Evaluating a patient-centered discrete choice experiment to quantify individual treatment preferences of newly diagnosed older adults with acute myeloid leukemia: A mixed methods study.

Abstract

256 Background: Numerous studies have demonstrated that physician perceptions of patient priorities and patients’ stated preferences differ substantially. Use of preference clarification using discrete choice experiments (DCEs) has been shown to improve preference-concordance for patients with solid tumors. Previously, we developed a DCE to describe preferences of survivors of acute myeloid leukemia (AML). We sought to evaluate this DCE among newly diagnosed patients with AML to inform shared decision-making. Methods: We used a sequential explanatory (quantitative to qualitative) mixed methods design to assess acceptability, feasibility, and content validity of the DCE to elicit individual preferences. Newly diagnosed older (≥ 60 years) adults with AML completed the DCE at the time of treatment decision. Patients and caregivers then completed semi-structured interviews to expand on quantitative findings. Based on initial results, healthy volunteers completed think-aloud sessions while completing the DCE and semi-structured interviews to further evaluate comprehensibility. Results: 47 participants were enrolled (18 patients [8 female, 10 male; aged 60-87], 16 caregivers, 15 healthy volunteers). Patients received best supportive care (n = 1), hypomethylating agents (n = 9), and high-intensity chemotherapy (n = 8). Feasibility/Acceptability: All patients completed the DCE (100% feasibility), and all reported that they answered according to their preferences (100% acceptability). Content Validity: 1) Relevance: 17/18 (94%) felt the DCE was relevant. 2) Comprehensiveness: In addition to the attributes in the DCE, provider recommendations and family considerations were noted to be critical in treatment decision-making. 3) Comprehensibility: Only 13/18 (72%) felt the DCE was easy to understand, and 14/18 (78%) felt it was easy to answer. Explanatory Interviews: Patients reported being in “shock,” “devastated” from receiving the diagnosis, feeling overwhelmed (“I was floored”) and not being able to reliably concentrate on the DCE due to these factors. Caregivers corroborated patients’ reports. Healthy volunteers also reported “information overload” and that the DCE was “technical” and the included outcome levels were “confusing.”. Conclusions: This DCE designed for older adults with AML was relevant, feasible, and acceptable. However, nearly 30% of patients reported difficulty understanding the DCE. Some reported not being able to attend to the complex tasks in the DCE, thereby compromising content validity of the measure. We suspect this was due to the distress caused by the diagnosis. This study demonstrates the challenge of developing valid preference elicitation instruments for newly diagnosed patients and highlights the need for extensive evaluation prior to clinical implementation.