Combining Information for Quantitative Decision-making in Drug Development

Abstract

Taking decisions during the development of a new drug requires combining many and varying pieces of information. The interconnections between these different pieces of information are often only partially known and in some cases merely conjectured. Despite this uncertainty, decisions must be taken – often with regard to the balance of risk and benefit – in order to make progress. A clear, consistent and efficient methodology for describing the structure of a project and the comprehensive state of knowledge at any point in time is required in order to add transparency to the decision process. A viable methodology must allow for a natural characterization of the uncertainty inherent in the drug development process and a relatively easy implementation. We present a possible solution that satisfies these requirements. The foundation of the proposed approach is based on a beta-binomial update mechanism well known in Bayesian statistics. This work provides a consistent framework for solving hierarchical multicriteria decision problems. Copyright © 2014 John Wiley & Sons, Ltd.