Development of a Bayesian Unit for Health Decision Sciences

This proposal will develop a Bayesian unit for health decision making. A Bayesian approach formalises the bringing together of data from structured sources/trials with existing knowledge and uses formal decision criteria to identify optimal courses of action for a given health related problem/question. The centre will prioritise three main areas for development; effectiveness research, calibration of disease models, and the choice of decision criteria to be used in making strategic health investment decisions.  
The effectiveness research strand includes the development of models to extend the meta-analytic approaches to include data from observational as well as randomised controlled trials. It also permits the incorporation of individual patient level data. Direct access will be provided to users (e.g. health services researchers) to outline tools for performing mixed treatment comparisons.  
The calibration of disease models allows outcomes of interventions such as screening, vaccination and treatment to be more accurately predicted. This has already been an essential component in previous work on the cost effectiveness of new technologies, such as HPV vaccination and population colorectal cancer screening. Calibration research for this proposal includes further development of approaches to take into account more recent advances in Bayesian computation and will allow a calibration for models for other diseases (e.g. Hepatitis C) for which treatment strategies are being assessed. 
The final strand on decision criteria analysis questions and formalises our current health decisions. The core work of the health related partner, the National Centre for Pharmacoeconomics (NCPE), is to advise health payers at national level.  Existing work has yielded quantitative evaluations of the cost effectiveness of pharmaceutical interventions, vaccines and medical devises.  A retrospective examination shows that the decision on whether to pay for a technology has differed depending on the nature of the disease, types of individuals affected and who the payer was. This strand will examine historical decisions in the international context with a view to formulating policy for payers in the future.

Award Date
22 March 2013
Award Value
Principal Investigator
Professor Cathal Walsh
Host Institution
University of Limerick
Research Leader Awards