A Rshiny Applet for Sequential, Multiple Assignment, Randomized Trials with a Survival Final Endpoint

• Activity Number: 388 - Statistical Methods in Handling Longitudinal Data and Temporal Effect
• Type: Contributed
• Date/Time: Wednesday, August 10, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Sasha Kravets* and Amy Ruppert and Jennifer Le-Rademacher and Sumithra Mandrekar
• Companies: Division of Epidemiology and Biostatistics, University of Illinois and Eli Lilly and Company and Mayo Clinic and Mayo Clinic
• Keywords: SMART; Adaptive Intervention ; RShiny; Power

Abstract

As personalized medicine continues to grow, adaptive interventions aim to personalize medicine through treatment sequences in which decisions to change therapy options are made using interim results to inform subsequent treatment decisions. Sequential, multiple assignment, randomized trial (SMART) designs are a method for building adaptive interventions. One of the goals of a SMART design is to compare adaptive treatment interventions, in which intermediate outcomes, denoted by tailoring variables, guide subsequent treatment decisions for individual patients. We focus our discussion on a SMART design with a binary tailoring variable and a survival final outcome. We discuss appropriate hazard rate assumptions for a SMART design, specifically, for a given first-stage therapy and prior to the tailoring variable assessment, hazard rates are assumed to be the same for all patients randomized to a treatment arm and can be established from historical data and calculated using an exponential survival model. After the tailoring variable assessment, individual hazard rates are assumed for each intervention path. We present a RShiny applet to estimate the power and sample size.