc’s scales, respectively, for Round 1. In Round 2, the corresponding IRRs increased to 0.63, 0.60, 0.59, 0.53, and 0.71. In both rounds, Tunc’s scale exhibited the highest IRR. For clinical assessments of HFS motor severity based on video observations, we recommend using Tunc’s scale because of its comparative reliability and because clinicians interpret the scale easily without modifications or the need for consensus training.