Meta-analyses of data arising from systematic reviews are increasingly used for evidence-based clinical and public health practice. Health-care professionals need to understand and critique this research design.

This course will present a detailed description of the meta-analysis process, discuss the strengths, potential bias and limitations of this design, and provide step-by-step guidance on how to actually perform and report a meta-analysis.

We will focus on issues relevant to meta-analyses of observational studies although the overall methodology is highly applicable to meta-analyses of randomized trials as well. In particular, there will be discussion about issues such as adjustment for confounders, aggregating data from different observational designs, assessing small-study effects.

By the end of the course, students will be able to:

• Understand the specific methodological issues raised by meta-analyses of observational studies
• Understand the principles of fixed-effect and random-effects models
• Examine sources of between-trial heterogeneity
• Assess small-study effects
• Incorporate risk of bias assessments into synthesis

All participants should bring a laptop, with Stata software installed. Basic knowledge of Stata is desirable.

• Hartung J, Knapp G, Sinha BK. Statistical Meta-Analysis with Applications. Wiley. 2008
• Schwarzer G, Carpenter JR, Ru¨cker G. Meta-Analysis with R. Springer 2015

Anna Chaimani, PhD

Anna Chaimani is a tenured Researcher at the Inserm METHODS Team – Sorbonne Paris Cité Epidemiology and Statistics Research Centre (UMR 1153), University of Paris. She is also a co-convenor of the Cochrane Statistical Methods group and of the Cochrane Comparing Multiple Interventions Methods group. She obtained her PhD in Epidemiology and Biostatistics in 2014 from the University of Ioannina in Greece and in 2016 she obtained a Chaire d'Excellence position from Université Sorbonne Paris Cité. Her research interests lie in the field of evidence synthesis with a particular focus on methodology about pairwise meta-analysis and network meta-analysis. She has worked on models for small-study effects in network meta-analysis, network meta-epidemiological models and methods for missing data. She also works on software development for meta-analysis and she is the author of the network graphs package in Stata.