Screening and diagnostic validity of the comprehensive cognitive domain assessment scale for mild cognitive impair-ment in the Chinese population with Parkinson disease
Objective To develop an appropriate comprehensive scale for mild cognitive impairment in Parkinson dis-ease(PD-MCI)due to various diagnostic methods for PD-MCI and a lack of consensus on the applicable neuropsychological scales for PD-MCI.Methods A total of 92 non-dementia PD patients with a confirmed diagnosis were enrolled in this study,and all patients were assessed using the comprehensive scale of five cognitive domains recommended by the Chinese version of PD-MCI guidelines and Movement Disorder Society(MDS).Based on the norms of healthy controls assessed by the same scale,the patients were divided into non-cognitive impairment PD group and PD-MCI group.The t-test was used to observe the consistency between groups;the receiver operating characteristic(ROC)curve was used to investigate the cut-off value of each scale with the highest sensitivity and specificity in identifying the impairment in each cognitive domain of PD-MCI;a binary lo-gistic regression analysis was used to determine the specific contents of the comprehensive cognitive domain assessment scale for PD-MCI.Results This study showed that the ten-point comprehensive cognitive domain assessment scale based on long-time memory in Brief Visuospatial Memory Test,Clock Drawing Test,digit span backward test,Trail Making Test A,Verbal Flu-ency Test,Trail Making Test B,Similarities,Boston Naming Test,long-time recall memory in Auditory Verbal Learning Test,and short-time recall memory in Wechsler Memory Scale-IV Logical Memory Test had a significant diagnostic validity in predict-ing PD-MCI,with an area under the ROC curve of 0.971(95%CI 0.941‒1.000).Conclusion The recommended cut-off value can be used to perform a subtype analysis of PD-MCI and observe the outcome and clinical specificity of different subtypes of PD-MCI,and therefore,it has a certain clinical significance.
国家科技期刊平台
NSTL
万方数据
