Estimating value of information: do we really have speed??

Abstract

Healthcare decision makers, such as the National Institute of Health and Care Excellence in the UK, use cost-effectiveness analysis and modelling to compare the costs and effects of disease management strategies. These analyses rely on limited evidence and decisions are often uncertain. Value of information (VoI) analysis quantifies the monetary value to decision makers of gathering further evidence. VoI requires nested Monte Carlo simulation to estimate the uncertain benefits of further research, which is computationally impractical for all but the simplest of cost-effectiveness models. Model regression and approximation approaches, including Gaussian processes, generalised additive models, and integrated nested Laplace approximation (INLA), have come into use as short-cut approaches to estimating VoI. Uptake has been boosted by the easy-to-use Sheffield Accelerative Value of Information (SAVI) online tool. However, these approaches may be unreliable for realistic models.

In my talk, I will explain these issues in greater detail and highlight problems with model regression and approximation when applied to realistic economic models. As an alternative, I will present novel adaptations of advanced Monte Carlo sampling schemes from computational finance to the estimation of VoI. These achieve the same accuracy and precision of standard Monte Carlo with lower computational cost by minimising the variance and bias of their VoI estimators. I will discuss both Quasi Monte Carlo and Multilevel Monte Carlo estimation of VoI and will apply them to several examples of cost-effectiveness models. Examples will include the recently published directly acting oral anticoagulants (DOACs) for prevention of stroke in atrial fibrillation, constructed data models in depression, and a comparison of the Ross procedure to biological and mechanical arotic valve replacement surgery. Results will be compared with estimates from SAVI and INLA.

Brief Biography

I started my academic journey with a degree in Maths from Trinity College Dublin before heading to Oxford for an MSc in mathematical finance. A brief stint at the National Centre for Pharmacoeconomics was enough to convince me to pursue a career in health economics and medical statistics, which led me to a PhD at the Biostatistics Unit in Cambridge. I’m now a Research Fellow in statistical modelling at the University of Bristol, mostly working with Professor Nicky Welton, and work as a consultant for Novartis, Hoffman-La Roche, and Pfizer Inc. My research interests are efficient value of information (VoI) analysis, VoI for adaptive trial designs, network meta-analysis on limited or disconnected evidence networks, population adjusted indirect comparisons, and structural uncertainty in cost-effectiveness models. I have a general interest in encouraging the use of R, rather than Excel, for cost-effectiveness analysis. I spend my ample free time with my wife and newly born baby boy.