Session Information
Date: Monday, October 8, 2018
Session Title: Rating Scales
Session Time: 1:15pm-2:45pm
Location: Hall 3FG
Objective: To investigate the measurement characteristics of the Unified Dyskinesia Rating Scale (UDysRS) using item response theory (IRT) approaches.
Background: The UDysRS has become the most frequently used assessment of levodopa induced dyskinesia (LID) in Parkinson’s disease. The UDysRS has 4 parts, each focused on one domain of dyskinesia, Historical Disability of ON Dyskinesia (Part 1); Historical Disability of OFF Dystonia (Part 2); Objective Impairment (Part 3) and Objective Disability (Part 4). Clinimetric examinations have demonstrated acceptable levels of internal consistency, inter-rater reliability, factor structure, sensitivity to change, and differential item functioning. To date, IRT clinimetric examinations have not be conducted. With this methodology, each UDysRS item is tested to determine how well it measures the dyskinesia domain of interest.
Methods: The Reference Standard database for the UDysRS was used, providing complete examinations on 253 patients with Parkinson’s disease. A graded response model IRT analysis with maximum likelihood parameter estimation was employed to examine item discrimination and item threshold. Item Fit S_X2 statistics were used to examine the adequacy of fit for individual items to the overall domain measurement.
Results: Item discrimination estimates were significant for all items in the UDysRS, and by Part, with Part 1, 3, 4, 2 ordered as the highest to lowest discrimination values. Item thresholds demonstrated adequate coverage of the summary domain of LID for all items. Individual S_X2 demonstrated adequate fit for most measures except the item assessing Handwriting (Part 1) and all four OFF Dystonia items (Part 2).
Conclusions: The UDysRS demonstrates acceptable levels of item discrimination, item threshold and item characteristic curves for all items with the exception of items assessing Historical Disability of OFF Dystonia. These findings suggest the 0 to 4 Likert-type scaling metric appears appropriate for items assessing on-dyskinesia, but alternate scaling might be indicated for items of Part 2.
To cite this abstract in AMA style:
C. Goetz, S. Luo, G. Stebbins. Item Response Theory Analysis of the Unified Dyskinesia Rating Scale Items [abstract]. Mov Disord. 2018; 33 (suppl 2). https://www.mdsabstracts.org/abstract/item-response-theory-analysis-of-the-unified-dyskinesia-rating-scale-items/. Accessed July 3, 2025.« Back to 2018 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/item-response-theory-analysis-of-the-unified-dyskinesia-rating-scale-items/