Start a shift table from the data — first_shift

The third kind of table the app's "first table" form starts: subjects by their baseline category (columns, under each group) and their worst category after baseline (rows) – a laboratory normal range (BNRIND / ANRIND) or toxicity grade (BTOXGR / ATOXGR) – one page a parameter, the percents of each column's subjects. It writes what such a table made by hand has (tflspec's LB example):

Usage

first_shift_table(
  x,
  output_id,
  path,
  data,
  subjects_path,
  subjects,
  population,
  group,
  params,
  baseline = "BNRIND",
  post = "ANRIND",
  worst = NULL,
  where = NA_character_,
  total = NA_character_,
  description = NA_character_
)

Arguments

x: A tflplanner.
output_id: The report.
path: The records' data file (data/adam/adlb.rds).
data: Its data.
subjects_path: The subject-level data file (data/adam/adsl.rds).
subjects: Its data.
population: The analysis set flag column of the subject-level data (SAFFL).
group: The column of the groups (TRT01A).
params: The parameters (PARAMCD values), one page each.
baseline, post: The baseline and the post-baseline category columns (BNRIND, ANRIND).
worst: The post-baseline categories from best to worst; NULL: normal ranges NORMAL, LOW, HIGH, else the categories' order.
where: The records counted (ONTRTFL == "Y"); NA all of them.
total: The heading of a column of each group's subjects whatever their baseline ("Total"); NA none.
description: The report's description.

Value

The tflplanner.

Details

the analysis set and its subjects (adsl_saf), and the parameters' records counted, kept to them (adlb_saf);
each subject's worst record of a parameter (adlb_saf_worst, R of its own): the last of worst, the post-baseline categories from best to worst;
one categorical analysis by parameter, group and baseline; with total, over the records and the same again with every baseline as total (a column of all the group's subjects);
the categories' order (code lists), the table (one page a parameter), and the column header (each group over its baselines; no N, which differs from one parameter's page to the next).