gtregression
.
gtregression is an R package that simplifies regression
modeling and generates publication-ready tables using the
gtsummary ecosystem. It supports a variety of regression
approaches with built-in tools for model diagnostics, selection, and
confounder identification—all designed to provide beginner and
intermediate R users with clean, interpretable output.
This package was created with the aim of empowering R users in low-
and middle-income countries (LMICs) by offering a simpler and more
accessible coding experience. We sincerely thank the authors and
contributors of foundational R packages such as gtsummary,
MASS, RISKS, dplyr, and
others—without whom this project would not have been possible.
Vision
At its core, gtregression is more than just a
statistical tool—it is a commitment to open access, simplicity, and
inclusivity in health data science. Our team is driven by the vision of
empowering researchers, students, and public health professionals in
LMICs through user-friendly, well-documented tools that minimize coding
burden and maximize interpretability.
We believe in the democratization of data science and aim to promote open-source resources for impactful and equitable research globally.
Features
- Supports multiple regression approaches:
- Logistic (logit)
- Log-binomial
- Poisson / Robust Poisson
- Negative Binomial
- Linear Regression
- Univariable and multivariable regression
- Confounder identification using crude and adjusted estimates
- Stepwise model selection (AIC/BIC/adjusted R²)
- Stratified regression support
- Formatted outputs using
gtsummary - Built-in example datasets:
PimaIndiansDiabetes2,birthwt,epil
Installation
# Install from CRAN
install.packages("gtregression")
# Or install the development version from GitHub
devtools::install_github("ThinkDenominator/gtregression")Quick Start
# Load necessary libraries
library(gtregression)
# Load example dataset
data("data_PimaIndiansDiabetes", package="gtregression")
# Convert diabetes outcome to binary and create categorical variables
pima_data <- data_PimaIndiansDiabetes |>
mutate(diabetes = ifelse(diabetes == "pos", 1, 0)) |>
mutate(bmi = case_when(
mass < 25 ~ "Normal",
mass >= 25 & mass < 30 ~ "Overweight",
mass >= 30 ~ "Obese",
TRUE ~ NA_character_),
bmi = factor(bmi, levels = c("Normal", "Overweight", "Obese")),
age_cat = case_when(
age < 30 ~ "Young",
age >= 30 & age < 50 ~ "Middle-aged",
age >= 50 ~ "Older"),
age_cat = factor(age_cat, levels = c("Young", "Middle-aged", "Older")),
npreg_cat = ifelse(pregnant > 2, "High parity", "Low parity"),
npreg_cat = factor(npreg_cat, levels = c("Low parity", "High parity")),
glucose_cat= case_when(glucose<=140~ "Normal", glucose>140~"High"),
glucose_cat= factor(glucose_cat, levels = c("Normal", "High")),
bp_cat = case_when(
pressure < 80 ~ "Normal",
pressure >= 80 ~ "High"
),
bp_cat= factor(bp_cat, levels = c("Normal", "High")),
triceps_cat = case_when(
triceps < 23 ~ "Normal",
triceps >= 23 ~ "High"
),
triceps_cat= factor(triceps_cat, levels = c("Normal", "High")),
insulin_cat = case_when(
insulin < 30 ~ "Low",
insulin >= 30 & insulin < 150 ~ "Normal",
insulin >= 150 ~ "High"
),
insulin_cat = factor(insulin_cat, levels = c("Low", "Normal", "High"))
) |>
mutate(
dpf_cat = case_when(
pedigree <= 0.2 ~ "Low Genetic Risk",
pedigree > 0.2 & pedigree <= 0.5 ~ "Moderate Genetic Risk",
pedigree > 0.5 ~ "High Genetic Risk"
)
) |>
mutate(dpf_cat = factor(dpf_cat,
levels = c("Low Genetic Risk",
"Moderate Genetic Risk",
"High Genetic Risk"))) |>
mutate(diabetes_cat= case_when(diabetes== 1~ "Diabetes positive",
TRUE~ "Diabetes negative")) |>
mutate(diabetes_cat= factor(diabetes_cat,
levels = c("Diabetes negative","Diabetes positive" )))
# Descriptive statistics table
exposures <- c("bmi", "age_cat", "npreg_cat", "bp_cat", "triceps_cat",
"insulin_cat", "dpf_cat")
# Create a descriptive table by diabetes category
des_tbl = descriptive_table(data= pima_data,
exposures = exposures,
by= "diabetes_cat")
# Check the data compatibility
dissect(pima_data)
# Univariable regression
uni_tbl = uni_reg(
data = pima_data,
outcome = "diabetes",
