报告时间：2025年11月14日 09:30-10:30

报告地点：苏州大学本部精正楼306

报告人：李扬，中国人民大学

摘要：

In the rapidly evolving domain of clinical research, multi-arm trials have increasingly emerged as pivotal methodologies for simultaneously assessing the efficacy of multiple therapeutic interventions. Typically, participants in multi-arm randomized controlled trials are allocated to various treatment arms or a control arm through complete randomization (CR). However, a significant limitation of CR is its inability to reliably achieve balanced distributions of baseline covariates among groups. To address this challenge, we propose a Sequential Multi-treatment Adaptive Randomized Trial (SMART) design. Specifically tailored to realistic clinical trial settings, where timely enrollment and immediate treatment allocation are critical, the SMART design facilitates the sequential assignment of participants based on their order of recruitment.  Through comprehensive numerical simulations, we demonstrate that the proposed SMART design exhibits substantial improvements over traditional randomization approaches with respect to covariate balance, computational efficiency, precision in estimating the average treatment effect, and enhanced statistical power. Moreover, applying the SMART design to a three-arm clinical trial aimed at comparing therapeutic strategies for chronic obstructive pulmonary disease further demonstrates its superior ability to achieve covariate balance and enhance the precision of statistical estimates.

报告人简介：

李扬，中国人民大学吴玉章特聘教授、博士生导师，学校交叉科学学术委员会副主任，入选国家级青年人才项目；担任国际统计学会Elected Member、中国商业统计学会副会长、中国统计学会常务理事、中国现场统计研究会常务理事、北京生物医学统计与数据管理研究会监事长等；主要从事模型选择与不确定性评价、复杂调查设计与分析、潜变量建模、试验设计与推断等领域研究，在国内外知名期刊发表论文九十余篇，承担国家自然科学基金、教育部重大项目、全国统计科学研究重大项目等。

邀请人：马学俊