报告人：邓宇昊（北京大学）

时间：2022年07月15日10:00-

腾讯会议ID：679 646 383

摘要：Clinical studies often encounter truncation by death, which may render certain outcomes undefined. Statistical analysis based solely on observed survivors may lead to biased results because the characteristics of survivors may differ between treatment groups. Under the principal stratification framework, a meaningful causal parameter called the survivor average causal effect is defined in the always-survivors subpopulation. This causal parameter may not be identifiable in observational studies, where treatment assignment and the survival or outcome processes are confounded by unmeasured factors. In this talk, we propose a novel method for handling unmeasured confounding when outcomes are truncated by death using a substitutional variable. The proposed method is applied to investigate the effects of allogeneic stem cell transplantation on leukemia relapse by appropriately adjusting transplantation-related mortality.

简介：邓宇昊，北京大学数学科学学院统计学专业博士候选人，本科毕业于北京大学数学科学学院概率统计系，主要研究方向为生物统计、因果推断、临床试验中的统计学方法，特别是被破坏的随机化试验和观察性研究，以第一作者身份在Biometrics、Statistics in Medicine等期刊发表多篇论文，引用量超200。

邀请人：穆春来

欢迎广大师生积极参与！