[R] How to bootstrap Kaplan-Miere estimator with 95% envelope

Frank E Harrell Jr f.harrell at vanderbilt.edu
Mon Sep 13 13:18:14 CEST 2004

wanr at ucalgary.ca wrote:
> Hi all,
> Given a typical right-censoring data which contains a time variable and a 
> censoring indicator. How do we bootstrap samples to obtain a 95% envelope 
> for the estimated cumulative hazard function?
> Thanks in advance.
> Rui

bootkm in the Hmisc package does not compute confidence intervals for 
the whole curve, only for one time point or one quantile of survival 
time, but it shows how to bootstrap K-M quickly using survfit.km in the 
survival package.

Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University

More information about the R-help mailing list