repeated measures with phreg

[msecic]

I have a dataset where the patients can be on up to two drugs.  So I have the time-to-event variable as my outcome and I want to compare the two drugs but account for the fact that a patient can be on either drug at a particular time.  How would I program this?

Thanks,

Michelle Secic

consult@secicstats.com

[Reeza]

You'd need to provide more information. A time dependent variable was already suggested and depending on your trial that could be the appropriate way.

Depending on the clinical context it may also make sense to separate it into 3 drugs, ie drug a, drug b, drug A/B.