Simulate AFT data with an outcome-dependent right-censored covariate

Generates a simulated dataset under an accelerated failure time (AFT) model with a right-censored covariate subject to outcome-dependent censoring. The function returns the true covariate X, censoring time C, observed covariate W = min(X, C), event indicator D = I(X <= C), and several derived quantities such as weights and conditional expectations.

Usage

data_aft(nSubjects = 10^3)

Arguments

nSubjects

Integer. Number of subjects to simulate. Defaults to 10^3.

Value

A data frame with nSubjects rows containing:

y

Continuous outcome.

Z

Standard normal covariate.

D

Event indicator I(X <= C).

A

Observed auxiliary covariate.

X

True covariate subject to censoring.

AX

A - X.

C

Censoring time.

W

min(X, C).

AW

A - W.

meanCYZ

Mean of C given (Y, Z) on the original scale.

meanXZ

Mean of X given (Z) on the original scale.

b

Subject ID.

myp_uniform

Random weights (Uniform(0.1, 0.9)).

myp_ywz_oracle

Oracle survival probability P(C >= W | Y, Z).

myp_ywz

Estimated survival probability using a Weibull AFT model.

meanXD0YZ

\(E[X | D = 0, Y, Z]\) computed by numerical integration.

AX_yz

A - X if D = 1, otherwise A - E[X | D = 0, Y, Z].

AX_z

A - X if D = 1, otherwise A - E[X | Z].