Skip to contents

Create Subgroup Analysis Table

Source: R/efficacy_subgroup.R
create_subgroup_table.Rd

Generates a tabular summary of subgroup analyses for HTA dossiers and Word documents. Provides the same information as create_forest_plot() in table format showing treatment effects (HR or OR) by subgroup with confidence intervals and interaction p-values.

Usage

create_subgroup_table(
  data,
  subgroups,
  min_subgroup_size = 20,
  endpoint_type = c("tte", "binary"),
  time_var = "AVAL",
  event_var = "CNSR",
  response_var = "AVALC",
  response_values = c("CR", "PR"),
  trt_var = "TRT01P",
  ref_group = NULL,
  conf_level = 0.95,
  show_interaction = TRUE,
  adjust_method = c("none", "holm", "hochberg", "hommel", "bonferroni", "BH", "fdr",
    "BY"),
  title = "Subgroup Analysis",
  autofit = TRUE
)

Arguments

data

ADaMData object or data frame

subgroups

Named list mapping variable names to display labels, e.g., list(AGEGR1 = "Age Group", SEX = "Sex", RACE = "Race")

min_subgroup_size

Numeric. Minimum subgroup size required before warnings are issued. Use NULL to disable warnings. Default is 20.

endpoint_type

"tte" for time-to-event (HR) or "binary" for binary outcomes (OR)

time_var

Time variable for TTE endpoints (default: "AVAL")

event_var

Event variable for TTE endpoints. If "CNSR", will be inverted automatically. Default: "CNSR"

response_var

Response variable for binary endpoints (default: "AVALC")

response_values

Values indicating response for binary endpoints (default: c("CR", "PR"))

trt_var

Treatment variable name (default: "TRT01P")

ref_group

Reference treatment group. If NULL, uses first level.

conf_level

Confidence level (default: 0.95)

show_interaction

Logical, calculate and show interaction p-values (default: TRUE)

adjust_method

Character. Method for adjusting interaction p-values for multiple comparisons. One of:

  • "none" (default): No adjustment

  • "holm": Holm-Bonferroni step-down (recommended for FWER control)

  • "hochberg": Hochberg step-up (controls FWER)

  • "hommel": Hommel's method (closed testing procedure, good for correlated tests)

  • "bonferroni": Bonferroni correction (conservative)

  • "BH" or "fdr": Benjamini-Hochberg (controls FDR)

  • "BY": Benjamini-Yekutieli (controls FDR under dependency) See adjust_pvalues for details.

title

Table title

autofit

Logical, whether to autofit column widths (default: TRUE)

Value

A ClinicalTable object

Examples

if (FALSE) { # \dontrun{
# TTE subgroup table
sg_table <- create_subgroup_table(
  data = adtte,
  subgroups = list(
    AGEGR1 = "Age Group",
    SEX = "Sex"
  ),
  endpoint_type = "tte",
  title = "Subgroup Analysis - Overall Survival"
)

# Binary endpoint subgroup table
sg_table <- create_subgroup_table(
  data = adrs,
  subgroups = list(SEX = "Sex"),
  endpoint_type = "binary",
  response_values = c("CR", "PR"),
  title = "Response Rate by Subgroup"
)

# With multiplicity adjustment for GBA dossiers
sg_table <- create_subgroup_table(
  data = adtte,
  subgroups = list(
    AGEGR1 = "Age Group",
    SEX = "Sex",
    RACE = "Race"
  ),
  endpoint_type = "tte",
  adjust_method = "holm",
  title = "Subgroup Analysis with Multiplicity Adjustment"
)
} # }