Bayesian meta-analysis to support decision making and policy
Get your free ticket here: https://buytickets.at/bayesianmixer/1862040
Abstract: Meta-analysis is the combination of information from studies that have been previously conducted. Often, we were not involved in those studies and so only have access to summary statistics. Also, there are many ways in which information can be lacking, we might suspect bias, and the studies might not be entirely compatible.
Bayesian methods can help here. Once we re-conceptualise meta-analysis as a statistical model, we can elaborate it in many different ways. I will compare several software options for this and give examples of applications found in researching my recent book on the subject with Professor Gian Luca Di Tanna.
Bio: Robert Grant is a statistician who has mostly worked on healthcare topics over the last 27 years. He is especially interested in Bayesian models, data visualisation and evidence synthesis.
His former roles were in hospital performance indicators, NHS guidelines, postgraduate teaching, epidemiology and health services research, and he has been a freelance trainer and consultant since 2017, with clients including the World Bank, Harvard Medical School and the Cabinet Office. He is one of the Stan developers, having written its Stata interface, and he teaches using Stan through the Royal Statistical Society each year.
