WM-voice: An international discrete choice experiment on treatment preferences in 1,455 patients with Waldenström macroglobulinemia

Abstract Introduction: Waldenström macroglobulinemia (WM) is a rare and incurable low-grade non-Hodgkin lymphoma. Therapeutic options are rapidly expanding, with no consensus on the preferred approach. Available options vary in their properties, such as efficacy, mode of administration, and side effects. An improved understanding of patients' preferences and trade-offs may guide individual management in clinical practice and set priorities in therapeutic innovation. To date, structured data on this topic in WM are limited to one small Dutch study. The WM-VOICE study was designed to quantify international WM patient preferences for treatment properties and identify preference variation across subgroups. Methods: This study employed a discrete choice experiment (DCE) amongst WM patients in Australia, Canada, the Netherlands, the United Kingdom (UK), and the United States of America (USA). Attributes and levels were selected through interviews with WM patients and expert hematologists. These included: response duration (2, 4, 6, or 8 years), administration mode (hospital infusions or pills at home), treatment duration (6 months, 2 years, or ongoing), temporary toxicities (grade 0-1, 2, or 3-4), persistent toxicities (grade 0-1 or 2), and risk of secondary malignancies (increased or not). Following DCE design generation and pilot testing, the final anonymous web-based survey was disseminated via patient organizations and clinics. It included demographic and clinical questions, as well as 16 DCE choice sets. Each set asked respondents to choose between 2 hypothetical unlabeled treatment options, defined by 6 key attributes. Repeated choice data were analyzed using a panel mixed logit model, yielding conditional relative importance (CRI), reflecting each attribute's influence on the likelihood of selecting a given option based on its levels, and willingness to trade, calculated by dividing level coefficients by the marginal utility of a 1-year response duration increase (assuming linear coding across the 2-, 4-, 6-, and 8-year levels). Results: A total of 1,455 WM patients participated in the DCE (Australia: n=209, Canada: n=219, the Netherlands: n=340, UK: n=214, USA: n=473). Most respondents (75%) were 61-80 years old, and 818 (56%) were male. With a median time since diagnosis of 6 years (IQR 3-10), 79% were previously treated for WM. In the DCE, all attribute levels had significant impact on respondents' treatment choices. Preferences were largely consistent across all 5 countries. Response duration and temporary toxicity had the highest importance (CRI 29.7%, 95% CI: 28.8–30.6 and 25.6%, 95% CI: 24.8–26.5, respectively), followed by secondary malignancy (17.7%, 95% CI: 16.8–18.6) and persistent toxicity (17.4%, 95% CI: 16.5–18.2). Administration mode (4.9%, 95% CI: 4.0–5.8) and treatment duration (4.6%, 95% CI: 3.8–5.4) had the lowest CRIs, favoring oral treatment at home over hospital infusions, and fixed-duration over ongoing therapy. Prior BTK inhibitor treatment significantly increased preference for oral over intravenous therapy (p=0.03). Other patient characteristics, including sex, age (≤70 vs >70 years), treatment status (untreated vs treated), prior chemotherapy, travel time to the hospital (<1 vs >1 hour), educational level, and current treatment status, did not impact choices. Within the context of the DCE, respondents were willing to trade 3.8 years and 5.6 years of response duration to move from short-term grade 3-4 toxicity to grade 2, or to grade 0-1, respectively. They would also trade 3.8 years to move from persistent grade 2 toxicity to grade 0-1, and 3.7 years to avoid an increased risk of secondary malignancy. Respondents accepted a 1.0 year reduction in response duration to take treatment orally instead of via infusions, 0.2 years to switch from ongoing to fixed-duration treatment of 2 years, and 1.0 year to switch to a 6-month treatment. Conclusion: The WM-VOICE study is the first to investigate international WM patients' treatment preferences. We identified 6 treatment properties significantly influencing choices, with strong similarity across 5 Western countries. Our results highlight that patients' choices are mainly driven by long response duration and avoidance of severe toxicity, whereas mode of administration and treatment duration are less influential. These results could be implemented in clinical practice to support shared decision-making and inform drug development and clinical trial design in WM.