| Objective. To evaluate proposed diagnostic procedures for detecting Parkinson's disease dementia (PD-D).;Background. Dementia occurs in 48--78% of Parkinson's disease (PD) patients. Diagnostic procedures for PD-D based on recent clinical criteria have been proposed by a Movement Disorders Society (MDS) Task Force. A brief, practical, bedside eight-item screening checklist derived from simple assessments with suggested cutoff points has not yet been validated in comparison to results of more extensive neuropsychological testing.;Methods/Design. Subjects were recruited from those scheduled for neuropsychological testing at two specialty PD centers, administered by a neuropsychologist as part of routine clinical care. Independent and blinded clinical neurologists also administered the screening checklist. Kappa coefficient was used to compare classification or diagnostic outcomes of the checklist and the neuropsychological battery.;Results. We tested ninety-one PD subjects, with mean age 66.3 (SD=9.7) and mean disease duration 8.8 years (SD=6.1). Seven subjects (7.7%) met all eight checklist criteria for probable PD-D, as compared to 15 (16.5%) diagnosed or classified with probable PD-D using the full neuropsychological assessment. There was moderate agreement between these two methods for determination of PD-D (Kappa=0.59, p<0.001). The screening checklist showed 100% specificity and 46.7% sensitivity for diagnosing PD-D as compared to the full neuropsychological assessment. The missed PD-D cases were largely due to two checklist items (positive depression screening and Mini Mental State Examination (MMSE) scores ≥26).;Conclusions/Relevance. Where subjects met screening checklist criteria, the designation of PD-D was 100% specific. However, for cases that did not meet these criteria, full neuropsychological testing was still needed to differentiate PD-D from milder cognitive impairment. If positive for PD-D, the screening checklist may therefore be used with confidence, but if negative, full neuropsychological testing is required if cognitive impairment is clinically expected. In our series of cases, the screening battery would have saved approximately 7 hours of testing time, because full neuropsychological testing takes over one hour compared to 15 minutes for the screening battery. Modification of cutoff values for the problem items on the checklist may allow adjustments to optimize specificity and sensitivity of this screening tool. |
