PURPOSE: Time and cost of de-novo, stated-preference survey development, data collection, and analysis often are barriers to incorporating patient preferences in decision making. In some therapeutic areas, there is enough preference data to begin evaluating and using previous studies to predict values in a new context. This workshop will focus on how to leverage existing preference evidence obtained in one or more treatment contexts to impute or “transfer” health-improvement benefits to answer different questions. Participants will learn how to assess the robustness of preference evidence bases to support benefit transfers and what statistical methods are available to conduct principled and transparent benefit transfers.

DESCRIPTION: Workshop attendees will obtain a working knowledge of available methods for conducting valid benefit transfers. The workshop will review a) how benefit-transfer methods have extensively been used in non-health applied economics, b) how familiar biostatistical tools can be adapted to benefit-transfer applications, and c) how multiple strategies were used in a recent health application. Dr. Johnson will chair the session and introduce the topic in the context of the extensive benefit-transfer literature in environmental economics (10 min.), Dr. Groothuis-Oudshoorn will show how established statistical methods for estimation, attribution, and prediction can apply to benefit transfers (15 min.), and Dr. Gonzalez will illustrate how these methods were applied to evaluate the preference evidence base in psoriasis and to derive consensus values and functions to predict maximum acceptable treatment risk for given therapeutic benefits and to evaluate the impact of cross-study factors (15 min.). Audience participation will include identifying problems and solutions for a hypothetical case study in inflammatory bowel disease (20 min). This interactive and informative workshop will be valuable to researchers, clinicians, and industry analysts who are interested making patient-preference evidence available for a wider range of decision-making applications.