作者: Hui Quan , Xuezhou Mao , Yujun Wu , Meehyung Cho , Peng-Liang Zhao
DOI: 10.1080/23737484.2015.1088419
关键词:
摘要: ABSTRACTCardiovascular event trials can be designed to allow adaptation based on interim results. In this article, we first discuss a strategy either accept noninferiority at an analysis or continue the study establish superiority. We demonstrate overall type I error rate control even when nominal significance level is applied both assessments two different datasets. also develop methodology for using blinded data timeline prediction via conditional probability future given available information piecewise exponential distribution. A cardiovascular trial used illustrate methods.