gtregression • gtregression

R package for health research

From model to manuscript. Fit regression and survival models, and get tables, forest plots and survival curves ready for your paper, without hand-formatting a single coefficient.

Get started Which function?

install.packages("gtregression")

What you can make

Manuscript tables Descriptive, crude and adjusted results merged into one table, with every footnote kept.

Forest plots Crude and adjusted estimates side by side, straight from the fitted models.

Survival analysis Kaplan-Meier curves, Cox and parametric models, competing risks and RMST.

Absolute effects Risk differences and the number needed to treat, crude or adjusted.

New in 1.2.0

Risk difference and NNT risk_difference() reports absolute risks, the risk difference and NNTB/NNTH, crude or adjusted.

More study designs Matched case-control, multilevel, ordinal, multinomial and competing-risk regression tables.

Richer survival tools Proportional-hazards plots, delayed entry, clustering, weights and adjusted survival curves.

Better merging Merge any number of tables, keep each block’s notes, and choose a table theme.

Read the full list of changes.

From data to manuscript

Prepare

Check variables, labels, levels and missing data.

dissect() Analysis-ready data

Describe

Build a clear baseline table before modelling.

descriptive_table() Table 1

Model

Fit crude, adjusted, stratified or survival models.

uni_reg() multi_reg() Effect estimates

Interpret

Check assumptions, confounding, interaction and fit.

check_ph() compare_models() A defensible model

Publish

Merge, visualise and export.

merge_tables() save_table() Manuscript-ready output

A five-minute example

library(gtregression)
library(dplyr)

data("data_birthwt", package = "gtregression")

birthwt <- data_birthwt |>
  mutate(
    race = factor(race, levels = c(1, 2, 3), labels = c("White", "Black", "Other")),
    smoke = factor(smoke, levels = c(0, 1), labels = c("No", "Yes")),
    ht = factor(ht, levels = c(0, 1), labels = c("No", "Yes")),
    low = factor(low, levels = c(0, 1), labels = c("Normal BW", "Low BW"))
  )

attr(birthwt$smoke, "label") <- "Smoking during pregnancy"
exposures <- c("age", "lwt", "race", "smoke", "ht")

# Baseline characteristics by outcome
desc <- descriptive_table(birthwt, exposures = exposures, by = "low",
                          show_dichotomous = "all_levels")

# One model per exposure, then one adjusted model
crude <- uni_reg(birthwt, outcome = "low", exposures = exposures, approach = "logit")
adjusted <- multi_reg(birthwt, outcome = "low", exposures = exposures, approach = "logit")

# One table for the paper, and a forest plot
table_2 <- merge_tables(desc, crude, adjusted,
                        spanners = c("Low birth weight", "Crude", "Adjusted"))
forest <- forest_reg(forest_df(crude, adjusted))

# Save
save_table(table_2, filename = "table-2", format = "docx")
save_forest(forest, filename = "forest-plot", format = "png")

Variable labels set with attr(x, "label") or labelled::var_label() appear in every table and plot. Each result keeps its fitted models (adjusted$models) and table data (adjusted$table_body), so nothing is hidden.

Find your way

Describe dataBaseline tables with row or column percentages.

Regression tablesLogistic, risk ratio, count, linear and Firth models.

Special designsMatched, multilevel, ordinal, multinomial and competing risks.

Survival analysisKaplan-Meier, Cox, parametric models and predictions.

Risk difference and NNTAbsolute effects for two-group comparisons.

Stratified analysisThe same model within each subgroup.

Confounding and interactionScreen confounders and test effect modification.

Causal mediationDirect and indirect effects with clear caveats.

DiagnosticsConvergence, collinearity, fit and model comparison.

VisualiseRegression plots, forest plots and survival curves.

Merge and customiseCombine tables, apply themes, style with flextable.

ExportWord, HTML, PDF and image output.

Built on trusted R packages

gtregression is a readable interface over established packages: stats, survival, MASS, risks and logistf fit the models; broom, sandwich and lmtest handle inference; flextable, officer and gt build the tables; ggplot2, patchwork and forestploter draw the figures. The fitted models stay inside every result, so each number can be traced back to its source.

Install

# CRAN
install.packages("gtregression")

# Development version
remotes::install_github("ThinkDenominator/gtregression")

Citation

If you use gtregression in your work, please cite it as:

Polani R, Eliyas SK, Sakthivel M, Kaviprawin M, Krishnamoorthy Y, Majella MG. gtregression: Tools for Creating Publication-Ready Regression Tables. Zenodo. https://doi.org/10.5281/zenodo.16905349