MoCA量表在中国人群中筛查轻度认知功能障碍最佳截断值的系统综述

发布时间：2018-05-02 16:10

本文选题：蒙特利尔认知评估量表 + 轻度认知功能障碍　；参考：《北京协和医学院》2017年硕士论文

【摘要】：目的蒙特利尔认知评估量表(MoCA)在使用汉语的华人筛查Petersen诊断的轻度认知功能障碍(MCI)准确性研究文献的系统综述,估算不同截断值下合并准确性指标并判断筛查MCI的最佳截断值。方法计算机检索 PubMed、EMbase、The Cochrane Library(2016 年 5期)、OVID、CNKI、维普、万方数据库和中国生物医学文献服务系统,搜集MoCA量表中文版本在中国中老年人群中以Petersen标准诊断MCI的相关诊断性试验,检索时限均为从2005年4月9日到2015年12月31日。由两位评价者独立筛选文献、提取资料,并采用QUADAS-2方法学工具评价纳入研究的质量后,对MoCA量表20/21-27/28截断值数据分别采用双变量模型(STATA14.0软件)和多变量模型(R3.3.2软件)进行合并准确性指标估计。综合考虑灵敏度、特异度及受试者工作特征曲线下面积、约登指数及上述准确性指标的可信区间,进行最佳截断值的判断。结果共纳入27个研究,这些研究的筛查背景包括社区人群和医院就诊者。这些研究共计5 755研究对象,各研究报告研究对象的平均年龄为60～80岁,按Petersen标准其中1 997人被诊断为MCI。对两个模型分析的MoCA20/21-27/28截断值准确性指标分别进行估计并比较判断最佳截断值。双变量模型最佳截断值判断为23/24-25/26,该区域敏感度分别为0.85(95%CI:0.66～0.94)、0.88(95%CI:0.77～0.94)、0.96(95%CI:0.93～0.97),特异度分别为 0.96(95%CI:0.80～0.99)、0.92(95%CI:0.74～0.98)、0.83(95%CI:0.75～0.89);多变量模型显示准确性各指标都比较高的截断值区域为 23/24-24/25,该区域敏感度分别为 0.83(95%CI:0.74～0.89)和 0.90(95%CI:0.85～0.94);特异度分别为 0.93(95%CI:0.87～0.96)、0.85(95%CI:0.75～0.91)。结论蒙特利尔认知评估量表在中国社区或医院就诊的中老年人群中筛查Petersen标准诊断的轻度认知障碍在23/24-25/26截断值区域有较高的诊断准确性。
[Abstract]:Objective to evaluate the accuracy of the Montreal Cognitive Assessment scale (MOCA) in the Chinese screening of mild cognitive impairment (MCI) in Chinese patients with mild cognitive impairment (MCI), and to estimate the combined accuracy indexes under different truncation values and to determine the best truncation value for screening MCI. Methods A computer-based search was conducted for the online database of OVIDC CNKI, Weip, Wanfang Database and the Chinese Biomedical document Service system in issue 5 of the Cochrane Library(2016. The Chinese version of the MoCA scale was collected for the diagnosis of MCI by Petersen standard in Chinese middle-aged and elderly population. The time limit for retrieval is from April 9, 2005 to December 31, 2015. The two reviewers independently sifted through the literature, extracted the data, and evaluated the quality of the studies by using the QUADAS-2 methodological tool. The data of 20 / 21-27 / 28 truncated values of MoCA scale were estimated by using bivariate model STATA14.0 software and multivariate model software R3.3.2 respectively. The best truncation value was determined by considering the sensitivity, specificity and area under the operating characteristic curve, the Jordan index and the confidence interval of the above accuracy index. Results A total of 27 studies were conducted with screening backgrounds including community populations and hospital visits. The average age of the subjects in each study was 6080 years old. According to Petersen criteria, 1 997 patients were diagnosed as MCIs. The MoCA20/21-27/28 truncation accuracy index of the two models is estimated and the best truncation value is compared. The best truncation value of the bivariate model is 23 / 24-25 / 26, the sensitivity of the region is 0.85 / 95, respectively. The sensitivity of the multivariable model is 0.85 / 95: 0.660.660.660.694/ 0.88 / 95CI: 0.770.770.770.94C: 0.939.97 respectively. The specificity of the model is 0.96 / 95: 0.800.990.992.95 / 95 CI: 0.74 / 0.980.95CIW 0.750.890.89; the multivariable model shows that the accuracy of each index is 2324-2425, and the sensitivity of this region is 0.839595 0.740.890.89) and 0.9995% 0.84%, respectively. The sensitivity of the multivariable model is 0.75% 0.75% 0.89%; the sensitivity of the multivariable model is 2324-2425%, and the sensitivity of the region is 0.8395 95% 0.74%. The specificity is 0.93% 0.87 0.96% 0.85 I 0.75 I 0.75 I 0.75 I 0.75 I 0.75 I 0.75 and I 0.75 I 0.75 I 0.75 0.91, respectively. Conclusion the Montreal Cognitive Assessment scale has a high diagnostic accuracy in the region of 23 / 24-25 / 26 truncation for mild cognitive impairment (MCI) in the elderly and middle-aged population in Chinese community or hospital.
【学位授予单位】：北京协和医学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R749.1

【参考文献】