icut {Epi} | R Documentation |
The follow up time from enter
to exit
is classified as
to wheter it is before cut
(Time
=0) or after
Time
=1). If cut
is between enter
and exit
,
the follow-up is split in two intervals, the first gets the value
cens
for the status.
icut( enter, exit, cut, fail = 0, cens.value = 0, data = data.frame(enter, exit, fail, cut), Expand = 1:nrow( data ), na.cut = Inf )
enter |
Date of entry. Numerical vector. |
exit |
Date of exit. Numerical vector. |
fail |
Indicator if exit status. |
cens.value |
Value to use for censoring status. |
cut |
Date where to cut follow-up. Numerical vector. |
data |
Dataframe of variables to carry over to the output dataframe. |
Expand |
Variable identifying original records. |
na.cut |
What value should be assigned to missing values of the
cutpoint. Defaults to Inf , so the inetrmediate event is
considered not to have occcurred. If set to -Inf , all persons
with missing cut are considered to have had an intermediate
event. If set to NA records with missing cut are
omitted from the result. |
The purpose of this function is to divide follow-up into pre- and post
some intermediate event like recurrence of disease, thus enabling
Follow-up for persons with a recurrence date (cut
) will be
split in two, with indication (in Time
) of what is pre and what
is post recurrence. This is typically what precedes a survival
analysis where recurrence is modelled as a time-dependent variable.
A data frame with one row per interval of follow up and columns given
in the data
argument, preceded by the columns:
Expand |
Identification of the rows from the input dataframe. |
Enter |
Entry date for the interval. |
Exit |
Exit date for the interval. |
Fail |
Failure indicator for the interval. |
Time |
Indicator variable for intervals after cut . |
Bendix Carstensen, Steno Diabetes Center, bxc@steno.dk, www.biostat.ku.dk/~bxc
one <- round( runif( 15, 0, 15 ), 1 ) two <- round( runif( 15, 0, 15 ), 1 ) doe <- pmin( one, two ) dox <- pmax( one, two ) # Goofy data rows to test possibly odd behaviour doe[1:3] <- dox[1:3] <- 8 dox[2] <- 6 dox[3] <- 7.5 # Some failure indicators fail <- sample( 0:1, 15, replace=TRUE, prob=c(0.7,0.3) ) # So what have we got data.frame( doe, dox, fail ) # Cut follow-up at 5 icut( doe, dox, fail, cut=5 )