Abstract
Group sequential tests have been effective tools in monitoring long term clinical trials. There have been several popular discrete sequential boundaries proposed for modeling interim analysis of clinical trials under the assumption of Brownian motion for the stochastic processes generated from test statistics. In this paper, we study five sequential boundaries under the fractional Brownian motion. The fractional Brownian includes the classic Brownian motion as a special case. We proposed a Bayesian method on how to estimate Hurst parameter, used simulated data to investigate the performance of boundary-based hypothesis-testing under either Bm or fBm assumption in both with and without “treatment effect” scenarios.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have