exposures = exposures,
approach = "logit"
)
# check models and summaries
uni_tbl$models
uni_tbl$model_summaries
# Plot univariable regression results
plot_reg(uni_tbl,
title = "Univariable Regression Results")
# multivariable regression
multi_tbl = multi_reg(
data = pima_data,
outcome = "diabetes",
exposures = exposures,
approach = "logit"
)
# check models and summaries
multi_tbl$models
multi_tbl$model_summaries
# Plot univariable regression results
plot_reg(multi_tbl,
title = "Multivariable Regression Results")
# combined plots
plot_reg_combine(
uni_tbl,
multi_tbl,
title = "Univariable vs Multivariable Regression Results")
# combine the tables
merge_table(des_tbl, uni_tbl, multi_tbl,
spanners = c("**Descriptive**",
"**Univariate**",
"**Multivariable**"))
# Save the table as a Word document
save_table(des_tbl, filename = "des_tbl", format = "docx")
save_docx(
tables = list(des_tbl, uni_tbl, multi_tbl),
filename = "Outputs.docx")
# Stratified regression
stratified_uni_reg(pima_data,
outcome= "diabetes",
exposures =c("bmi", "insulin_cat", "age_cat", "dpf_cat"),
approach = "logit",
stratifier = "glucose_cat")
stratified_multi_reg(pima_data,
outcome= "diabetes",
exposures =c("bmi", "insulin_cat", "age_cat", "dpf_cat"),
approach = "logit",
stratifier = "glucose_cat")
# Check model convergence
check_convergence(pima_data,
exposures = exposures,
outcome = "diabetes",
approach = "logit",
multivariate = F)
check_convergence(pima_data,
exposures = exposures,
outcome = "diabetes",
approach = "logit",
multivariate = T)
# identify confounders
identify_confounder(pima_data,
outcome = "diabetes",
exposure = "npreg_cat",
potential_confounder = "bp_cat",
approach = "logit")
# check interactions
interaction_models(pima_data,
outcome,
exposure = "bmi",
effect_modifier = "glucose_cat",
covariates = c("insulin_cat", "age_cat", "dpf_cat"),
approach = "logit")Key Functions
Descriptive & Compatibility Tools
| Function Name | Purpose |
|---|---|
descriptive_table() |
Summarise exposures by outcome groups |
dissect() |
Check outcome-exposure compatibility |
Regression Functions - Fit univariate and multivariable models
| Function Name | Purpose |
|---|---|
uni_reg() |
Univariable regression (OR/RR/IRR/β) |
multi_reg() |
Multivariable regression |
Regression Functions by stratifier
| Function Name | Purpose |
|---|---|
stratified_uni_reg() |
Stratified univariable regression |
stratified_multi_reg() |
Stratified multivariable regression |
Model Diagnostics & Selection
| Function Name | Purpose |
|---|---|
check_convergence() |
Evaluate model convergence and max fitted values |
select_models() |
Stepwise model selection (AIC/BIC/adjusted R²) |
Confounding & Interaction
| Function Name | Purpose |
|---|---|
identify_confounder() |
Confounding assessment via % change or MH method |
interaction_models() |
Compare models with and without interaction terms |
Plots & Exports
| Function Name | Purpose |
|---|---|
plot_reg() |
Forest plot for a single regression model |
plot_reg_combine() |
Side-by-side forest plots for uni/multi models |
modify_table() |
Customize column labels or output structure |
save_table() |
Export table to .html, .csv,
.docx
|
save_docx() |
Save table as Word document (.docx) |
save_plot() |
Save plot as .png, .pdf, etc. |
merge_tables() |
Combine descriptive and regression tables |
Conclusion
The gtregression package simplifies regression coding
and produces publication-ready tables with interpretation notes. It
enables beginners to explore a variety of regression models with ease,
transparency, and reproducibility. Explore the documentation for each
function to discover additional options and customization features.
Citation
If you use gtregression in your work, please cite it
as:
Rubeshkumar, P., Eliyas, S. K., Sakthivel, M., Krishnamoorthy, Y., & Majella, M. G. (2025). ThinkDenominator/gtregression: CRAN v1.0.0 (CRAN). Zenodo. https://doi.org/10.5281/zenodo.16905350
Acknowledgements
The gtregression package icon uses the “Hearts” symbol created by Kim Sun Young from The Noun Project, used under the Creative Commons Attribution (CC BY 3.0) license.