Understanding patient preferences for treatment of acute leukemia in the unites states: A discrete choice experiment

Abstract

Abstract BACKGROUND: Acute leukemia is a cancer of the white blood cells that progresses rapidly and aggressively. For adult patients, failure to achieve remission after first-line treatment (“refractory disease”) or recurrence following remission (“relapse”) is associated with poorer outcomes and difficult treatment choices. These treatment choices typically involve complicated trade-offs between efficacy, quality of life (QoL) and treatment burden and therefore, treatment decisions may be influenced by individual patient preferences and priorities. Quantitative preference studies can inform regulatory, clinical and shared decision-making; however, evidence remains limited. METHODS: An online discrete choice experiment (DCE) was conducted to quantify relapsed/refractory treatment preferences among adults living with acute leukemia in the United States (US). The design of the DCE was informed by formative qualitative research consisting of online bulletin boards and refined using individual think-aloud pre-testing interviews. This resulted in the inclusion of five attributes: mode of administration, QoL during treatment, chance of responding to treatment, duration of treatment response and QoL during treatment response. In the DCE, participants were presented with 12 choice tasks comparing two unlabelled hypothetical treatments. Participants were eligible if they were adults that had a diagnosis of acute leukamia and were recruited via patient advocacy groups. The choice data were analysed using latent class models to enable an exploration of preference heterogeneity, and model coefficients were used to estimate relative attribute importance (RAI) scores. RESULTS: A total of 88 patients in the US completed the DCE. The mean age of the sample was 60, with an average of three years since diagnosis at the time of survey completion. The sample was predominantly female (59%) and acute lymphoblastic leukemia (ALL) was the most common subtype (46%), followed by acute myeloid leukemia (AML; 38%) and acute promyelocytic leukemia (APL; 16%). At the time of the survey, most participants were either not receiving active treatment (36%) or were on maintenance therapy (24%). The latent class analysis identified three distinct clusters of preferences amongst the sample. Cluster 1 (with a class membership of 38%) were focused predominantly on efficacy, focusing almost entirely on the chance of responding to treatment (RAI=69.4%) over all other attributes. In contrast, the most important attribute in cluster 2 (class membership 12%) was QoL during response (RAI=38.8%), followed by duration of response (RAI=31.9%) and chance of response (RAI=21.3%). Finally, in cluster 3 (the largest cluster with 50% class membership) preferences were generally more balanced across the attributes. However, compared to the other clusters, MOA was valued more highly (RAI=15.5%) and QoL during response was not valued as highly (RAI=10.1%). The variation in RAI scores across these clusters indicate significant preference heterogeneity across the acute leukemia patient population. One of the only similarities across clusters was that QoL during treatment was not considered important in any cluster (RAI<6%). CONCLUSIONS: This study highlights the diversity of treatment preferences among people with acute leukemia in the US, revealing three distinct subgroups with differing priorities in the relapsed/refractory treatment context. While many patients prioritise the chance of responding to treatment, others are far more focused on quality of life. Recognising this heterogeneity may support more individualised treatment discussions and better align care with patient values. These findings contribute to the growing body of evidence supporting the role of patient preferences in treatment planning and decision-making in oncology.