nmatools

R package providing one-stop wrappers for network meta-analysis (NMA) using the netmeta package: a scripted pipeline that fits a network and writes a complete set of results and publication-ready figures to disk, a suite of publication visualizations, and an interactive CINeMA + ROB-MEN Shiny GUI for rating confidence in the results.

Documentation

The full user manual lives in docs/manual/. Start at the manual home for reading paths, or jump to a chapter below.

Getting started

Chapter	Topic
1. Installation and quick start	Requirements, install, five-minute NMA
2. Data formats	Input formats, w2i_trials, helpers, project setup

Scripted pipeline

Chapter	Topic
3. The NMA pipeline	netmetawrap() arguments, outputs, overrides, subnetworks
4. Batch runs and rare events	run_nma_batch() and the Mantel-Haenszel rare-event workflow
5. Transitivity	plot_transitivity() covariate plots

GUI guide

Chapter	Topic
6. GUI overview	Launch modes and workflow
7. Configuration tab	Loading data, NMA and ROB-MEN settings, running
8. CINeMA domains	Domain tabs ①③④⑤⑥
9. ROB-MEN	Domain ② and the ROB-MEN assessment
10. Report and export	Summary grid, figures, ZIP bundle

Visualizations

Chapter	Topic
11. Colored league tables	color_league() / color_league_multi()
12. Colored forest and network graphs	color_forest() / color_netgraph() / palettes
13. Kilim and Vitruvian plots	kilim() / vitruvian()

Evidence frameworks

Chapter	Topic
14. Evidence frameworks	min_context() / part_context()

Reference

Chapter	Topic
15. Troubleshooting and FAQ	Common problems by symptom
16. Function reference	All exported functions and the literature reference list

Install

# Install from GitHub (recommended)
# install.packages("remotes")   # if not yet installed
remotes::install_github("ykfrkw/nmatools", force = TRUE)

Dependencies (netmeta, meta, dplyr, writexl, magick, and others) are installed automatically. Requires R ≥ 4.1 and a recent netmeta (≥ 3.x); see Chapter 1.

Quick start

A single netmetawrap() call fits the NMA for one outcome and writes every result — the fitted model, consistency tests, a league table, and a set of publication-ready plots — to ./outputs/{outcome}/.

library(nmatools)

d <- load_w2i()   # bundled W2I insomnia sample data (Furukawa et al. 2024)

netmetawrap(
  data            = d,
  studlab         = "id",
  treat           = "t",
  outcome         = "remission_lt",  # also names the output sub-directory
  n               = "n",
  event           = "r",
  sm              = "OR",
  reference.group = "Pharmacotherapy",
  small.values    = "undesirable"
)
# → results are written to outputs/remission_lt/

See Chapter 3 for the full output-file table and override hooks, and Chapter 4 to run many outcomes at once with run_nma_batch().

Interactive GUI (CINeMA + ROB-MEN)

cinema() launches a Shiny application for interactive confidence assessment, implementing CINeMA (Nikolakopoulou et al. 2020) and ROB-MEN (Chiocchia et al. 2021). Launch it empty or pre-load a data frame, then work left to right through the domain tabs to the Report and export bundle — see Chapters 6–10.

library(nmatools)

cinema()                                          # launch empty; upload in the GUI
cinema(load_w2i(), format = "binary", effect_measure = "OR")   # pre-load from R

Sample data

w2i_trials — arm-level data from the Furukawa et al. (2024) W2I NMA. Three treatments (CBT-I, Combination, Pharmacotherapy) for chronic insomnia; four binary outcomes (remission and dropout at long-term and post-treatment). The column dictionary is documented in Chapter 2.

d  <- load_w2i()            # arm-level trial data
ci <- w2i_cinema_path()     # path to bundled CINeMA CSV (remission_lt only)

Furukawa Y, Sakata M, Furukawa TA, Efthimiou O, Perlis M. Initial treatment choices for long-term remission of chronic insomnia disorder in adults: a systematic review and network meta-analysis. Psychiatry Clin Neurosci. 2024;78(11):646-653. https://doi.org/10.1111/pcn.13730

Citation

nmatools is a wrapper around netmeta and meta. If you use it in published research, please also cite those packages directly. The full literature reference list is in Chapter 16.

citation("netmeta")
citation("meta")