Compute jackknife pseudo-observations of the cause-specific cumulative incidence for competing risks

jackknife.competing.risks2(object, times, cause, mr)

Arguments

object: A survfit object, with competing events
times: Times at which the cumulative incidence is computed, must be length 1
cause: Value indicating for which cause the cumulative incidence is to be computed, it must match one of the values available in object (see example)
mr: Model response, the result of a call to Surv, or a matrix with two columns: "time" (observed follow up time) and "status" (0 = censored, 1, ..., k = event types)

Value

A vector of jackknifed pseudo-observations of the cause-specific cumulative incidence at time times

Examples


sfit.cuminc <- survival::survfit(survival::Surv(etime, event) ~ 1, data = mgus2)
mrs <- with(mgus2, Surv(etime, event))
pseudo.obs <- jackknife.competing.risks2(sfit.cuminc, times = 200, cause = "pcm", mrs)
mean(pseudo.obs)
#> [1] 0.09190984
# agrees with
summary(sfit.cuminc, times = 200)
#> Call: survfit(formula = survival::Surv(etime, event) ~ 1, data = mgus2)
#> 
#>  time n.risk n.event Pr((s0)) Pr(pcm) Pr(death)
#>   200    124     938    0.225  0.0919     0.683