netmeta: Network Meta-Analysis using Frequentist Methods
Official Git repository of R package netmeta
Authors
Gerta Rücker, Ulrike Krahn, Jochem König, Orestis Efthimiou, Annabel Davies, Theodoros Papakonstantinou, Guido Schwarzer
Description
R package netmeta (Balduzzi et al., 2023) provides frequentist methods for network meta-analysis and supports Schwarzer et al. (2015), Chapter 8 "Network Meta-Analysis".
Available network meta-analysis models
-
frequentist network meta-analysis (Rücker, 2012; Rücker & Schwarzer, 2014);
-
additive network meta-analysis for combinations of treatments (Rücker, Petropoulou et al., 2020);
-
network meta-analysis of binary data using the Mantel-Haenszel method or the non-central hypergeometric distribution (Efthimiou et al., 2019).
Methods to present results of a network meta-analysis
-
network graphs (Rücker & Schwarzer, 2016);
-
forest plots;
-
league tables with network meta-analysis results;
-
tables with network, direct and indirect estimates looking similar to the statistical part of a GRADE table for a network meta-analysis (Puhan et al., 2014).
Methods to rank treatments
-
rankograms and ranking by the Surface Under the Cumulative RAnking curve (SUCRA) (Salanti et al., 2011);
-
ranking of treatments by P-scores (frequentist analogue of SUCRAs without resampling) (Rücker & Schwarzer, 2015);
-
partial order of treatment rankings ('poset') and Hasse diagram for 'poset' (Carlsen & Bruggemann, 2014; Rücker & Schwarzer, 2017).
Methods to evaluate network inconsistency
-
split direct and indirect evidence to check consistency (Dias et al., 2010);
-
net heat plot and design-based decomposition of Cochran's Q (Krahn et al., 2013).
Additional methods
-
contribution of direct comparisons to network estimates (Papakonstantinou et al., 2018; Davies et al., 2022)
-
importance of individual studies measured by reduction of precision if removed from network (Rücker, Nikolakopoulou et al., 2020)
-
'comparison-adjusted' funnel plot (Chaimani & Salanti, 2012);
-
measures characterizing the flow of evidence between two treatments (König et al., 2013).
Installation
release:
Current officialinstall.packages("netmeta")
Current development version on GitHub:
Installation using R package remotes:
install.packages("remotes")
remotes::install_github("guido-s/netmeta", ref = "develop")
Bug Reports:
You can report bugs on GitHub under Issues.
or using the R command
bug.report(package = "netmeta")
(which is not supported in RStudio).