I am researching factors associated with timely access to medical care and treatment for breast cancer patients. The two response variables of interest are "time for the first consultation" and "time for the first treatment after consultation". These two are measured on same individual as well. I have modeled these responses seperately using Cox proportional hazards model.

Now I am searching for a bivariate response model that can be used in this case. I couldn't find one. Does anyone has an idea on this?

  • $\begingroup$ Why can't you just describe the two models separately? They are addressing fundamentally different things. $\endgroup$
    – AdamO
    Commented Jun 7, 2018 at 16:23

1 Answer 1


The appropriate method for your data is that recurrent event survival analysis. There are a various of method but you could use cox with frailty to account for the correlation .


You could refer to the following paper: Penalized Survival Models and Frailty by Terry M. Therneau, Patricia M. Grambsch and V. Shane Pankratz

The paper explains the kidney data example in R code.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.