Hello,

I am trying to design a study in which the primary endpoint is time to event.

Due to some constraints, the study needs to be a non-inferiority one. How can I check for non-inferiority using a Cox model ? I know how to do it when I check mean difference, or other simple things, but log rank tests and Cox models seems complicated. I couldn't find any data online. Does anyone has experience with this kind of analysis ?

cheers