Compare Model-Based CIF to the Aalen-Johansen Estimator

Computes the nonparametric Aalen-Johansen CIF (via cmprsk) for each competing-risks cause, overlays it on the fitted Beta-Danish CIF, and returns the Aalen-Johansen times/estimates, the fitted CIF values on a common grid, and Gray's CIF-equality test where applicable.

Usage

cif_compare(fit, tmax = NULL, n_grid = 160, plot = TRUE)

Arguments

fit

A fitted object of class "bd_competing".

tmax

Optional upper time for evaluation; default the 95th percentile of observed times.

n_grid

Number of time points on the evaluation grid (default 160).

plot

Logical; if TRUE (default) draws a panel of overlays.

Value

A list with elements tgrid, cif_fit (data frame long format), cif_aj (data frame long format) and optionally gray_test.

Details

Requires cmprsk (Suggests).

Examples

if (FALSE) { # \dontrun{
set.seed(1)
T1 <- rbetadanish(200, 1.2, 1.5, 1.0, 0.4)
T2 <- rbetadanish(200, 1.0, 2.0, 1.0, 0.2)
C  <- stats::rexp(200, 0.05)
time  <- pmin(T1, T2, C)
cause <- ifelse(time == C, 0L, ifelse(T1 <= T2, 1L, 2L))
fit <- fit_bd_competing(time = time, cause = cause)
cif_compare(fit)   # requires cmprsk to be installed
} # }