BackgroundThe number of targeted antileukemic therapies for the treatment of AML has increased over the past decade. However, for patients who are not candidates for high intensive chemotherapy (HIC), a better understanding is needed of what treatment attributes patients and physicians value and how their preferences differ to potentially improve treatment satisfaction.ObjectiveTo understand the impact of treatment attributes on treatment selection, we quantified the preferences of: a) patients newly diagnosed with AML who are not candidates for HIC; and, b) the physicians who treat them.MethodsA discrete choice experiment (DCE) was performed to quantify the extent to which treatment attributes impact patients' and physicians' treatment decisions in various scenarios. A literature review was conducted to identify treatment attributes important to patients with AML. Following the review, one-on-one phone interviews were conducted with patients (United States [US], n=3; United Kingdom [UK], n=3) and physicians (US, n=2; UK, n=2) to finalize the attributes and levels included in the DCE. Patients who were eligible to participate in the online survey were adults with a self-confirmed AML diagnosis, who had not relapsed or been refractory to treatment, had not received a stem cell transplant, and had not received HIC or met one of the following criteria: aged ≥75 years, diagnosis of congestive heart failure, chronic kidney disease, or other types of cancer, or ECOG score of 3 or 4. Physicians included hematologists/ oncologists treating >5 patients with AML over the past year. The web-based DCE included choice cards showing 2 hypothetical treatment profiles with 6 attributes (chance of 2-year overall survival [OS], average quality of life [QoL], risk of serious infections, risk of grade 3/4 nausea, chance of achieving transfusion independence, and duration of hospitalization per year) at varying levels. Participants chose a preferred treatment for each choice card. Conditional logit regression models were used to estimate preference weights and to analyze the impact of treatment attributes on participants' choices.ResultsThe DCE was completed by 77 patients newly diagnosed with AML who had not received HIC (US, n=47; UK, n=30) and 145 physicians (US, n=48; UK, n=52; Canada, n=29; Australia, n=16). Mean patient age was 71.4 years; 51.9% were female. Mean (SD) time since AML diagnosis was 8.3 (8.2) months. Most physicians were hematologists (81.4%) and saw a median of 30 AML patients yearly. For patients, duration of hospitalization (decrease from 6 to 2 weeks/year) was the most important attribute followed by average QoL (increase from 50 to 85 on a 100-point QoL scale) and chance of 2-year OS (increase from 15% to 40%; Figure). Based on these findings, we estimated that patients were willing to accept a decrease in 2-year OS (from 40% to 15%) or an increase in risk of serious infections (from 5% to 20%) to decrease time spent hospitalized (from 6 to 2 weeks per year). For physicians, chance of 2-year OS (from 15% to 40%) was the most important attribute followed by average QoL (increase from 50 to 85 on a 100-point scale), risk of serious infections (from 20% to 5%), and risk of grade 3/4 nausea and vomiting (from 20% to 1%; Figure). Based on these findings, we estimated that physicians were willing to accept an increased risk of grade 3/4 nausea and vomiting (from 10% to 20%) in exchange for decreased time in hospital (from 6 to 2 weeks per year) and increased chance of achieving transfusion independence (from 35% to 55%) when other treatment attributes remained stable.ConclusionSignificant differences in treatment attribute importance for patients with newly diagnosed AML who had not received HIC were observed between patients and physicians. Patients most valued treatments that reduced hospitalization duration while physicians most valued treatments that improved chance of 2-year OS. These differences highlight the importance of a shared decision-making process when choosing treatments for patients with AML ineligible for HIC. However, given the variability among individual patients, it may be particularly worthwhile for physicians to initiate a discussion with patients prior to treatment selection to determine what treatment attributes each patient values most. Treatment selection could then be tailored based on attributes most valued by the patient and likely lead to improved treatment satisfaction. [Display omitted] DisclosuresZhou: Astellas Pharma, Inc.: Consultancy. Yang: Astellas Pharma, Inc.: Consultancy. Song: Astellas Pharma, Inc.: Consultancy. Marshall: Astellas Pharma, Inc.: Consultancy; Arthur JE Child Chair: Other: Indirectly related salary support. Griffin: Novartis: Patents & Royalties: Post marketing royalties from midostaurin; Astellas Pharma, Inc.: Consultancy. Saini: Astellas Pharma, Inc.: Consultancy, Honoraria. Shah: Astellas Pharma, Inc.: Current Employment; University of Michigan School of Public Health Department of Health Management and Policy Alumni Board: Other: Chair-Elect.