cox.ipw: Missing data IPW Cox

In scheike/timereg: Flexible Regression Models for Survival Data

Missing data IPW Cox

Description

Fits an Cox-Aalen survival model with missing data, with glm specification of probability of missingness.

Usage

cox.ipw(
  survformula,
  glmformula,
  d = parent.frame(),
  max.clust = NULL,
  ipw.se = FALSE,
  tie.seed = 100
)

Arguments

survformula	a formula object with the response on the left of a '~' operator, and the independent terms on the right as regressors. The response must be a survival object as returned by the ‘Surv’ function. Adds the prop() wrapper internally for using cox.aalen function for fitting Cox model.
glmformula	formula for "being" observed, that is not missing.
d	data frame.
max.clust	number of clusters in iid approximation. Default is all.
ipw.se	if TRUE computes standard errors based on iid decompositon of cox and glm model, thus should be asymptotically correct.
tie.seed	if there are ties these are broken, and to get same break the seed must be the same. Recommend to break them prior to entering the program.

Details

Taylor expansion of Cox's partial likelihood in direction of glm parameters using num-deriv and iid expansion of Cox and glm paramters (lava).

Value

returns an object of type "cox.aalen". With the following arguments:

iid	iid decomposition.
coef	missing data estiamtes for weighted cox.
var	robust pointwise variances estimates.
se	robust pointwise variances estimates.
se.naive	estimate of parametric components of model.
ties	list of ties and times with random noise to break ties.
cox	output from weighted cox model.

Author(s)

Thomas Scheike

References

Paik et al.

Examples

### fit <- cox.ipw(Surv(time,status)~X+Z,obs~Z+X+time+status,data=d,ipw.se=TRUE)
### summary(fit)

scheike/timereg documentation built on Jan. 25, 2023, 3:43 p.m.