Sample Size and Power Computations Methods for Two-Stage Randomized Trial, with Focus on Time-to-Event Data

• Activity Number: 63 - Power and Sample Size: Methods and Applications
• Type: Contributed
• Date/Time: Monday, August 3, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Rouba Chahine* and Inmaculada Aban and Dustin Long
• Companies: University of Alabama at Birmingham and University of Alabama at Birmingham and University of Alabama at Birmingham
• Keywords: Random Trial; Choice; Preference

Abstract

A standard component of randomized clinical trials is informed consent, in which participants are made aware of all procedures or treatments. This knowledge may decrease compliance or reduce tolerance for inconveniences or difficulties when participants receive non-preferred treatments. A two-stage randomized design (2SRCT) incorporates participants’ preference. First, participants are randomized to the random (RT) or choice (CT) group. Second, RT group participants are randomized to one of two treatments; and those in the CT group choose between treatments. An inadequate sample size can lead to an underpowered study that fails to detect clinically relevant effects, while an unnecessarily large sample size can waste resources. Power analysis methods to calculate appropriate sample size for 2SRCT based on the ANOVA approach have been proposed for Normal and Binary outcomes. We investigate this approach to any outcome whose distribution satisfies regulatory conditions and use simulation to evaluate its performance for time-to-event data using uncensored and truncated Weibull distributions.