Dear scholars, in my thesis I have two event of interest, inhospital complication and death. At conception, I thought once patients developed inhospital complication they will be censored for death....when read on survival analysis I found this action will inflate hazard ratio and will overestimate the incidence of the second outcome of interest.....please can you help me on how to approach this data? thank you
It is included in Stata's survival models, which version do you have? the command is stcrreg.
Version 13...I think the command is for competing risk analysis. Ok, I inserted inhospital complication as failure variable and death as competing risk, then when I run for result it shows me I made an error...says competing risks events must be stset as censored....how can I handle this?
Behailu,
You should have only one variable with all possible failures
ex.
failuretype, where
0= censored
1= failure
2= competing event
then you stset your data as usual
stset timevar, failure(failuretype==1), time0(time0var)
And in the regression command you define the competing event
stcrreg indepvar1 indepvar 2 indepvarN , compete(failuretype==2)
- Badges
- Science topic
- Similar topics
- Psychology