Skip to contents

Subgroup forest plot

Source: R/builders-figure.R
fig_forest.Rd

Treatment effect, comparator versus reference, overall and within each level of subgroup_var. With time-to-event data (CNSR present) the effect is a Cox hazard ratio on a log scale, and the survival package is required. Otherwise it is a difference in means (Welch t-test) at the last visit. Subgroups that cannot be estimated are dropped and listed in the tlf_dropped attribute.

Usage

fig_forest(
  df,
  ctx = fig_ctx(),
  subgroup_var = NULL,
  ref_arm = NULL,
  comp_arm = NULL,
  ci = 0.95,
  point_labels = TRUE,
  base_size = NULL,
  config = NULL
)

Arguments

df

Analysis data (BDS).

ctx

A fig_ctx() list.

subgroup_var

Subgroup column. NULL uses the first of SEX, AGEGR1, RACE, REGION present.

ref_arm, comp_arm

Reference and comparator arms (default: first and last arm in sort order).

ci

Confidence level.

point_labels

Print the estimate and CI next to each row?

base_size

Font size; NULL uses the configuration.

config

A tlf_config(); NULL resolves through tlf_get_config().

Value

A ggplot.

See also

Other figure builders: fig_boxplot(), fig_edish(), fig_km(), fig_lsmeans(), fig_meanci(), fig_shift(), fig_spaghetti(), fig_waterfall()

Examples

p <- fig_forest(ex_adlb[ex_adlb$PARAMCD == "ALT", ], subgroup_var = "SEX")
class(p)
#> [1] "ggplot2::ggplot" "ggplot"          "ggplot2::gg"     "S7_object"      
#> [5] "gg"