Can you get confidence intervals for a cause-specific cumulative incidence with survfit?

58 Views Asked by At

I have a competing risks data set and am trying to obtain, for a specific set of explanatory variables, the cumulative incidence at a specific time and a confidence interval for the value.

Using the rotterdam dataset from the survival package as an example:

library(survival)
rott <- rotterdam

rott <- rott %>%
  mutate(time=pmin(dtime,rtime)) %>%
  mutate(outcome = ifelse(death==1|recur==1,1,0)) %>%
  mutate(status = ifelse(death==1,2,ifelse(recur==1,1,0)))

model <- coxph(Surv(time,as.factor(status))~hormon+age, data=rott, id=pid)

predict <- survfit(model, newdata=c(hormon=1, age=50))

summary(predict,time=2000)$pstate

This gives me the cumulative incidence for the competing risks at time 2000, but I can't see how to obtain a confidence interval for this.

In the non-competing risk setting for the survival I know you can use $upper and $lower but trying to find the equivalent here.

0

There are 0 best solutions below