[R] predicting cumulative hazard for coxph using predict
Zainab Jamal
zainab_jamal at yahoo.com
Tue Feb 24 00:23:35 CET 2009
Hi
I am estimating the following coxph function with stratification and frailty where each person had multiple events.
m<-coxph(Surv(dtime1,status1)~gender+cage+uplf+strata(enum)+frailty(id),xmodel)
> head(xmodel)
id enum dtime status gender cage uplf
1 1008666 1 2259.1412037 1 MA 0.000 0
2 1008666 2 36.7495023 1 MA 2259.141 0
3 1008666 3 0.3026852 1 MA 2295.891 0
4 1008666 4 46.7493287 1 MA 2296.193 0
5 1008666 5 61.9772917 1 MA 2342.943 0
6 1008666 5 100.9964815 1 MA 2404.920 0
If I estimate predict(m,newdata,type="lp") does the output include xbeta or does it include xbeta+frailty term?
I want to predict the cumulative hazard for each person for their next event at a given time in future.What is the best way to compute it?
Currently I am doing it by pulling the different elements and putting them together using my own code.
For example, I get the cumulative baseline hazard from the basehaz function
bh<-basehaz(m)
bh$hazard for the strata in which the predicted event is and the time that matches to the prediction time.
xbeta from predict function
xbeta<-predict(m,newdata,type="lp")
fval<-m$frail
Regards
Zainab
More information about the R-help
mailing list