R help - Jul 2004 - Bootstrap of Kaplan-Meier Estimate for clustered data

Is there an R-procedure to calculate a Kaplan Meier estimate with 
confidence bounds for correlated survival times (with correlated groups of 
observations  which can be accounted for with the "cluster" argument
in the
cox regression). I would need it to estimate the survival time of dental 
implants, where each patient has several implants.

Thanks,

Martin

I would like to take advantage of the  vectorized 'ifelse' command.
Given a
filter number of 'filter', if the nm'th value in a n x m matrix I
would
like to perform the following.

newMatrix <- ifelse(myMat <=filter, filter+1, "existing nm'th
value").

What do I use to express the "existing nm'th value" in this
expression if I
want to test an nm th value is less than 'filter', if false use the 
existing nm'th value.

This does not seem to be documented in the help().

Peter W.