Short Course 4

Network Meta-Analysis with R

Room: TBD (Sunday, 8 July 2017 from 9:00 am – 5:00 pm)


  • Gerta Rücker (University of Freiburg) Freiburg, Germany
  • Guido Schwarzer (University of Freiburg) Freiburg, Germany
  • James R. Carpenter (London School of Hygiene & Tropical Medicine) London, UK

Summary of Course

Meta-analysis is central to the increasing drive for evidence based decision making in evaluating therapies and developing health policies; appropriate statistical methods are key to this.

Network meta-analysis combines and ranks evidence from multiple treatments. This is a complex task; the analyst is likely to be faced with challenges such as (i) complex models; (ii) heterogeneity; (iii) possible inconsistency of direct and indirect evidence; (iv) small study effects; (iv) handling aggregate and IPD data. This is reflected in the rapid development of statistical methodology in this area. Alongside this, the increasing importance of network meta-analysis in decision making is reflected by the increasing number of publications of medical applications and their consideration by the GRADE working group (

This course provides a comprehensive introduction to network meta-analysis and its application using R. Interspersing lectures with computer practicals for the students, we begin by describing fixed and random effects network meta-analysis models alongside measures of heterogeneity and inconsistency. We then move to more advanced topics including ranking of treatments, splitting of direct and indirect evidence, and decomposition of heterogeneity due to treatment designs. All methods are applied to real data examples with different outcomes and data formats.


  1. Introduce key concepts in network meta-analysis, and associated statistical models.
  2. Discuss the main threats/difficulties with network meta-analysis in practice.
  3. Describe statistical methods for assessing and addressing these.
  4. Present and illustrate extensions that are the subject of on-going research. 

Learning Outcomes

At the end of the course, participants should:

  1. Have a clear understanding of the rationale and key statistical models in network meta-analysis.
  2. Be confident in using R to prepare datasets, fit and correctly interpret standard network meta-analysis models.
  3. Be able to produce and understand appropriate plots and tables using R.
  4. Understand the concepts and be able to carry out more advanced methods like evidence splitting.


The latter part of the course requires a Masters level education to benefit fully from the material.

Participants are expected to be familiar with the basic concepts of meta-analysis for pairwise comparisons. Furthermore, a basic knowledge of R is required to successfully participate in R practicals. Intending participants could usefully complete sessions 1-3 of our IBC2014 short course ‘Meta-analysis and its implementation in R’, which is free to IBS member at

About the Instructors

The instructors are co-authors of the book ‘Meta-Analysis with R’ in the Springer Use-R! series.

They presented a short course ‘Meta-analysis and its implementation in R’ at IBC2014 in Florence, Italy. This course was both very well attended and received very positive feedback, particularly the use of extensive computer practicals in R, which participants ran on their laptops. Because this was the most popular short course at IBC Florence with 30 participants, it was recorded and made available on the IBS website:

James R. Carpenter is a professor of medical statistics at the LSHTM and Programme Leader in Methodology at the MRC Clinical Trials Unit, London. He is a fellow of the UK Higher Education Academy, holding a diploma in teaching in higher education (with distinction) from the Institute of Education, University of London. He has extensive experience with Masters level teaching, having served as an external examiner at Masters level in the UK and as Chair of Examiners for the MSc Medical Statistics at the LSHTM. In addition he has presented and co-presented numerous courses, principally on missing data, internationally (including at IBC2012, ENAR 2013, RSS 2013).

Gerta Rücker is a senior statistician at the Institute for Medical Biometry and Statistics, University of Freiburg, Germany. She holds a diploma and a PhD in mathematics and has extensive experience in teaching, primarily for students and researchers in medicine and related fields. In a seminal paper, Gerta Rücker introduced methods from graph theory to frequentist network meta-analysis.

Guido Schwarzer is a senior statistician and head of IT at the Institute for Medical Biometry and Statistics, University of Freiburg, Germany. He is author of ‘R package meta’ and co-author of ‘R packages netmeta and metasens’. He is an established researcher in the area of meta-analysis and lead statistician of several Cochrane reviews. He has long-term experience in teaching medical students as well as researchers.

Gerta Rücker and Guido Schwarzer participated in several workshops on meta-analysis methods; both in German (workshops of Cochrane Germany in Freiburg, pre-conference workshop at DAGStat 2016) and in English (Master’s course Medical Biometry/Biostatistics at Heidelberg University 2010 – 2016, Cochrane Colloquium 2015). 


Schwarzer, G, Carpenter JR, Rücker G (2015). Meta-analysis with R. Springer International Publishing. Springer contact: Eva Hiripi (; Discount: 33.3%.