mets (version 1.3.4)

gofM.phreg: GOF for Cox covariates in PH regression

Description

Cumulative residuals after model matrix for Cox PH regression p-values based on Lin, Wei, Ying resampling.

Usage

gofM.phreg(
  formula,
  data,
  offset = NULL,
  weights = NULL,
  modelmatrix = NULL,
  n.sim = 1000,
  silent = 1,
  ...
)

Arguments

formula

formula for cox regression

data

data for model

offset

offset

weights

weights

modelmatrix

matrix for cumulating residuals

n.sim

number of simulations for score processes

silent

to keep it absolutely silent, otherwise timing estimate will be prduced for longer jobs.

...

Additional arguments to lower level funtions

Author

Thomas Scheike and Klaus K. Holst

Details

That is, computes $$ U(t) = \int_0^t M^t d \hat M $$ and resamples its asymptotic distribution.

This will show if the residuals are consistent with the model. Typically, M will be a design matrix for the continous covariates that gives for example the quartiles, and then the plot will show if for the different quartiles of the covariate the risk prediction is consistent over time (time x covariate interaction).

Examples

library(mets)
data(TRACE)
set.seed(1)
TRACEsam <- blocksample(TRACE,idvar="id",replace=FALSE,100)

dcut(TRACEsam)  <- ~. 
mm <- model.matrix(~-1+factor(wmicat.4),data=TRACEsam)
m1 <- gofM.phreg(Surv(time,status==9)~vf+chf+wmi,data=TRACEsam,modelmatrix=mm)
summary(m1)
if (interactive()) {
par(mfrow=c(2,2))
plot(m1)
}

m1 <- gofM.phreg(Surv(time,status==9)~strata(vf)+chf+wmi,data=TRACEsam,modelmatrix=mm) 
summary(m1)

## cumulative sums in covariates, via design matrix mm 
mm <- cumContr(TRACEsam$wmi,breaks=10,equi=TRUE)
m1 <- gofM.phreg(Surv(time,status==9)~strata(vf)+chf+wmi,data=TRACEsam,
		  modelmatrix=mm,silent=0)
summary(m1)