Forest Plots for Bayesian Meta Analyses in R • bayesfoRest

Overview

bayesfoRest is an R package for creating publication-ready forest plots from Bayesian meta-analyses. It provides a comprehensive solution for visualizing meta-analytic results with support for various effect measures, subgroup analyses, and risk of bias assessments.

Key Features

Bayesian Forest Plots: Create forest plots from brms models with customizable aesthetics
Multiple Effect Measures: Support for OR, HR, RR, IRR, MD, and SMD
Subgroup Analysis: Built-in support for subgroup meta-analyses
Risk of Bias Integration: Seamlessly incorporate risk of bias assessments (RoB2, ROBINS-I, ROBINS-E, QUADAS-2)
Shrinkage Visualisation: Display shrinkage effects with density plots or point intervals
ROPE Support: Add Region of Practical Equivalence (ROPE) to plots
Extensive Customisation: Colours, fonts, labels, and layout

Installation

You can install the development version of bayesfoRest from GitHub:

# install.packages("devtools")
devtools::install_github("BLMoran/bayesfoRest")

Basic Usage

Binary Outcome Forest Plot (Odds Ratio)

bayes_forest(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  xlim = c(0.1, 3.5)
)

Add Subgroups

bayes_forest(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  subgroup = T,
  xlim = c(0.1, 3.5)
)

Add Risk of Bias Columns

bayes_forest(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  subgroup = T,
  add_rob = T,
  xlim = c(0.1, 3.5)
)

Add Risk of Bias Legend

bayes_forest(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  subgroup = T,
  add_rob = T,
  add_rob_legend = T,
  rob_tool = "rob2",
  xlim = c(0.1, 3.5)
)

Exclude Studies with High Risk of Bias

bayes_forest(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  subgroup = T,
  add_rob = T,
  add_rob_legend = T,
  rob_tool = "rob2",
  exclude_high_rob = T,
  xlim = c(0.1, 3.5)
)

Standalone Risk of Bias Plot

# Create standalone RoB visualization
rob_plot(
  data = binary_outcome,
  rob_tool = "rob2",
  add_rob_legend = TRUE
)

Overall Plot

# Create overall plot
overall_plot(
  data = binary_outcome,
  model = model_bin,
  measure = "OR",
  study_var = "Author",
  incl_tau = T,
  incl_mu_prior = T,
  incl_tau_prior = T,
  mu_xlim = c(0.1, 3.5),
  tau_xlim = c(0, 1),
  plot_arrangement = "vertical"
)

Sensitivity Plot

# Create sensitivity plot with different priors and meta-analysis models
sensitivity_plot(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  study_var = Author,
  rob_var = Overall,
  exclude_high_rob = T,
  incl_pet_peese = T,
  incl_mixture = T,
  incl_bma = T,
  model_bma = robma_fits_bin,
  add_probs = F,
  priors = list(
    vague = prior(normal(0, 10), class = "Intercept"),
    weakreg  = prior(normal(0, 1),  class = "Intercept"),
    informative = prior(normal(0, 0.5), class = "Intercept")),
  xlim = c(0.25, 1.5),
  add_null_range = T
)

Sensitivity Plot with Probabilities

# Create sensitivity plot with different priors and meta-analysis models with probabilities
sensitivity_plot(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  study_var = Author,
  rob_var = Overall,
  exclude_high_rob = T,
  incl_pet_peese = T,
  incl_mixture = T,
  incl_bma = T,
  model_bma = robma_fits_bin,
  add_probs = T,
  priors = list(
    vague = prior(normal(0, 10), class = "Intercept"),
    weakreg  = prior(normal(0, 1),  class = "Intercept"),
    informative = prior(normal(0, 0.5), class = "Intercept")),
  xlim = c(0.25, 1.5),
  add_null_range = T
)

ECDF Plot

# Create ECDF plot with different meta-analysis models
ecdf_plot(
  model = model_bin,
  data = binary_outcome,
  measure = "OR",
  study_var = Author,
  rob_var = Overall,
  priors = list(
    weakreg  = prior(normal(0, 1),  class = "Intercept")),
  exclude_high_rob = T,
  incl_pet_peese = T,
  incl_mixture = T,
  incl_bma = T,
  model_bma = robma_fits_bin,
  xlim = c(0.25, 1.5),
  add_null_range = T,
  prob_reference = "null"
)

Citation

If you use bayesfoRest in your research, please cite:

@Manual{,
  title = {bayesfoRest: Forest Plots for Bayesian Meta-Analyses in R},
  author = {[Benjamin Moran and Thomas Payne]},
  year = {2025},
  note = {R package version 0.0.0.9000},
  url = {https://github.com/BLMoran/bayesfoRest},
}

Dependencies

bayesfoRest depends on several R packages:

brms for Bayesian model fitting
metafor for meta-analysis preparation
gt for creating tables
patchwork for combining plots
ggplot2 for plotting
ggdist for density plotting
tidybayes for tidy workflow
dplyr, tidyr, purrr for data manipulation
paletteer for colour palettes
fontawesome for risk of bias icons

Feedback, Issues and Contributing

We welcome feedback, suggestions, issues and contributions. Please feel free to contact either Ben or Tom with any feedback. For any bugs, please file it here with a minimal code example to reproduce the issue. Pull requests can be made here. Please note that the bayesfoRest project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

License

This package is licensed under the GPL-3 License.

Acknowledgments

bayesfoRest builds upon the excellent work of the brms , metafor, gt, patchwork, ggdist, tidybayes and the tidyverse suite of packages. Without the work of the R community and their contributions, bayesfoRest would not be possible.

bayesfoRest: Forest Plots for Bayesian Meta Analyses in R