Do We Need Advanced Modeling Methods for Equity-Informative Economic Evaluations?

ISSUE: There is an increased interest in incorporating health equity considerations in the value assessment of health technologies. Distributional cost-effectiveness analysis (DCEA) is an intuitively appealing extension of conventional cost-effectiveness analysis to quantify health equity impacts and facilitate potential trade-offs between improving total health and reducing health inequalities. However, there is debate about the feasibility of implementing a DCEA, especially for newer health technologies. This session will discuss whether advanced modeling methods are beneficial or even needed for DCEA, and whether these methods can help with some of the evidence challenges.

OVERVIEW: In practice, DCEA relies on mathematical modeling that integrates different sources of evidence to estimate expected outcomes and opportunity costs for the compared interventions by subgroup defined according to individual and non-individual factors such as race/ethnicity, socioeconomic status, and geography. The need to model distributional effects has implications for the model structure and simulation approach, as well as evidence requirements. After introducing the session and an opening statement on the relevance of DCEA (Avanceña, ~11 minutes), panel members will present key features of individual-level and discrete event simulation of (theoretical) relevance for DCEA (Degeling), value of information (VOI) in the context of DCEA (Kunst), and estimation of model input parameters (Jansen) (~8 minutes each). Subsequently, the panel members and audience will have a debate lead by the moderator (25 minutes). In addition to questions from the audience, the following questions will be discussed: Do we need individual-level or discrete event simulation modeling for DCEA? Are we replacing structural uncertainty with parameter uncertainty by using more complex models, and would that provide any benefit? What are the most important model parameters, and what are the implications for data collection and parameter estimation?