Methods for designing and analyzing multiple arms survival trials that incorporate patient's treatment choice are needed. In these trials, patients are randomized into two groups, random and choice. Participants in the choice group choose their treatment, which is not a current standard practice in randomized clinical trials. In this paper, we propose a new method based on the likelihood function to design and analyze these trials with time to event outcomes in the presence of non-informative right censoring. We use simulations to evaluate the methods for Weibull outcomes, complete and censored. Finally, we provide an illustration for designing a study in which we discuss some design considerations and demonstrate the methods.
Keywords: Likelihood function; censoring; clinical trial designs; preference; survival; treatment choice.
© 2023 Informa UK Limited, trading as Taylor & Francis Group.