Practical Considerations for Incorporating Equity More Explicitly in Australian Health Technology Assessment Processes

Globally, advances in methods to examine equity alongside economic evaluations are being considered for application in health technology assessment (HTA) processes which inform decisions to publicly fund pharmaceuticals and medical services. In this practical application, we focus on the Australian context and explore how several of these methods might be used to incorporate equity more explicitly in HTA decisions. Specifically, we describe distributional cost-effectiveness analysis, extended cost-effectiveness analysis, equity weighting, multi-criteria decision analysis, mathematical programming and other quantitative approaches. We consider the feasibility and suitability of each method considering contextual factors, highlight challenges, describe current use in Australia and make recommendations for their application. We argue that there are opportunities to use aspects of several different methods in the Australian context which would illuminate how costs (including out of pocket costs) and benefits are distributed across the population, how normative concerns for equity influence the calculated social value of a new technology, how equity criteria complement other measures of value and how the technology impacts non-health outcomes. However, some of the stated requirements for the methods, such as using the same social groupings for all analyses, applying cost-effectiveness thresholds and the application of algorithms may restrict the usefulness of the evidence generated in a context where there are many priority populations of interest, several aspects of equity are considered important and there is a well-established deliberative approach to decision-making. We recommend that the equity implications of new health technologies should be universally evaluated. However, a tailored and transparent approach should be taken whereby specific equity analyses consider the context of the technology and the chosen methods are well justified.