[R-sig-eco] Mean survival time (and possibly its confidence interval) for interval-censored data (using "interval" R package)

[Pierre THIRIET]

Dear all,

I would like to get mean survival time (and possibly its confidence 
interval) for interval-censored data.
Using "interval package", I am fitting the non-parametric maximum 
likelihood estimate for the distribution (i.e. generalization of the 
Kaplan-Meier estimate). But then, I don't know how to compute the mean 
survival time from the icfit object.
For getting a raw estimate of the mean survival times, I plot the NPLME 
and look at the time interval where survival porbability equals 0.5. Is 
there a better way for getting mean survival time, and is there a way 
for getting its confidence interval?

library(interval)
data(bcos)
icout<-icfit(Surv(left,right,type="interval2")~treatment, data=bcos)
plot(icout)
abline(h=0.5)

I thank you in advance,
Pierre