In this post, we would like to review the idea of meta-analysis and compare a traditional, frequentist style, random effects meta-analysis to Bayesian

Examining Meta-Analysis

submited by
Style Pass
2024-11-27 16:00:03

In this post, we would like to review the idea of meta-analysis and compare a traditional, frequentist style, random effects meta-analysis to Bayesian methods. We will do this using the meta R package and a Bayesian analysis conducted with R but actually carried out by the Stan programming language on the back end. We will use a small but interesting textbook data set of summary data from eight randomized controlled trials. These studies examined the effectiveness of the calcium channel blocker Amlodipine compared to a placebo in improving work capacity in patients with angina. The data set and the analysis with the meta package come directly from the textbook by Chen and Peace (2013) . Although there are several R packages that are capable of doing the Bayesian meta-analysis, we chose to work in rstan to demonstrate its flexibility and hint how one might go about doing a large complex study that doesn’t quite fit with pre-programmed paradigms.

To follow our examples, you really don’t need to know much more about meta-analysis than the definition offered by Wikipedia contributors (2024) : “meta-analysis is a method of synthesis of quantitative data from multiple independent studies addressing a common research question”. If you want to dig deeper, a good overview of the goals and terminology can be found in Israel and Richter (2011) . The online textbook Doing Meta-Analysis in R: A Hands-on Guide by Mathias Harrier and Ebert (2021) will take you a long way in performing your own work.

Leave a Comment