fastkm: fastkm
In mbrueckner/AHR: Estimation and Testing of Average Hazard Ratios

Description Usage Arguments Details Value See Also Examples

Fast Kaplan-Meier estimator

1 2	fastkm(time, status, ltrunc = rep.int(0, length(time)), left.limit = FALSE, eval = time)

`time`	vector of right-censored survival times
`status`	censoring indicator for each element of `time` (0 = right-censored, 1 = event)
`ltrunc`	vector of left-truncation times
`left.limit`	indicates wether estimated survival function is left continuous
`eval`	points at which the estimated survival function should be evaluated

This function calculates the Kaplan-Meier estimator for right-censored survival data, at arbitrary time points. It can handle left-truncated and/or right-censored data with ties. Avoids the overhead of the survfit or prodlim functions by stripping away most of the features not needed here.

A list containing the vectors time, surv and variance, and n.atrisk

survfit and prodlim

T <- rexp(100)
C <- rexp(100)
Y <- pmin(T, C)
D <- T <= C
sort(fastkm(Y, D)$surv, decreasing=TRUE)
# should be exactly the same as
fit <- survfit(Surv(Y, D) ~ 1)
f <- approxfun(fit$time, fit$surv, f=0, rule=2, yleft=1)
f(fit$time)