time2event (version 0.1.0)

tcomp.risk: Competings Risks Regression with time-to-event data as covariates.

Description

Fits a semiparametric model for the cause-specific quantitie with time-to-event data as covariates.

Usage

tcomp.risk(formula, na.time=c("remove","censor"), verbose=FALSE, data = sys.parent(), cause, times = NULL, 	Nit = 50, clusters = NULL, est = NULL, fix.gamma = 0, gamma = 0, 	n.sim = 0, weighted = 0, model = "fg", detail = 0, interval = 0.01, 	resample.iid = 1, cens.model = "KM", cens.formula = NULL, 	time.pow = NULL, time.pow.test = NULL, silent = 1, conv = 1e-06, 	weights = NULL, max.clust = 1000, n.times = 50, first.time.p = 0.05, 	estimator = 1, trunc.p = NULL, cens.weights = NULL, admin.cens = NULL, 	conservative = 1, monotone = 0, step = NULL)

Arguments

formula

a formula object, with the response on the left of a '~' operator, and the terms on the right. The response must be a survival object as returned by the `Event' function. The status indicator is not important here. Time-invariant regressors are specified by the wrapper const(), and cluster variables (for computing robust variances) by the wrapper cluster(). In case that time-to-event data are covariates, use the wrapper time() to indicate the time-to-event data as covariates.

na.time

a missing-data filter function for time-to-event covariates. The option 'remove' will remove all the data with 'NA', while the option 'censor' will treat the missing data as censored and then replace with the logest time. Default is 'remove'.

verbose

logical. Should R report extra information on progress? Default is 'FALSE'.

data

a data.frame with the variables.

cause

For competing risk models specificies which cause we consider.

times

specifies the times at which the estimator is considered. Defaults to all the times where an event of interest occurs, with the first 10 percent or max 20 jump points removed for numerical stability in simulations.

Nit

number of iterations for Newton-Raphson algorithm.

clusters

specifies cluster structure, for backwards compability.

est

possible starting value for nonparametric component of model.

fix.gamma

to keep gamma fixed, possibly at 0.

gamma

starting value for constant effects.

n.sim

number of simulations in resampling.

weighted

Not implemented. To compute a variance weighted version of the test-processes used for testing time-varying effects.

model

"additive", "prop"ortional, "rcif", or "logistic".

detail

if 0 no details are printed during iterations, if 1 details are given.

interval

specifies that we only consider timepoints where the Kaplan-Meier of the censoring distribution is larger than this value.

resample.iid

to return the iid decomposition, that can be used to construct confidence bands for predictions

cens.model

specified which model to use for the ICPW, KM is Kaplan-Meier alternatively it may be "cox"

cens.formula

specifies the regression terms used for the regression model for chosen regression model. When cens.model is specified, the default is to use the same design as specified for the competing risks model.

time.pow

specifies that the power at which the time-arguments is transformed, for each of the arguments of the const() terms, default is 1 for the additive model and 0 for the proportional model.

time.pow.test

specifies that the power the time-arguments is transformed for each of the arguments of the non-const() terms. This is relevant for testing if a coefficient function is consistent with the specified form $A_l(t)=beta_l t^time.pow.test(l)$. Default is 1 for the additive model and 0 for the proportional model.

silent

if 0 information on convergence problems due to non-invertible derviates of scores are printed.

conv

gives convergence criterie in terms of sum of absolute change of parameters of model

weights

weights for estimating equations.

max.clust

sets the total number of i.i.d. terms in i.i.d. decompostition. This can limit the amount of memory used by coarsening the clusters. When NULL then all clusters are used. Default is 1000 to save memory and time.

first.time.p

first point for estimation is pth percentile of cause jump times.

n.times

only uses 50 points for estimation, if NULL then uses all points, subject to p.start condition.

estimator

default estimator is 1.

trunc.p

truncation weight for delayed entry, P(T > entry.time | Z_i), typically Cox model.

cens.weights

censoring weights can be given here rather than calculated using the KM, cox or aalen models.

admin.cens

censoring times for the administrative censoring

conservative

set to 0 to compute correct variances based on censoring weights, default is conservative estimates that are much quicker.

monotone

monotone=0, uses estimating equations montone 1 uses

step

step size for Fisher-Scoring algorithm.

Value

returns the same object as that of comp.risk(). See comp.risk() for details

Details

The funciton tcomp.risk is an extention of the function comp.risk for time-to-event covariates. If the model has no time-to-event covariates, tcomp.risk will print the warning sign 'No time-varying covariate!!!' and then do exactly the same procedure as comp.risk does. If the model has time-to-event covariates, the time-to-event covaraites should be wrapped with time() by placing the right-hand side of a ~ operator. In particular, the wrapper time(a1,b1,a2,b2,a3,b3,...) will be used with time-to-event covariates, where $ai$ and $bi$, $i=1,2,...$ are time-to-event and status, respectively. See comp.risk for other details.

References

S. Kim (2016). time2event: an R package for the analysis of event time data with time-to-event data as covariates. Wayne State University/Karmanos Cancer Institute. Manuscript.

Examples

  data(bmtelder)

  # convert to data with time-to-event data as covariates
  # nrm with cgvhd
  tnrm2data = time2data(c("nrm.t","nrm.s"),c("cgvhd.t","cgvhd.s"),bmtelder)$data

  # no time-varying analysis with 'comp.risk'
  set.seed(3927)
  cr2r = comp.risk(Event(nrm.t,nrm.s)~cgvhd.s+cond+donor,data=bmtelder,
  			cause=1,resample.iid=1,n.sim=1000,model="additive")
  cr2r.pred = predict(cr2r,X=1)
  plot(cr2r.pred)

  # time-varying analysis with 'comp.risk'
  set.seed(3927)
  nt.cr2r = comp.risk(Event(start,end,nrm.s)~cgvhd.s+cond+donor,data=tnrm2data,
  			cause=1,resample.iid=1,n.sim=1000,model="additive")
  nt.cr2r.pred = predict(nt.cr2r,X=1)
  plot(nt.cr2r.pred)

  # time-varying analysis with 'tcomp.risk'
  set.seed(3927)
  t.cr2r = tcomp.risk(Event(nrm.t,nrm.s)~time(cgvhd.t,cgvhd.s)+cond+donor,data=bmtelder,
  			cause=1,resample.iid=1,n.sim=1000,model="additive")
  t.cr2r.pred = predict(t.cr2r,X=1)
  plot(t.cr2r.pred)