Interrater and Intrarater Reliability of the TWSTRS ‐Delphi Motor Severity Scale for Cervical Dystonia

BACKGROUND: Cervical dystonia (CD) motor severity is typically quantified with the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS). An expanded version of the Motor Severity section of the TWSTRS ("TWSTRS-Delphi") has been previously evaluated for its clinimetric properties but not for inter- and intrarater reliability. OBJECTIVE: In this study we quantified inter- and intrarater reliability of the Motor Severity section of the TWSTRS-Delphi for CD. METHODS: Ten raters familiar with the TWSTRS-Delphi and ten raters unfamiliar with it independently scored standardized videos from 90 patients with CD. We computed one-way random intraclass correlation coefficient (ICC, absolute agreement) for total and item scores. We repeated assessments and calculations with a subset of raters to assess intrarater reliability. RESULTS: Interrater reliability for the Total score was good for Familiar raters (ICC = 0.697) and fair for Unfamiliar raters (ICC = 0.566). Most individual items achieved fair-excellent interrater agreement in both groups (Familiar 0.415-0.763; Unfamiliar 0.416-0.769). Sagittal Shift, Duration, and Sensory Tricks showed poor agreement. Intrarater reliability was fair-excellent for most items in both groups (Familiar 0.494-0.927; Unfamiliar 0.413-0.877), but poor for Sagittal Shift and Duration among Familiar raters. CONCLUSIONS: The TWSTRS-Delphi yields reproducible Total and item scores across movement disorder raters. The data pinpoint Sagittal Shift, Duration, and Sensory Tricks as the principal sources of between-rater variability, defining immediate priorities for scale modification and focused training.