I have a cox model with two main effects (x,y) and two interaction terms (xt, yt) where t is time. I am having some trouble interpreting the hazard ratios. Does the hazard ration become more of an instantaneous hazard rate when you have time-dependent coefficients in the model?

Thank you

  • $\begingroup$ Could you please say more about what you mean by the interaction terms with time and how you are representing them to the survival-analysis software? That’s a potential source of error as explained in this vignette on pages 19-20. Depending on what you mean by the interactions with time it might be impossible to interpret the hazard ratios in a useful way. $\endgroup$ – EdM May 31 '19 at 22:13
  • $\begingroup$ Notwithstanding the previous comment, there will not be one HR (unless the interaction effect = 0), the HR will depend on the time, HR(x,t) = h(x+1,t)/h(x,t) $\endgroup$ – adibender Jun 4 '19 at 13:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.