Sample size and power calculations based on generalized linear mixed models with correlated binary outcomes
Sample size and power calculations based on generalized linear mixed models with correlated binary outcomes
复制标题
DOI：
10.1016/j.cmpb.2008.03.001
复制
发表时间：
2008-08-01
影响因子：
6.1
通讯作者：
Houck, Patricia R.
中科院分区：
文献类型：
--
作者：
Dang, Qianyu;Mazumdar, Sati;Houck, Patricia R.
广义线性混合模型（GLIMMIX）提供了一种强大的技术来模拟具有不同类型分布的相关结果。该模型现在可以很容易地实现与SAS PROC GLIMMIX在9.1版。对于二元结果，惩罚拟似然（PQL）或边际拟似然（MQL）的线性化方法为固定效应提供了相对准确的方差估计。使用GLIMMIX基于这些线性化方法，我们推导出公式的功效和样本量的纵向设计随着时间的推移磨损。我们发现，功效和样本量估计值取决于受试者内相关性和随机效应的大小。在这篇文章中，我们列出了纵向研究中通常用于检验假设的最小样本量表。模拟研究被用来比较结果。我们还提供了SAS宏的Web链接，我们开发了该宏来计算相关二元结果的功效和样本量。(C)2008爱思唯尔爱尔兰有限公司保留所有权利。
The generalized linear mixed model (GLIMMIX) provides a powerful technique to model correlated outcomes with different types of distributions. The model can now be easily implemented with SAS PROC GLIMMIX in version 9.1. For binary outcomes, linearization methods of penalized quasi-likelihood (PQL) or marginal quasi-likelihood (MQL) provide relatively accurate variance estimates for fixed effects. Using GLIMMIX based on these linearization methods, we derived formulas for power and sample size calculations for longitudinal designs with attrition over time. We found that the power and sample size estimates depend on the within-subject correlation and the size of random effects. in this article, we present tables of minimum sample sizes commonly used to test hypotheses for longitudinal studies. A simulation study was used to compare the results. We also provide a Web link to the SAS macro that we developed to compute power and sample sizes for correlated binary outcomes. (C) 2008 Elsevier Ireland Ltd. All rights reserved.