Bayesian Network Meta-Analysis for Health Technology Assessment and Evaluation for Investigative Treatment
Baoguang Han, Eli Lilly and Company
Keywords: Bayesian, Network Meta-analysis, Health Technology Assessment, Indirect Comparison, Drug Development
Evidence-based decision making in health technology assessment and drug development often requires comparisons of all available treatments to assess their efficacy and safety profiles. With the lack of information on direct, head-to-head comparison of different treatments, network meta-analysis (NMA) can provide useful evidence for selecting the best treatment strategy for health care intervention or for assessing the benefit of an investigative treatment vs. alternative treatments. Bayesian approach offers a flexible framework for NMA, given its ability to explore a wide range of modeling structures, take into account of multiple sources of evidence and provide straightforward probability statements around the parameters of interest useful for decision makers. Properly implemented, Bayesian methods improve the efficiency and accuracy of decision making and enhance trial design, resulting in efficient use of historical and current data. In the setting of NMA, a specific advantage of the Bayesian methods is that the posterior probability distribution allows ranking of competing interventions. We will describe Bayesian framework for NMA, and a web-interface tool called BAyesian Tool for Meta-Analysis of Networks (BATMAN), which allows NMA for both continuous and binary outcomes and is flexible to adjust for study-level variations. We will discuss applications of Bayesian NMA in health technology appraisals and evaluation of investigative treatment.
Important Dates & Deadlines
- October 9 - 11, 2013