Framing Benefits in Decision Aids: Effects of Varying Contextualizing Statements on Decisions About Sacubitril-Valsartan for Heart Failure

Abstract

Background. Presenting numeric data alone may result in patients underappreciating clinically significant benefits. Contextualizing statements to counter this may raise concern about absence of neutrality. These issues arose during construction of a decision aid for sacubitril-valsartan, a heart failure medication associated with a ∼3% absolute reduction in 2-year mortality that carries high out-of-pocket cost. A contextualizing statement framing this as a “pretty big benefit” was incorporated. The impact of statements like this within decision aids is unknown. Objective. This online Qualtrics survey sought to deepen understanding of benefit framing by testing the impact of varying contextualizing statements within a decision aid for sacubitril-valsartan. Design. Participants were randomly assigned to receive one of six abbreviated versions of a decision aid for sacubitril-valsartan that varied only by contextualizing statement (ranging from strongly neutral to strongly positive and using relative and absolute risk reductions). Participants were asked to answer questions regarding the likelihood of taking the medication at a cost of $50/month and their perception of the drug’s benefits. Results. A total of 1873 participants who were demographically similar to the heart failure population completed the survey. Fifty-four percent were willing to take sacubitril-valsartan at $50/month. Each of the five experimental contextualizing statements was compared with the baseline version; no significant differences were observed in reported likelihood of taking sacubitril-valsartan. After controlling for demographics and covariates, group assignment did not predict likelihood of taking the medication. Higher income, better self-reported health status, and younger age were associated with increased likelihood of taking sacubitril-valsartan. Limitations. This study used a hypothetical scenario and evaluated one method of delivering contextualizing statements. Conclusions. Contextualizing statements as tested within this decision aid did not affect decision making.