DOCSEARCHbyAlgolia
Submit Docs

gtsummary R Package

repository·main·Indexed 22 days ago

https://github.com/ddsjoberg/gtsummary

An R package for creating publication-ready analytical and summary tables. It provides tools like tbl_summary() for descriptive statistics, tbl_regression() for summarizing regression models, and tbl_merge() for presenting multiple models side-by-side. Tables can be exported to formats including .png, .html, .docx, .rtf, .tex, and .ltx via gt.

Tokens
864
Snippets
5
Records
5
Agent score
29%

What's inside gtsummary

  2. Merge regression tables with tbl_merge()

    main

    You can present multiple regression models side-by-side using tbl_merge(). This function takes a list of gtsummary tables and allows you to add a tab_spanner to group them under a common header.

    library(survival)

# build survival model table
t2 <-
  coxph(Surv(ttdeath, death) ~ trt + grade + age, trial) |> 
  tbl_regression(exponentiate = TRUE)

# merge tables
tbl_merge_ex1 <-
  tbl_merge(
    tbls = list(t1, t2),
    tab_spanner = c("**Tumor Response**", "**Time to Death**")
  )

  3. Save gtsummary tables to files

    main

    To save a table to a specific file format, convert it to a gt object using as_gt() and then use gt::gtsave(). Supported extensions include .png, .html, .docx, .rtf, .tex, and .ltx.

    tbl |> 
  as_gt() |> 
  gt::gtsave(filename = ".") # use extensions .png, .html, .docx, .rtf, .tex, .ltx

  4. Create a summary table with tbl_summary()

    main

    Use tbl_summary() to create descriptive statistics tables from data frames or tibbles. It automatically detects variable types (continuous, categorical, dichotomous) and includes missingness information. You can use the include argument to select specific variables and the by argument to split the table by a grouping variable.

    library(gtsummary)

# Basic summary table
table1 <- trial |> 
  tbl_summary(include = c(age, grade, response))

# Customized summary table
table2 <-
  tbl_summary(
    trial,
    include = c(age, grade, response),
    by = trt, # split table by group
    missing = "no" # don't list missing data separately
  ) |> 
  add_n() |> # add column with total number of non-missing observations
  add_p() |> # test for a difference between groups
  modify_header(label = "**Variable**") |> # update the column header
  bold_labels()

  5. Summarize regression models with tbl_regression()

    main

    Use tbl_regression() to display regression model results. It automatically identifies common models (like logistic or Cox proportional hazards) and pre-fills appropriate headers (e.g., Odds Ratio or Hazard Ratio). Use exponentiate = TRUE to display results as exponentiated coefficients.

    mod1 <- glm(response ~ trt + age + grade, trial, family = binomial)

t1 <- tbl_regression(mod1, exponentiate = TRUE)