Objective: To investigate the measurement characteristics of the MDS-UPDRS using item response theory (IRT) approaches.

Background: The MDS-UPDRS has become the most frequently used assessment of PD.  Clinimetric examinations of the 4 parts of the scale have demonstrated acceptable levels of internal consistency, inter-rater reliability, factor structure, sensitivity to change, minimal clinical important change, and differential item functioning.  To date, the measurement characteristics have not been studied using IRT. 

Methods: The validated International MDS-UPDRS Translation database was used, providing complete examinations on 5793 persons from 14 languages.  A graded response model IRT analysis with maximum likelihood parameter estimation was employed to examine item discrimination, or the strength of the relationship between the item and the measured domain, and item threshold, or the level of item response to the overall severity of the measured domain for each Part of the MDS-UPDRS.  Item characteristic curves were used to examine the fit between scaling metric and overall domain severity.  Mplus (version 7) was used for these analyses.

Results: Item discrimination estimates were significant for all items in each of the 4 Parts of the MDS-UPDRS.  Item thresholds demonstrated adequate coverage of the individual domain for each Part with the exception of the item assessing Dopamine Dysregulation Syndrome in Part 1.  Individual item characteristic curves showed that the 0 to 4 Likert scaling approach provide a valid metric for assessing the individual Part domain severity for most items.  However, these curves for Part 3 tremor measures (kinetic, postural, resting) indicated that the 0 to 4 metric did not provide optimal information on overall domain of PD motor function.

Conclusions: The MDS-UPDRS demonstrates acceptable levels of item discrimination and item threshold for all items with the exception of the assessment of Dopamine Dysregulation Syndrome.   Additionally, the 0 to 4 Likert-type scaling metric appears appropriate for most items, although alternate scaling might be indicated for Part 3 tremor assessments.

« Back to 2017 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/item-response-theory-analysis-of-the-mds-updrs-items/