how to interpret coefficient of TVCs interacting with log(time)

[John Daddo]

Dear friends,

Happy new year! Really glad to have found this site. I am running a Cox model as follows:

stset time, id(id) failure(death==1)

stcox x1 x2, tvc(x3 x4) nohr texp(ln(_t))

I would tremendously appreciate any suggestions on how to compute the percentage change in the hazard rate of death attributable to a change in the value of x3 which is a time-varying indicator variable on whether the patient received a medicine administered at different times. Are there any ready commands in Stata for this task? Thank you so much!

[John Daddo]

Dear friends,

Happy new year! Really glad to have found this site. I am running a Cox model as follows:

stset time, id(id) failure(death==1)

stcox x1 x2, tvc(x3 x4) nohr texp(ln(_t))

I would tremendously appreciate any suggestions on how to compute the percentage change in the hazard rate of death attributable to a change in the value of x3 which is a time-varying indicator variable on whether the patient received a medicine administered at different times. Are there any ready commands in Stata for this task? Thank you so much!

Also, my data consist of multiple records of about 500 patients, observed over multiple points of time. Many thanks again!

[John Daddo]

Sorry all, I think I have solved my problem. I really did not need to use tvc here.