Optimal Clinical Trial Designs for Personalized Medicine

• Activity Number: 88
• Type: Contributed
• Date/Time: Sunday, July 31, 2011 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Hongjian Zhu*+ and Hongyu Zhao and Feifang Hu
• Companies: Yale University and Yale University and University of Virginia
• Keywords: Clinical trials ; Optimal allocation ; covariates ; linear models ; randomization design

Abstract

Personalized medicine is a rapidly advancing field of healthcare which tailors treatments to individual patients and promises greater precision and effectiveness than traditional medicine. As the development of genomics, personalized medicine is increasingly relying on the genetic data, since mutation is believed to be the root causes of many complex diseases, such as cancer, heart disease, and diabetes. In this paper, we study a multiple regression model specially designed for genomics-based personalized medicine. Our objective is to find the optimal allocation in terms of efficiency and implement the optimal allocation by adaptive randomization procedures. Based on a new perspective on analyzing linear models, we derive the optimal allocation in various types of linear models. A new adaptive procedure is used to target the optimal allocation. The proposed method is easy to interpret and implement in applications. Simulation results show that the proposed method is superior to traditional randomization procedures in terms of efficiency.