Repository files navigation

Themultinma package implements network meta-analysis, networkmeta-regression, and multilevel network meta-regression models whichcombine evidence from a network of studies and treatments using eitheraggregate data or individual patient data from each study (Phillippo etal. 2020; Phillippo 2019). Models are estimated in a Bayesian frameworkusing Stan (Carpenter et al. 2017).

You can install the released version ofmultinma fromCRAN with:

install.packages("multinma")

The development version can be installed fromR-universe with:

install.packages("multinma",repos= c("https://dmphillippo.r-universe.dev", getOption("repos")))

or from source onGitHubwith:

# install.packages("devtools")devtools::install_github("dmphillippo/multinma")

Installing from source requires that therstan package is installedand configured. See the installation guidehere.

A good place to start is with the package vignettes which walk throughexample analyses, seevignette("vignette_overview") for an overview.The series of NICE Technical Support Documents on evidence synthesisgives a detailed introduction to network meta-analysis:

Dias, S. et al. (2011). “NICE DSU Technical Support Documents 1-7:Evidence Synthesis for Decision Making.”National Institute forHealth and Care Excellence. Available fromhttps://www.sheffield.ac.uk/nice-dsu/tsds.

Multilevel network meta-regression is set out in the following methodspapers:

Phillippo, D. M. et al. (2020). “Multilevel Network Meta-Regressionfor population-adjusted treatment comparisons.”Journal of the RoyalStatistical Society: Series A (Statistics in Society),183(3):1189-1210. doi:10.1111/rssa.12579.

Phillippo, D. M. et al. (2024). “Multilevel network meta-regressionfor general likelihoods: synthesis of individual and aggregate datawith applications to survival analysis”.arXiv:2401.12640.

Themultinma package can be cited as follows:

Phillippo, D. M. (2025).multinma: Bayesian Network Meta-Analysis ofIndividual and Aggregate Data. R package version 0.8.1.9000, doi:10.5281/zenodo.3904454.

When fitting ML-NMR models, please cite the methods paper:

Phillippo, D. M. et al. (2020). “Multilevel Network Meta-Regressionfor population-adjusted treatment comparisons.”Journal of the RoyalStatistical Society: Series A (Statistics in Society),183(3):1189-1210. doi:10.1111/rssa.12579.

For ML-NMR models with time-to-event outcomes, please cite:

Phillippo, D. M. et al. (2024). “Multilevel network meta-regressionfor general likelihoods: synthesis of individual and aggregate datawith applications to survival analysis”.arXiv:2401.12640.