Hypothesis Testing of Covariate-Adaptive Randomized Clinical Trials Under Generalized Linear Models

• Activity Number: 598
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Wei Ma*+ and Feifang Hu
• Companies: University of Virginia and University of Virginia
• Keywords: Covariate-adaptive design ; clinical trials ; logistic regression ; conservative tests ; Pocock and Simon's marginal procedure ; stratified permuted block design

Abstract

Covariate-adaptive designs provide a way to balance important covariates in randomized clinical trials. Recently, it has been demonstrated theoretically that the traditional hypothesis testing is usually too conservative under linear regression model. In this talk, we further study the properties of hypothesis testing based on generalized linear models, especially on logistic regression for binary responses, under covariate-adaptive designs. Theoretically we prove that hypothesis testing is usually conservative in terms of small type I errors under the null hypothesis for logistic regression under a large class of covariate-adaptive designs, which includes the many popular covariate-adaptive designs. Numerical studies are performed to assess Type I errors and power comparison. Some adjusting methods are also studied and recommended.