Guidance or Misdirection? Unpacking the Role of Feedback in Health Preference Assessments
Speaker(s)
Genie M1, Reed S2, Gonzalez J3, Ozdemir S1
1Duke University, Durham, NC, USA, 2Duke University’s School of Medicine, Durham, NC, USA, 31Duke Clinical Research Institute, Cary, NC, USA
Presentation Documents
OBJECTIVES: Discrete choice experiment (DCE) surveys frequently incorporate tutorials, comprehension questions, and practice questions to prepare respondents for the choice questions. This study investigated the influence of providing feedback to respondents on a dominated-choice question on subsequent choice behavior.
METHODS: A DCE was used to assess patient preferences for medical devices to manage heart failure among 626 patients. Respondents were given a dominated-choice question where Device A and Device B offered no benefits but carried risks (i.e., suboptimal option) compared to a “No Device” option. Among those selecting a suboptimal option (N=340), half were randomized to receive feedback, along with an opportunity to revise their choice while the other half did not receive feedback. The effect of feedback on preference for opting out (i.e., No Device) and choice consistency was examined using multinomial and heteroscedastic multinomial logit models.
RESULTS: Among those receiving the feedback (N=170), 71% continued to choose the suboptimal option in the dominated-choice question even after they received feedback. Respondents who received the feedback were more likely to choose the “No Device” option in the subsequent choice questions (p=0.002) when it was not clearly inferior. Notably, in the following choice task, 17.7% of the respondents who received feedback chose the "No Device" option compared to 8.2% of those who did not receive feedback (p=0.034). Contrary to our expectations, providing feedback led to a reduction in choice consistency.
CONCLUSIONS: While the provision of feedback can influence choice behavior in DCEs, most respondents in our study maintained their initial choice. This may stem from a misunderstanding of the options presented, or beliefs/hope that a device, by virtue of it being available, would afford some benefit not represented in the choice question. The study reveals potential unintended consequences of providing feedback in DCEs, suggesting that feedback alone may not be sufficient to “train” survey participants.
Code
MSR75
Topic
Patient-Centered Research
Topic Subcategory
Patient Behavior and Incentives, Stated Preference & Patient Satisfaction
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Medical Devices