[R] How to specify "newdata" in a Cox-Modell with a time dependent interaction term?

[Jürgen Biedermann]

Dear Mr. Therneau, Mr. Fox, or to whoever, who has some time...

I don't find a solution to use the "survfit" function (package: 
survival)  for a defined pattern of covariates with a Cox-Model 
including a time dependent interaction term. Somehow the definition of 
my "newdata" argument seems to be erroneous.
I already googled the problem, found many persons having the same or a 
similar problem, but still no solution.
I want to stress that my time-dependent covariate does not depend on the 
failure of an individual (in this case it doesn't seem sensible to 
predict a survivor function for an individual). Rather one of my effects 
declines with time (time-dependent coefficient).

For illustration, I use the example of John Fox's paper "Cox 
Proportional - Hazards Regression for Survival Data".
http://cran.r-project.org/doc/contrib/Fox-Companion/appendix-cox-regression.pdf

Do you know any help? See code below

Thanks very much in advance
Jürgen Biedermann

#----------------------------------------
#Code

Rossi <- 
read.table("http://cran.r-project.org/doc/contrib/Fox-Companion/Rossi.txt", 
header=T)

Rossi.2 <- fold(Rossi, time='week',
     event='arrest', cov=11:62, cov.names='employed')

# see below for the fold function from John Fox

# modeling an interaction with time (Page 14)

mod.allison.5 <- coxph(Surv(start, stop, arrest.time) ~
     fin + age + age:stop + prio,
     data=Rossi.2)
mod.allison.5

# Attempt to get the survivor function of a person with age=30, fin=0 
and prio=5

newdata.1 <- 
data.frame(unique(Rossi.2[c("start","stop")]),fin=0,age=30,prio=5,Id=1,arrest.time=0)
fit <- survfit(mod.allison.5,newdata.1,id="Id")

Error message:

 >Fehler in model.frame.default(data = newdata.1, id = "Id", formula = 
Surv(start,  :
   Variablenlängen sind unterschiedlich (gefunden für '(id)')

--> failure, length of variables are different.

#-----------------------------------------------------------------
fold <- function(data, time, event, cov,
     cov.names=paste('covariate', '.', 1:ncovs, sep=""),
     suffix='.time', cov.times=0:ncov, common.times=TRUE, lag=0){
     vlag <- function(x, lag) c(rep(NA, lag), x[1:(length(x)-lag)])
     xlag <- function(x, lag) apply(as.matrix(x), 2, vlag, lag=lag)
     all.cov <- unlist(cov)
     if (!is.list(cov)) cov <- list(cov)
     ncovs <- length(cov)
     nrow <- nrow(data)
     ncol <- ncol(data)
     ncov <- length(cov[[1]])
     nobs <- nrow*ncov
     if (length(unique(c(sapply(cov, length), length(cov.times)-1))) > 1)
         stop(paste(
             "all elements of cov must be of the same length and \n",
             "cov.times must have one more entry than each element of 
cov."))
     var.names <- names(data)
     subjects <- rownames(data)
     omit.cols <- if (!common.times) c(all.cov, cov.times) else all.cov
     keep.cols <- (1:ncol)[-omit.cols]
     nkeep <- length(keep.cols)
     if (is.numeric(event)) event <- var.names[event]
     times <- if (common.times) matrix(cov.times, nrow, ncov+1, byrow=T)
         else data[, cov.times]
     new.data <- matrix(Inf, nobs, 3 + ncovs + nkeep)
     rownames <- rep("", nobs)
     colnames(new.data) <- c('start', 'stop', paste(event, suffix, sep=""),
         var.names[-omit.cols], cov.names)
     end.row <- 0
     for (i in 1:nrow){
         start.row <- end.row + 1
         end.row <- end.row + ncov
         start <- times[i, 1:ncov]
         stop <- times[i, 2:(ncov+1)]
         event.time <- ifelse (stop == data[i, time] & data[i, event] == 
1, 1, 0)
         keep <- matrix(unlist(data[i, -omit.cols]), ncov, nkeep, byrow=T)
         select <- apply(matrix(!is.na(data[i, all.cov]), ncol=ncovs), 
1, all)
         rows <- start.row:end.row
         cov.mat <- xlag(matrix(unlist(data[i, all.cov]), 
nrow=length(rows)), lag)
         new.data[rows[select], ] <-
             cbind(start, stop, event.time, keep, cov.mat)[select,]
         rownames[rows] <- paste(subjects[i], '.', seq(along=rows), sep="")
         }
     row.names(new.data) <- rownames
     as.data.frame(new.data[new.data[, 1] != Inf &
         apply(as.matrix(!is.na(new.data[, cov.names])), 1, all), ])
     }
#-----------------------------------------------------------------