Bayesian Modeling for Economic Evaluation Using “Real-World Evidence”

Abstract

This chapter presents an example of the use of Bayesian modeling with “real world evidence”. Real World Evidence (RWE) is a terminology increasingly often used to describe observational data obtained using Electronic Health Records. Primary care databases can be generally defined as collections of de-identified individual level data from a network of general practices across a given jurisdiction. Real world evidence is potentially exciting means to bring external information into analyses based on immature data. The use of RWE has great potential, particularly when embedded in a wider economic modeling, which is based on the statistical component as the, arguably, important building block, but that at the same time decisively move forward from the static point of view of finding statistical significance, in favour of a more comprehensive decision-making approach. The Bayesian machinery has the advantage of allowing for a principled integration of data sources, which in cases such as this might allow a substantial “regularization” of the resulting inference.