Censored least absolute deviation (CLAD) regression using CRQ in library QUANTREG

I would like to perform a CLAD regression with

y = EQ-5D-5L utility scores (bounded from above by 1.0)
x = various patient characteristics

I already found out that I need to use CRQ in the library QUANTREG, but I couldn't figure out the specifics so far. My questions are:

Do I need to use the Powell method?
If so, how do I specify "yc" (censoring times) if I do not have a time-variable but a 0/1 censoring variable?

This is the code I tried but I keep getting the notification "Event times can not exceed ctimes for right censoring", because for patients with a utility score >0 and <1 the score is higher than the 0/1 yc variable which I created.

daten <- read.table ("P:/XXX.csv", header=TRUE, sep=";")

attach(daten)

x=cbind(factor(qlq) , AGE , SEX)

daten$c <- 1

daten$d <- ifelse (daten$UTILITY<1,0,1)

yc <- daten$d

y <- daten$UTILITY

clad <- crq (Curv(UTILITY, d, "right") ~ x, tau=0.5, method="Powell", data=daten)

Thank you in advance!

Solution

If anyone ever comes across the same obstacle: for every y, yc needs to be the value at which y is censored, not a 0/1 censoring indicator.

In my case (y=utility score of the EQ-5D-5L) yc needs to be 1.

The following command does the trick: daten$d <- rep(1.000,377) (because I have 377 observations)