Convenience function to prepare an outcome object from adtte-like data sets, either for tte or binarized endpoint.
Usage
build_out_tte(
data = NULL,
file = NULL,
filter = NULL,
cut = NULL,
unit = "unit",
label = ".out",
columns = NULL
)Arguments
- data
adtte-like tibble, e.g. output of haven::read_sas('adtte.sas7bdat')
- file
path to ADaM-like tte data set, e.g. files/adtte.sas7bdat, ignored if data is provided
- filter
character vector to be applied to the data set (e.g. to select PARAMCD or subset to particular population)
- cut
optional numeric, required for binarized version from tte data. Note that cut must provided in the same unit as the time column (e.g. AVAL)
- unit
only relevant if outcome should be binarized, i.e. if
cutis provided.unitshould describe the unit of the numeric (time) column in data set ass well as ofcut, and will be included in outcome values- label
optional label for the binarized outcome variable, defaults to '.out'
- columns
list defining the mapping of required information to columns in the data with default entries. See details for defaults.
Value
A tibble with column .id and either an additional character column .out for the binarized version
or with the addition of the pair .time and .status for tte outcomes.
Observations with missing values in either .time or .status will be removed.
In this case, a notification with the number of removed observations will be printed to the console.
Note that sample sizes may differ for binarization and tte outcome, as subjects are dropped if the observed time is censored and below cut.
Details
Note that unit is solely a description to be included in the outcome value (e.g. "event in first 2 year(s)").
No conversion of the time is done to unit.
Column mapping defaults are chosen to match ADaM adtte data sets:
id'SUBJID'time'AVAL'censor'CNSR'
If status instead of censoring indicator should be used, specify e.g. status = 'event' (defaults to NULL).
If both are defined, status is used, censor is ignored.