Bayesian Meta-Analysis in Stata

In systematic literature reviews and meta-analysis, researchers frequently encounter difficulties in the evidence base that prevent the usual estimators from being used. For example, there may be reason to believe that results reported by some studies are biased. Other studies might not have reported the statistics required for traditional meta-analysis. Bayesian models for meta-analysis can accommodate these and other problems in a flexible modelling methodology. They also produce results in a probabilistic form that has been shown to support decision makers.

In this course, participants will be introduced to Bayesian methods using the sampling algorithms in Stata’s flexible bayesmh command, no previous experience of Bayes is therefore required. We will show how to re-conceptualise meta-analysis from a weighted average estimator to a probabilistic model, before teaching the code needed for basic common effect and random effect meta-analyses. We will then progress to more sophisticated models for network meta-analysis (multiple intervention choices) and unreported statistics.

At the end of the course, participants will be able to autonomously implement (with the help of the Stata routine templates specifically developed for the course) the appropriate methods, given both the nature of their data and the analysis in hand, within their own research.

The course is of particular interest to researchers and professional working in Biostatistics, Business Administration, Economics, Education, Management, Marketing, Psychology, Public Health and Social Sciences. Due to its introductory nature, it is however, also accessible to individuals, regardless of their respective disciplines or fields, who need to acquire the requisite toolset to apply Bayesian Meta-Analysis to their own data. During the course, theoretical concepts are reinforced by applied examples, in which the course tutor discusses current research issues, highlighting potential pitfalls and the advantages of individual techniques.

A working knowledge of the basic principles of biostatistics and epidemiology, as well as a basic knowledge of the statistical software Stata. Knowledge of the arguments treated in our Meta-Analysis course will significantly facilitate participation in this course.

SESSION I

1. A primer on Bayesian methods and the bayesmh command in Stata
2. Meta-analysis as a model using likelihood (and prior)
3. Basic models using Stata

SESSION II

1. Arm-based and network meta-analysis
2. Dealing with biases and unreported statistics

USEFUL REFERENCES

Grant, R. & Di Tanna G. (2025) Bayesian Meta-Analysis: a practical introduction. Chapman and Hall/CRC.

We are currently adding the finishing touches to our 2026 training calendar. We therefore ask you to check our website regularly or contact us at training@tstat.eu should the dates for the course you are interested in not be published yet. You will then be contacted via email as soon as the dates are available.