Search

The aim of this paper is two-fold: to propose and construct a Bayesian meta-analysis model and to apply it to re-analyse the results of a recently-published meta-analysis concerning the efficacy of bone marrow-derived cells (BMC) transplantation in patients with heart diseases. The results based on the conventional and the Bayesian frameworks are compared and discussed. The outcome of interest is the combined weighted mean difference (WMD) in left ventricular ejection fraction (LVEF) between patients treated with BMC and their controls. Unlike the conventional approach, the proposed Bayesian model allows researchers to integrate quantifiable prior evidence (e.g., expert opinions) with published data. The conventional model showed that BMC could bring modest but statistically significant effect to adult patients with heart diseases. The Bayesian models based on concentrated and diffuse priors provided similar conclusion, although the results were heavily influenced by the selected priors. This suggests that the likelihood for the meta-analyses was rather "weak", thus one must interpret the published results and conclusion with extra care. More studies need to be conducted before a foregone conclusion can be made. The Bayesian models provided more insights to the problem and the nature of data selected for meta-analysis. [ABSTRACT FROM AUTHOR]