Hi!

I have done a Cox regression on survival data. Unfortunately, the Schoenfeld residuals are not constant over time. In my model, therefore, I have included an interaction term between exposure and follow up time which significantly differs from 1 meaning there is an annual change to the HR I obtain from the model (which is the HR at time 0). If I instead want to center the model at say 5 years or 10 years to get the HR at this specific follow up time. Does anyone know how to do this?


Best wishes,

David Bergman