Cox model IPCW cumulative incidence estimate

Estimates the marginal cumulative incidence of event type 1 using a weighted Aalen-Johansen estimator, where the weights are the IPCW weights stored in the p_notcens column (as added by add_ipcw_weights_cr()).

Usage

cuminc_ipcw_cr(
  data_long,
  esttimes = seq(from = 0, to = 10, length.out = 100),
  extend = TRUE,
  covariate = "z1"
)

Arguments

data_long: A long-format data frame with IPCW weights, as returned by add_ipcw_weights_cr(). Must contain columns tstart, tstop, delta, id, and p_notcens.
esttimes: Numeric vector of times at which to return estimates. Defaults to 100 equally spaced points from 0 to 10.
extend: Logical. If FALSE, estimates beyond the minimum of the stratum-specific maximum follow-up times are set to NA. Default is TRUE.
covariate: Character string. Name of the covariate column, used only when extend = FALSE to compute the truncation time. Default is "z1".

Value

A numeric vector of IPCW cumulative incidence estimates at esttimes.

Examples

set.seed(42)
dat <- sim_data_cr(n = 200, censoring = "baseline")
dat_long <- wide_to_long_cr(dat)
dat_long <- add_ipcw_weights_cr(dat_long, strat = "no")
#> Warning: Loglik converged before variable  3 ; beta may be infinite. 
cuminc_ipcw_cr(dat_long, esttimes = seq(0, 5, 0.5))
#>  [1] 0.0000000 0.2525084 0.3656652 0.4349979 0.4489364 0.4768644 0.4768644
#>  [8] 0.4768644 0.4847719 0.4847719 0.4847719