mets (version 1.3.11)

sim_phreg: Simulation of Output from Cox Model

Description

Simulates data that looks like fit from a Cox model. Automatically censors data for the highest value of the event times by using cumulative hazard.

Usage

sim_phreg(
  cox,
  n,
  data = NULL,
  Z = NULL,
  rr = NULL,
  strata = NULL,
  entry = NULL,
  extend = TRUE,
  cens = NULL,
  rrc = NULL,
  ...
)

Value

Data frame with simulated event times, status, and covariates.

Arguments

cox

Output from coxph or phreg model fitting.

n

Number of simulations.

data

Data frame to extract covariates for simulations (draws from observed covariates).

Z

Design matrix instead of data.

rr

Vector of relative risks for Cox model.

strata

Vector of strata.

entry

Delayed entry variable for simulation.

extend

Extend possible stratified baselines to largest endpoint.

cens

Censoring specification (matrix = cumulative hazard, scalar = rate).

rrc

Relative risks for Cox-type censoring.

...

Arguments for rchaz (e.g., entry-time).

Author

Thomas Scheike

Examples

data(sTRACE)
nsim <- 100
coxs <-  phreg(Surv(time,status==9)~strata(chf)+vf+wmi,data=sTRACE)
set.seed(100)
sim3 <- sim_phreg(coxs,nsim,data=sTRACE)
head(sim3)
cc <-   phreg(Surv(time,status)~strata(chf)+vf+wmi,data=sim3)
cbind(coxs$coef,cc$coef)
plot(coxs,col=1); plot(cc,add=TRUE,col=2)

Z <- sim3[,c("vf","chf","wmi")]
strata <- sim3[,c("chf")]
rr <- exp(as.matrix(Z[,-2]) %*% coef(coxs))
sim4 <- sim_phreg(coxs,nsim,data=NULL,rr=rr,strata=strata)
sim4 <- cbind(sim4,Z)
cc <-   phreg(Surv(time,status)~strata(chf)+vf+wmi,data=sim4)
cbind(coxs$coef,cc$coef)
plot(coxs,col=1); plot(cc,add=TRUE,col=2)