Preliminary Psychometric Review of Neurologists' Speech Ratings on the Unified Parkinson's Disease Rating Scale.

PURPOSE: The purpose of this study was to conduct a preliminary psychometric evaluation of the speech item of the Movement Disorder Society-sponsored revision of the Unified Parkinson's Disease Rating Scale (MDS-UPDRS) Part III. METHOD: Twenty neurologists specializing in movement disorders evaluated the speech-unscripted and sentence reading-of people with Parkinson's disease (PD) on the MDS-UPDRS Part III. The concurrent validity of the speech item was evaluated by comparing the ratings to a previously collected estimate of the speakers' intelligibility during sentence reading. The reproducibility of the ratings was evaluated in terms of reliability (intra- and interrater) and agreement. RESULTS: MDS-UPDRS Part III speech ratings based on sentence reading were moderately and negatively correlated with previously collected estimates of intelligibility. No such relationship was found for MDS-UPDRS Part III speech ratings based on unscripted speech. Individual correlations for each rater ranged widely, with some showing weak, negligible, or positive relationships with intelligibility. Intrarater reliability was high, but interrater reliability was fair for both speech tasks. Movement disorders specialists agreed on MDS-UPDRS speech ratings, on average, less than 70% of the time. CONCLUSIONS: A critical communication outcome-intelligibility during unscripted speech-is not captured by MDS-UPDRS Part III speech ratings, suggesting weak concurrent validity of this widely used metric of speech function. The reproducibility of MDS-UPDRS Part III speech ratings across providers is fair. These speech ratings may be insufficient to assess the speech function of people with PD and identify those in need of speech services.