Background: Drug-induced dyskinesias are a common and disabling clinical problem in the long-term management of Parkinson disease (PD). Their management and the development of new treatments rely on rigorous and meaningful dyskinesia measurement. Although clinician-based approaches exist, patient-based measures are limited.
Method: Potential rating scale items concerning daily activities affected by dyskinesias were generated from patients, literature review, and expert opinion. The resulting 42-item questionnaire was administered to 98 patients known to have problematic dyskinesias; 72 patients were invited to complete it twice for test-retest reliability (trt). Rasch analysis guided scale development. Results were cross-validated using traditional psychometric methods by examining scaling assumptions (item means and variances, item-total correlations), reliability (Cronbach alpha, trt), and validity (factor analysis). External validation was performed against standard dyskinesia measures: blinded video rating using modified Goetz and Abnormal Involuntary Movements Scales (AIMS), and Unified PD Rating Scale (UPDRS) questions 32-34.
Results: Response rates were high. Fourteen items were removed because of high missing data. The remaining items were Rasch analyzed. Two items were removed because of misfit. The resulting 26 items formed a clinically and statistically conformable set. Traditional psychometric criteria were satisfied and external validation showed good correlation with the UPDRS items and moderate to good correlation with objective dyskinesia measures.
Conclusion: The 26-item Parkinson Disease Dyskinesia Scale (PDYS-26) satisfied multiple criteria for reliable and valid measurement. Correlations with objective measures suggest that it captures related but not identical constructs. As a patient-derived scale that generates linear measurements, it could complement existing clinician-based dyskinesia measures.