Dementia in Parkinson's disease: usefulness of the pill questionnaire.

The Level I algorithm for the diagnosis of dementia associated with Parkinson's disease (PD-D) recommended by the Movement Disorder Society task force includes a Pill Questionnaire to determine the impact of cognitive decline on daily activities. The objective of this study was to test the performance of the Pill Questionnaire as a screening tool for the detection of dementia (all-cause) in patients with PD and to test the performance of another functional scale substituting the Pill Questionnaire for the diagnosis of "probable PD-D" (pPD-D). Data were collected from 529 patients who had PD in Hoehn and Yahr stages 1 through 5. The measures used include the Scales for Outcomes in PD-Motor (SCOPA-Motor), scales for psychiatric complications, the Mini Mental State Examination, the Clinical Impression of Severity Index, and the Pill Questionnaire. The SCOPA-Motor functional subscale score was categorized as "impact" or "no impact" of PD on daily activities. According to clinical judgment, 13.3% of patients had dementia. For detecting dementia, the Pill Questionnaire had 89% accuracy, although its positive predictive value was 55%. Performance was worse with the categorized SCOPA-Motor subscale. According to the Movement Disorder Society task force criterion, 85 patients (16.1%) had pPD-D. When the Pill Questionnaire was substituted by the categorized SCOPA-Motor subscale, the modified algorithm showed sensitivity, specificity, and accuracy indexes over 90% but had positive predictive value of 66% for pPD-D diagnosis. Although the Pill Questionnaire demonstrated acceptable basic properties as a screening tool for dementia, its positive predictive value was low. The SCOPA-Motor subscale cannot be proposed as a substitute for the Pill Questionnaire.