Skip to contents

The analyses of a forest plot: the figure's own ARD

Source: R/fig_model.R
tfl_fig_forest_analyses.Rd

A forest plot prints numbers – the hazard ratio, its interval, N – so they are statistics of an ARD, computed by cards / cardx and traceable in ARS, not fitted inside the figure program. This gives the rows of an ARD definition the plot needs: one analysis data (the parameter's records of the population, the subgroup variables added from the population's data) and one custom analysis per estimate – a Cox model of all subjects (cardx::ard_regression()), and one within each subgroup variable (cards::ard_strata()). tfl_fig_template("forest_hr") returns them as the design's attribute analyses; a study manager writes them to the figure's ARD definition.

Usage

tfl_fig_forest_analyses(
  data = "ADTTE",
  param = "OS",
  pop = "FASFL",
  group = "TRT01P",
  subgroups = c("SEX", "AGEGR1"),
  time = "AVAL",
  censor = "CNSR",
  id = "HR"
)

Arguments

data

The time-to-event dataset (ADTTE).

param

Its parameter (PARAMCD).

pop

The population (analysis set), by its id or its flag (FASFL). The analysis data is named <data>_<population>_<param> (adtte_fas_os), the population without its flag's FL: the rule tflplanner names a data by.

group

The treatment variable; its first level is the reference.

subgroups

The subgroup variables (of the population's data), a character vector or one string with , between them.

time, censor

The time and the censoring indicator (1 = censored).

id

The analysis id of the overall estimate; a subgroup's is <id>_<variable>.

Value

A list of two data frames, analysis_data and analyses, with the columns of those sheets they fill (no output_id).

Examples

tfl_fig_forest_analyses("ADTTE", "OS", "FASFL", "TRT01P", c("SEX", "AGEGR1"))$analyses
#>   analysis_id                      label method         data
#> 1          HR Hazard ratio, all subjects custom adtte_fas_os
#> 2      HR_SEX        Hazard ratio by SEX custom adtte_fas_os
#> 3   HR_AGEGR1     Hazard ratio by AGEGR1 custom adtte_fas_os
#>                                                                                                                                                                                            code
#> 1                                                                cox <- survival::coxph(survival::Surv(AVAL, CNSR == 0) ~ TRT01P, data = data)\ncardx::ard_regression(cox, exponentiate = TRUE)
#> 2    cards::ard_strata(\n  data,\n  .strata = SEX,\n  .f = ~ cardx::ard_regression(\n    survival::coxph(survival::Surv(AVAL, CNSR == 0) ~ TRT01P, data = .x),\n    exponentiate = TRUE\n  )\n)
#> 3 cards::ard_strata(\n  data,\n  .strata = AGEGR1,\n  .f = ~ cardx::ard_regression(\n    survival::coxph(survival::Surv(AVAL, CNSR == 0) ~ TRT01P, data = .x),\n    exponentiate = TRUE\n  )\n)