mets (version 1.3.7)

resmean.phreg: Restricted mean for stratified Kaplan-Meier or Cox model with martingale standard errors

Description

Restricted mean for stratified Kaplan-Meier or stratified Cox with martingale standard error. Standard error is computed using linear interpolation between standard errors at jump-times. Plots gives restricted mean at all times. Years lost can be computed based on this and decomposed into years lost for different causes using the cif.yearslost function that is based on integrating the cumulative incidence functions. One particular feature of these functions are that the restricted mean and years-lost are computed for all event times as functions and can be plotted/viewed. When times are given and beyond the last event time withn a strata the curves are extrapolated using the estimates of cumulative incidence.

Usage

resmean.phreg(x, times = NULL, covs = NULL, ...)

Arguments

x

phreg object

times

possible times for which to report restricted mean

covs

possible covariate for Cox model

...

Additional arguments to lower level funtions

Author

Thomas Scheike

Examples

library(mets)
data(bmt); bmt$time <- bmt$time+runif(408)*0.001
out1 <- phreg(Surv(time,cause!=0)~strata(tcell,platelet),data=bmt)

rm1 <- resmean.phreg(out1,times=10*(1:6))
summary(rm1)
par(mfrow=c(1,2))
plot(rm1,se=1)
plot(rm1,years.lost=TRUE,se=1)

## comparing populations, can also be done using rmstIPCW via influence functions
rm1 <- resmean.phreg(out1,times=40)
e1 <- estimate(rm1)
e1
estimate(e1,rbind(c(1,-1,0,0)))

## years.lost decomposed into causes
drm1 <- cif.yearslost(Event(time,cause)~strata(tcell,platelet),data=bmt,times=10*(1:6))
par(mfrow=c(1,2)); plot(drm1,cause=1,se=1); plot(drm1,cause=2,se=1);
summary(drm1)

## comparing populations, can also be done using rmstIPCW via influence functions
drm1 <- cif.yearslost(Event(time,cause)~strata(tcell,platelet),data=bmt,times=40)
summary(drm1)
## first cause 
e1 <- estimate(drm1)
estimate(e1,rbind(c(1,-1,0,0)))