Dear members,
I have a data set where the hazard proportionality assumption does not hold for a Cox regression. The failures were also much tied. I tried to use both the "TVC" and "exactm" options. I got an error message that "TVC" and "exactm" options can't be used simultaneously. However, the data seem more meaningful with "exactm" option.
My question is to know if the "exactm" option also handles the issue of non proportionality of the risk. In case it does not, how to handle both non PH and tied failures issues in Stata?
Thanks for your help!
Eric
I have a data set where the hazard proportionality assumption does not hold for a Cox regression. The failures were also much tied. I tried to use both the "TVC" and "exactm" options. I got an error message that "TVC" and "exactm" options can't be used simultaneously. However, the data seem more meaningful with "exactm" option.
My question is to know if the "exactm" option also handles the issue of non proportionality of the risk. In case it does not, how to handle both non PH and tied failures issues in Stata?
Thanks for your help!
Eric