PURPOSE:

HTA bodies are expected to make decisions on the clinical efficacy and/or cost effectiveness, by examining the potential value in improving final patient relevant outcomes. For conditions with high disease burden and unmet treatment needs, data collection on final outcomes may delay urgent patient access to new treatments. In these circumstances, surrogate outcomes may substitute for and predict a relevant final patient outcome. However, their use can result in decision uncertainty and inaccurate technologies’ assessment.

Several frameworks have been proposed to increase transparency in the use of surrogate outcomes. However, most of these frameworks have not been extended to cover the impact of surrogacy use to the modelling assumptions in economic evaluations supporting HTA submissions.

DESCRIPTION:

The workshop will start with an introduction to the concept of surrogacy analysis and the importance to consider the totality of clinical and statistical evidence on the surrogacy relationship with final outcomes (10 min, Marc Buyse). Then, a newly developed web-based decision tool to support HTA agencies and payers when faced with surrogate endpoint evidence and address areas of uncertainties will be presented (15 min, Oriana Ciani). The workshop will continue with presenting the main statistical methodologies (risk equations, multi-state models and partition survival models) to accommodate for the use of surrogate outcomes in the economic modelling, the pros and challenges for each method and their implementation in HTAs (15min, Bart Heeg). Finally, Mario Ouwens will provide a summary of discussants’ recommendations on when and how surrogate outcomes can be reliably supporting HTA submissions. The session will conclude with an interactive discussion including the use of real time polling questions by presenting three hypothetical cases of HTA submissions using surrogate outcomes and potential impact on final decision making (15min). This workshop will benefit statisticians, payers and industry intending to use surrogate methodologies in submissions.