|
Abstract:
|
With the advent of immunotherapy, in which non-proportional hazard is present, traditional logrank test no longer provides optimal power in detecting treatment effect, and Cox proportional-hazard model no longer appropriately describes treatment effect. Many researches have proposed novel and powerful tests, but they may rely on prior knowledge of the survival curve of the experimental drug. In this presentation, we explore a new method that directly models the non-proportional hazard ratio function. Simulation results from sensitivity analysis and comparisons with commonly used methods will be presented.
|
