INCORPORATING STAKEHOLDER PREFERENCES INTO THE DEVELOPMENT AND PRESENTATION OF VALUE OF INFORMATION (VOI) ANALYSES TO INFORM REAL-WORLD RESEARCH INVESTMENT DECISIONS IN A US CANCER CLINICAL TRIALS GROUP

Purpose: Our objective is to use Value of Information (VOI) methods to help prioritize clinical trial research investment decisions within SWOG, a large US-based cancer clinical trials cooperative group.  Here we describe our process for incorporating SWOG decision makers’ preferences into generating and presenting VOI information that would meet SWOG’s need for timely decision-making.

Method: We used an iterative approach to develop a feasible and efficient process to generate and present VOI information to stakeholders.  We engaged directly with SWOG stakeholders during an in-person VOI training session and presented our initial process and results for 2 trial proposals recently reviewed by SWOG VOI results were developed using a minimal modeling framework to facilitate timely calculations. Extension of this work to 12 clinical trials reviewed by SWOG is currently underway.

Result: Stakeholder’s input about the types of information and the modes of presentation that they would find most compelling led to several modifications to our initial approaches, which mirrored VOI results and presentations styles typical in academic journals.  Based on this feedback, we assumed that the decision to adopt or reject a new treatment would be made on the basis of the trial’s primary endpoint, as is commonly done in the US.  Next, we calculated the value of additional evidence for trial proposals using three different value scales: (a) the primary endpoint of the trial, (b) quality-adjusted life years (QALYs), and, (c) expected net benefits.  The expected net benefits scale included disease and treatment-related costs in the projected outcomes of different decisions, but costs were not included in the decision to adopt or reject the treatment.  Lastly, we assumed that the implementation of research findings into clinical practice would require significant differences to reflect decision makers’ general aversion to adopt promising, but highly uncertain health technologies.  We will present the VOI results for the 12 historical studies using this approach.

Conclusion: VOI analyses are potentially useful to aid in the prioritization of clinical research. We worked with decision makers to develop a process for generating VOI estimates that is transparent and more closely aligned with real-world decision processes in the US. Future work will formally assess stakeholders’ acceptance of these methods and whether they influence decision-making in prospective clinical trial research investment decisions within SWOG.