Single arm trials with a synthetic control arm built from RWD

Fri, Sept 17, 2021, 09:00 – 10:00AM (PDT), 12:00 – 1:00PM (EST)

Abstract: 

Randomized clinical trials (RCT) are the gold standard for approvals by regulatory agencies. However, RCT's are increasingly time consuming, expensive, and laborious with a multitude of bottlenecks involving volunteer recruitment, patient truancy, and adverse events. An alternative that fast tracks clinical trials without compromising quality of scientific results is desirable to more rapidly bring therapies to consumers. We propose a model-based approach using nonparametric Bayesian common atoms models for patient baseline covariates. This specific class of models has two critical advantages in this context: (i) The models have full prior support, i.e., allow to approximate arbitrary distributions without unreasonable restrictions or shrinkage in specific parametric families; (ii) inference naturally facilitates a reweighting scheme to achieve equivalent populations. We prove equivalence of the synthetic and other patient cohorts using an independent separate verification. Failure to classify a merged data set using a flexible statistical learning method such as random forests, support vector machines etc. proves equivalence. We implement the proposed approach in two motivating case studies.

Speaker:

Peter Mueller is Professor of Statistics and Mathematics at UT Austin. He obtained a doctorate in statistics from Purdue University in 1991. He served on the faculty in ISDS, at Duke University (1991-2001), in the Department of Biostatistics at M.D. Anderson Cancer Center (2001-2011) and at the Department of Mathematics and the Department of Statistics and Data Science at University of Texas at Austin (2011-current), and as Adjunct Professor at the Department of Biostatistics at M.D. Anderson Cancer Center (2011-current). Mueller has published over 150 refereed articles in statistics, biostatistics and other journals. He is an elected fellow of the ASA, the IMS, and ISBA, recipient of the Zellner award (ISBA) and served as president of ISBA, and as chair of ASA/SBSS. He works on Bayesian inference, with a focus on nonparametric Bayesian methods, simulation based methods, optimal design and multiple comparison procedures. He is interested in applications in biostatistics and bioinformatics, including in particular Bayesian clinical trial design, hierarchical models, population PK/PD models, inference for histone modifications and tumor heterogeneity. He is an elected fellow of the ASA, the IMS, and ISBA, and served as president of ISBA, and as chair of ASA/SBSS.

Strategic Alliance: