- 單位 : 統計學系
- 分類 : 演講公告
- 點閱 : 229
- 日期 : 2019-07-03
時 間：108年6 月4日（星期二下午14:10~15:00）
主講人：許根寧 首席生物統計師 (藥明康德)
講 題：A Group Sequential Multiple Comparisons for Confirmatory Clinical Trials
A pharmaceutical company may conduct a clinical trial considering more than one experimental treatment compared with the concurrent control. Such the trial design will arise the multiplicity issue, that is, if each comparison is performed at the significant level α, then the overall type I error cannot be maintained at α. For a clinical trial, especially a Phase III confirmatory clinical trial, strongly controlling the overall type I error rate at α is required by the associated authority. In addition, the group sequential design has been widely used in clinical trials, in which the hypothesis testing would be performed based on the available efficacy data in the course of trial. If the interim analysis presents the statistically significant result, the trial could be terminated early. Repeatedly performing the hypothesis testing based on the cumulative data will also arise the multiplicity issue. In this talk, I will illustrate the procedure of group sequential multiple comparisons with a real-life phase III confirmatory clinical trial. In the trial, two experimental treatments would be compared with the concurrent control, and one interim analysis would be performed. In this talk, I will also talk a little about the role of biostatistician in clinical trials.