ecif.pred: Excess CIF prediction based on newdata and geepack::geese...

In cribosch/matchsurv: Excess Risk Estimation for Matched Cohort Survival Data

Description

Excess CIF prediction based on newdata and geepack::geese estimates.

Usage

ecif.pred(model = NULL, times, formula, dataset, strata.levels = NULL,
  coefs = NULL, vcov = NULL, link = "log")

Arguments

model	geese object. To be defined if coefs and vcov are null.
times	vector of timepoints as the one used to estimate the GEE model
formula	model formula
dataset	new data
strata.levels	if CIF predicted for different strata, define strata levels
coefs	coefficient estimates (model$beta). To be specified if model is NULL
vcov	coefficient variance and covariance matrix (model$vbeta). To be specified if model is NULL, together with coefs.
link	link from the gee model

Value

dataset with predicted values; ready to be used with ggplot2

Examples

dcif<-sim.data.MatchCR(1000,5)
tp<-c(0.5,1,2,5,10,15,25)
setdcif1<-compcomp(timereg::Event(time=FALSE,time2=time,cause=cause)~X1+X2, 
                   data=dcif, cluster=i, idControl=j, time.points=tp,
                   cens.formula=NULL, event=1)
exc.cif.mod1<-geepack::geese(Rt~-1+factor(h)+X1+X2,
                    data=setdcif1,
                    family="gaussian", #error distribution
                    mean.link = "log", #link function for Rt
                    corstr="independence", #correlation structure
                    id=clust.num, #cluster vector
                    weights=weights #censoring weights 
                    )
### prediction: 
af<-paste0("-1+factor(h)+X1+X2") #model formula
newd<-data.frame(expand.grid(h=tp,X1=c(0,1),X2=c(0.8,1.5,2.5))) # newdata
# define the different subjects for whom the excess risk is predicted
strata.levels<-factor(1:6, levels=1:6,
                      labels =paste0(rep("X1=",6),
                      expand.grid(X1=c(0,1),X2=c(0.8,1.5,2.5))[,1], 
                      rep(", X2=",6),
                      expand.grid(X1=c(0,1),X2=c(0.8,1.5,2.5))[,2]))
pred.exc.cif<-ecif.pred(exc.cif.mod1,times = tp,dataset = newd,
                        formula = af, strata.levels = strata.levels)
head(pred.exc.cif,8)

cribosch/matchsurv documentation built on Aug. 15, 2019, 11:55 a.m.