kyotil (version 2015.11-13)

make.timedep.dataset: Create Dataset for Time-dependent Covariate Proportional Hazard Model Analaysi

Description

Returns a data frame that is suitable for time-dependent covariate Cox model fit.

Usage

make.timedep.dataset(dat, X, d, baseline.ageyrs, t.1, t.2 = NULL)

Arguments

dat

data frame

X

string. Name of the followup time column in dat. Unit needs to be years.

d

string. Name of the followup time column in dat.

baseline.ageyrs

string. Name of the followup time column in dat.

t.1

numerical. Cutoff for age group

t.2

numerical. Second cutoff for age group

Value

Returns a data frame with the following columns added: tstart, tstop, .timedep.agegrp, .baseline.agegrp

tstart

left bound of time interval

tstop

right bound of time interval

.timedep.agegrp

time-dependent age group

.baseline.agegrp

baseline age group

Details

The function assumes that the followup length is such that only one change of age group is possible.

References

Therneau, T. and Crowson, C. Using Time Dependent Covariates and Time Dependent Coefficients in the Cox Model. A vignette from the R package surival. Fong, Y. and Halloran, M.E. Time-varying Age Group Analysis in Vaccine Efficacy Trials. In prep.

Examples

library(survival)

n=3000; followup.length=5; incidence.density=0.015; age.sim="continuous"

dat.0=sim.dat.tvarying.two(n, followup.length, incidence.density, age.sim, seed=1)
dat=subset(dat.0, for.non.tvarying.ana, select=c(ptid, X, d, baseline.age, trt))
dat.timedep = make.timedep.dataset (dat, "X", "d", "baseline.age", 6)
coxph(Surv(tstart,tstop,d) ~ trt*.timedep.agegrp, dat.timedep)