Hi everybody,
I'm fairly new to survival analysis with R and have some questions how to
apply and interpret the coxph and related functions:
I have time-dependent covariates with several measurements per subject with
constant delta t. The covariates change in each time step.
I fitted the following model:
fit <- coxph(Surv(start, stop, event) ~ ratePO + rateC + BLamp + BLP80 +
cluster(subjectID), data=dat)
and get
n= 1081, number of events= 10
coef exp(coef) se(coef) robust se z
Pr(>|z|)
ratePO -0.50189 0.60539 0.25195 0.17696 -2.836 0.004565 **
BLamp -0.05340 0.94800 0.02877 0.01470 -3.632 0.000281 ***
rateC 0.82888 2.29076 0.38111 0.30110 2.753 0.005908 **
BLP80 0.19425 1.21440 0.11377 0.04320 4.497 6.9e-06 ***
Rsquare= 0.035 (max possible= 0.056 )
Likelihood ratio test= 38.62 on 4 df, p=8.362e-08
Wald test = 46.71 on 4 df, p=1.756e-09
Score (logrank) test = 31.03 on 4 df, p=3.022e-06, Robust = 13.91
p=0.007582
Now for a single subject I want to use this model in the sense of a
classifier to predict the event probability at time t.
Do I have to use the function predict.coxph? Which of the types is the right
one? Or can I also use survfit for this purpose?
I know these are pretty basic questions, but I'm kind of lost here.
Any comments would be of great help.
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
