Summary of Concomitant Medications

# ------------------------------------------------------------------------------
# t-conmeds — Summary of Concomitant Medications (Number of Subjects)
# ------------------------------------------------------------------------------
# Target: Table 14-7.04 of the CDISCPILOT01 clinical study report (27 June 2006).
#
# WHAT COUNTS AS CONCOMITANT. Every ADCM record counts. ADCM carries PREFL,
# ONTRTFL and FUPFL, and a narrower reading — medications taken during treatment
# only — is available and would give a different table. The reference report
# counts every recorded medication (77, 74 and 78 subjects), and that is what this
# display reproduces. The choice is stated in the display's own footnotes because
# it is a definition, not a detail: restricting to ONTRTFL = "Y" drops roughly two
# thirds of the records.
#
# CMCLAS is the anatomical therapeutic class as coded on the source records, and
# CMDECOD the coded medication name. 6085 of 7510 records are coded "UNCODED" in
# both, and they are reported as such rather than hidden: 86%, 83% and 92% of
# subjects have at least one medication that was never coded, which is a property
# of this study's data that a reader should see.
#
# `group: [TRT01P]` — planned treatment, as in the rest of this display group and
# as in the reference report. Twelve subjects received a treatment other than the
# one planned.
# ------------------------------------------------------------------------------
id: t-conmeds
title: "Summary of Concomitant Medications (Number of Subjects)"
regulatory_id: CMT01
type: table
dataset: adcm
analysis_set: safety
denominator: adsl
group:
  - TRT01P
total: false
analyses:
  - name: any_conmed
    method: subject_count
    label: "Subjects receiving at least one concomitant medication"
    variable: CMFL
  - name: by_class_term
    method: hierarchical_count
    label: "Concomitant medications by therapeutic class and coded medication"
    hierarchy: [CMCLAS, CMDECOD]

id: t-conmeds
title: "Summary of Concomitant Medications (Number of Subjects)"
study: CDISCPILOT01
population_label: "Safety Analysis Set"
cutoff: "2014-07-01"
source: "Source: adcm, adsl (pharmaverseadam). Data cut-off: 2014-07-01."
footnotes:
  - "Treatment groups are PLANNED treatment (TRT01P). Twelve subjects in this study received a treatment other than the one they were randomised to; grouping by actual treatment (TRT01A) would give column sizes of 86, 96 and 72."
  - "A subject is counted once per therapeutic class and once per coded medication, however many records they have. Class counts are therefore not the sum of the medication counts beneath them."
  - "Every recorded medication counts, whether it was taken before, during or after treatment. Restricting to medications taken on treatment (ONTRTFL = 'Y') would count a much smaller set of records."
  - "UNCODED is not a therapeutic class. It is how this study's data records a medication that was never coded to a dictionary term, and it is reported rather than dropped."
  - "Classes, and the medications within a class, are ordered by the largest subject count in any treatment group, ties alphabetically. The reference document this display targets ordered them alphabetically by class instead."
  - "Percentages are based on the number of subjects in the safety analysis set for each treatment group."
columns:
  order: [Placebo, "Xanomeline Low Dose", "Xanomeline High Dose"]
format:
  digits:
    # `N:` and `n:` are deliberately absent. R's YAML reader resolves BOTH bare
    # `N` and bare `n` to the boolean FALSE, so declaring the two together is a
    # duplicate-key parse error and declaring either alone silently keys the digit
    # plan under "FALSE". Counts default to zero decimals, which is what is wanted.
    p: 1
rows:
  - { analysis: any_conmed, variable: CMFL, level: "Y", pattern: n_pct,
      label: "Subjects receiving at least one concomitant medication" }
  - { label: "Therapeutic class and coded medication, n (%)", section: true }
  - { analysis: by_class_term, type: hierarchical, levels: [CMCLAS, CMDECOD],
      pattern: n_pct }
variants:
  post_text: {}
  in_text:
    title: "Concomitant Medications Taken by at Least 5% of Subjects in Any Treatment Group"
    filter:
      min_pct: 5
    footnotes:
      - "Reduced rendering of the same analysis results dataset as the post-text display: a coded medication is shown when at least 5% of subjects in at least one treatment group took it, and a therapeutic class is shown when at least one of its medications is."