How Do Members of the Duchenne and Becker Muscular Dystrophy Community Perceive a Discrete-Choice Experiment Incorporating Uncertain Treatment Benefit? An Application of Research as an Event

Background Best-worst scaling methods have been used in several Duchenne and Becker muscular dystrophy (DBMD) studies to quantify patient and caregiver priorities and preferences and promote patient-focused drug development (PFDD). We sought to assess the extent to which different members of the DBMD community would accept a discrete-choice experiment (DCE) that incorporates uncertainty regarding individual-level benefit. Methods A community advisory board encouraged the development and testing of a DCE to further examine treatment preferences in DBMD and to facilitate the inclusion of a policy-relevant uncertainty attribute. The DCE assessed preferences across a primary outcome (muscle strength) and several risks (uncertainty regarding treatment benefit, kidney damage risk, and fracture risk). The single instrument was tested among adult patients, caregivers, and professionals at the national Parent Project Muscular Dystrophy annual meeting. The DCE was analyzed using conditional logit. Instrument acceptability was evaluated using a previously developed set of questions assessing ease of understanding and answering, and if answers reflected the respondents' real preferences. We proposed a 75% agreement rate as a threshold of acceptability, and used a Z score to assess if this was met, exceeded, or rejected. Results A total of 161 people completed the survey including 9 patients, 87 caregivers, and 65 professionals. Patients reported high acceptability across all evaluation items (p values > 0.21). Caregivers and professionals exceeded the benchmark of acceptability on understanding and reflecting real preferences (p < 0.001). Professionals met the benchmark (p = 0.08) for ease of answering, but caregivers did not (p < 0.01). DCE results demonstrated thatall groups made meaningful trade-offs, with patients being less tolerant of risks than either caregivers or professionals (p < 0.001), and with no significant difference between caregivers and professionals (p = 0.46). Conclusions This study demonstrates the acceptable application of a single instrument across a multi-stakeholder population that used a complex preference method and included a policy-relevant uncertainty variable. Ease of answering was lowest among caregivers, but a post-hoc analysis revealed that it was most difficult for those with children under the age of 10, while those with older children met the threshold. The success of this study has laid the foundation for a global study of DBMD preferences using this method.