论文部分内容阅读
多水平模型是既适用于计量资料、又适用于二分类或多分类资料的统计方法。本文通过对各类型的实例数据,分别应用多水平模型、协方差分析和CMH卡方检验方法进行分析,探讨多水平模型在多中心临床试验中心效应分析中的应用。结果表明,在计量资料的中心效应分析中,协方差分析较多水平模型更容易发现中心效应;对分类资料的中心效应,多水平模型更为敏感。可见不同分析方法对中心效应的分析结果不相同,在实际运用中应结合数据自身的结构特点和研究目的,并根据各种方法的适用条件选择分析方法。
The multi-level model is a statistical method that applies both to metrology data and to bi-class or multi-class data. In this paper, we use the multi-level model, covariance analysis and CMH chi-squared test to analyze the data of various types of cases to explore the application of multi-level model in the center effect analysis of multi-center clinical trials. The results show that in the central effects analysis of measurement data, the covariance analysis is more likely to find the central effect in more horizontal models; it is more sensitive to the central effects of classification data and the multi-level model. Therefore, different analysis methods have different results on the analysis of central effects. In practical application, the data should be combined with its own structural characteristics and research purposes, and the method of analysis should be selected according to the applicable conditions of various methods.