重复测量数据的常用统计分析方法  

Statistical methods for repeated measurement data in scientific research

作　　者：冯国双[1] Feng Guoshuang(Big Data and Engineering Research Center,Beijing Children′s Hospital,Capital Medical University/National Center for Children′s Health,Beijing 100045,China)

机构地区：[1]国家儿童医学中心首都医科大学附属北京儿童医院大数据和工程研究中心,北京100045

出　　处：《中华预防医学杂志》2020年第7期804-812,共9页Chinese Journal of Preventive Medicine

摘　　要：重复测量数据是医学中常见的一种数据,该数据的分析不能简单对每个时间点分别比较,而应采用专业的统计分析方法。本文介绍了目前重复测量数据常见的三种方法,即重复测量方差分析、广义估计方程和多水平模型,并基于实例对三种方法的具体软件实现、结果解读进行了详细解释,同时对三种方法实际应用的情形进行了分析比较,以期临床科研人员能够正确分析重复测量数据,提高数据的分析效率。Repeated measurement data is a common type of data in medicine,whichcan not be simply compared at each time point,and a professional statistical analysis method should be used to analysis this kind of data.Three common statistical methods were introduced for repeated measurement data,including repeated measurement analysis of variance,generalized estimation equations and multilevel models.The implementation of specific software and related results for the three methods based on some cases were also explainedin the article.Additionally,we compared the actual application of the three methods,in order to help clinical researchers to analyze repeated measurement data correctly and to improve their efficiency of data analysis.

关 键 词：模型 统计学 数据说明 统计 重复测量数据 

分 类 号：R195.1[医药卫生—卫生统计学]