Center for Networked Biomedical Research in Neurodegenerative Diseases (CIBERNED), Carlos III Institute of Health, Madrid, Spain.
Department of Statistics, Center of Human and Social Sciences, Spanish Council for Scientific Research, Madrid, Spain.
Mov Disord. 2020 Jun;35(6):969-975. doi: 10.1002/mds.28002. Epub 2020 Mar 27.
The primary validation of the Movement Disorder Society Non-Motor Rating Scale was recently published, but 2 important structural analyses were not included. The objective of this study was to examine the structural characteristics of the Movement Disorder Society Non-Motor Rating Scale by factor and cluster analyses.
Data came from the validation study, an international multicenter cross-sectional study of 402 Parkinson's disease patients. Demographic and clinical data, the Movement Disorder Society Non-Motor Rating Scale, and Hoehn and Yahr staging were used. Exploratory and confirmatory factor analysis and nonhierarchical cluster analysis were performed.
The exploratory factor analysis showed that all 13 domains of the Movement Disorder Society Non-Motor Rating Scale, except 1, and the Non-Motor Fluctuations subscale performed as unidimensional (variance explained: 0.36, sleep and wakefulness; -0.82, orthostatic hypotension). The confirmatory factor analysis could be carried out in 9 domains and showed that 6 of them and the Non-Motor Fluctuations subscale adjusted to the model satisfactorily according to the root mean square error of approximation. Furthermore, all domains had comparative fit index values >0.95, except depression and pain (both, 0.94) and sleep and wakefulness (0.90). The Non-Motor Fluctuations subscale showed satisfactory root mean square error of approximation (0.07), but a low comparative fit index value (0.91). A 5-cluster solution, correctly classifying 96% of the cases, was found.
Overall, most subscales of the Movement Disorder Society Non-Motor Rating Scale are unidimensional, and although each subscale is distinct in terms of content covered, factors and clusters that are of clinical relevance are discernible and contribute to our understanding of possible nonmotor subtypes in Parkinson's disease. © 2020 International Parkinson and Movement Disorder Society.
运动障碍协会非运动症状评分量表的初步验证最近已经公布，但有 2 个重要的结构分析并未包含在内。本研究的目的是通过因子和聚类分析来检验运动障碍协会非运动症状评分量表的结构特征。
数据来自验证研究，这是一项针对 402 例帕金森病患者的国际多中心横断面研究。使用了人口统计学和临床数据、运动障碍协会非运动症状评分量表、Hoehn 和 Yahr 分期。进行了探索性和验证性因子分析以及非层次聚类分析。
探索性因子分析显示，运动障碍协会非运动症状评分量表的除 1 个以外的所有 13 个领域和非运动波动子量表都表现为单维性（方差解释：0.36，睡眠和觉醒；-0.82，直立性低血压）。验证性因子分析可以在 9 个领域进行，并表明其中 6 个领域和非运动波动子量表根据近似均方根误差调整到模型中令人满意。此外，除抑郁和疼痛（均为 0.94）和睡眠和觉醒（0.90）外，所有领域的比较拟合指数值均>0.95。非运动波动子量表的近似均方根误差（0.07）令人满意，但比较拟合指数值较低（0.91）。发现了一个 5 聚类的解决方案，可以正确分类 96%的病例。
总体而言，运动障碍协会非运动症状评分量表的大多数子量表都是单维的，尽管每个子量表在涵盖的内容方面都有所不同，但可以辨别出具有临床相关性的因素和聚类，这有助于我们理解帕金森病中可能存在的非运动亚型。© 2020 国际帕金森病和运动障碍协会。
