Title: bmeta: An R package for Bayesian meta-analysis
Authors: Tao Ding - University College London (United Kingdom) [presenting]
Abstract: It is often the case that important research questions are studied more than once by different research teams at different locations and the outcomes from small studies can be diverse and conflicting. This may result in difficulties in decision making. However, combining available information from multiple sources to generate an integrated result may provide more indications. Meta-analysis is a commonly used statistical approach to realise this goal by integrating results from independent studies and is considered to play an essential role in evidence-based medicine. Since most applied implementations of meta-analysis are conducted under the Frequentist paradigm, there is the need to develop a package using Bayesian meta-analytic methods given its advantages (e.g. the observed data can be complemented by some prior belief and it takes fuller account of the uncertainties related to both parameter values and models). Therefore, the bmeta package for R is created. The b in the name stands for Bayesian.