[R] censor=FALSE and id options in survfit.coxph

[Terry Therneau]

Yes, it is a bug.  Thanks for providing a complete example.  I'll look into it, but leave 
for a week's vacation in a couple of hours and have some other pressing tasks.

Terry T.


Terry,

I recently noticed the censor argument of survfit.  For some analyses it greatly reduces the size of the resulting object, which is a nice feature.

However, when combined with the id argument, only 1 prediction is made.  Predictions can be made individually but I'd prefer to do them all at once if that change can be made.

Chris