Model mis-specification and overestimation of the intraclass correlation coefficient in cluster randomized trials

被引:12
|
作者
Giraudeau, B
机构
[1] Fac Med Tours, INSERM, CIC 202, F-37032 Tours, France
[2] Univ Paris 07, INSERM, Biostat & Epidemiol Clin, U 717, Paris, France
关键词
bias; cluster randomized trial; intraclass correlation coefficient; mixed-effects model; one-way random effects model;
D O I
10.1002/sim.2260
中图分类号
Q [生物科学];
学科分类号
07 ; 0710 ; 09 ;
摘要
Intraclass correlation coefficient (ICC) estimates must be provided when reporting the results of a cluster randomized trial. This study demonstrates that estimating this parameter with one-way ANOVA and an underlying mixed-effects statistical model leads to biased estimates. The bias depends on the effect size of the studied treatment. Copyright (c) 2005 John Wiley & Sons, Ltd.
引用
收藏
页码:957 / 964
页数:8
相关论文
共 50 条