Objective: To assess differential item functioning (DIF) in the Unified Dyskinesia Rating Scale (UDysRS) to determine if the duration of Parkinson’s Disease (PD) and/or duration of dyskinesia substantially bias any item scores.

Background: Assessing DIF is a core validation step for rating scales. If DIF is present in an item, interpretation of the item must consider influences from the tested covariates. In this study, DIF occurs if item-score probability among patients with similar levels of dyskinesia differ by the duration of PD and/or duration of dyskinesia. DIF can be uniform (covariates have consistent levels of influence on item-score across all levels of dyskinesia) or non-uniform (covariates have varying levels of influence on item-score across different levels of dyskinesia).

Method: Using UDysRS translation databases (N=2,313), both uniform and non-uniform DIF from duration of PD and/or duration of dyskinesia were tested. Unidimensionality of UDysRS was first confirmed using confirmatory factor analysis and requiring the Confirmatory Fit Index (CFI) > 0.90 with Root Mean Square Error Approximation (RMSEA) < 0.10. DIF analysis was conducted using two independent latent models: the iterative hybrid ordinal logistic regression/item response theory (graded response model) approach implemented in R package lordif and the MIMIC model. DIF in an item was confirmed if both methods independently identify DIF at a significance level determined using a Bonferroni correction for multiple comparisons. McFadden pseudo R^2 was used to measure clinical relevancy (DIF magnitude) for items identified with DIF with the following cutoff points: <0.035=negligible, 0.035-0.07=moderate, >0.07=large. An item with DIF was considered to be clinically relevant if it exceeded the negligible rating. These procedures were repeated using the interaction between PD duration and dyskinesia duration. The combined impact of multiple items identified with DIF were examined using the differential test function (DTF) index.

Results: Most items did not show DIF based on PD and dyskinesia duration. For all items where DIF was identified, the magnitude statistic showed negligible effect.

Conclusion: The absence of significant DIF suggests the UDysRS can effectively measure dyskinesia severity without influence from the duration of PD and the duration of dyskinesia.

« Back to 2019 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/differential-item-functioning-in-the-unified-dyskinesia-rating-scale-udysrs-based-on-duration-of-parkinsons-disease-and-dyskinesia/