PREFER-DMT: A Discrete Choice Experiment Framework for Eliciting Stakeholder Preferences on Disease-Modifying Therapy Implementation in European Healthcare Systems

Background: Health technology assessment (HTA) bodies across Europe are preparing to evaluate Alzheimer’s disease-modifying therapies (DMTs), yet no standardised framework exists for eliciting and comparing stakeholder preferences on DMT implementation across different healthcare systems. Existing discrete choice experiment (DCE) evidence on Alzheimer’s treatment preferences is limited to the US context and focuses narrowly on treatment characteristics, neglecting the implementation factors diagnostic pathways, monitoring requirements, geographic access, waiting times that are critical for European healthcare systems with capacity constraints. Objective: To present the PREFER-DMT (PREFerences for European Reform of Disease-Modifying Therapy) framework: a standardised DCE protocol designed for multi-country deployment across European healthcare systems, enabling cross-national comparison of stakeholder preferences for DMT implementation. Methods: The PREFER-DMT framework comprises: (1) a nine-attribute, three-domain DCE design covering treatment characteristics (efficacy, ARIA risk, administration, duration), implementation factors (diagnostics, monitoring, geographic access, waiting time), and cost; (2) a four-stakeholder sampling strategy (patients/caregivers, clinicians, HTA professionals, general public) with country-specific sample size calculations; (3) a Bayesian D-efficient experimental design (Ngene) with 36 choice tasks blocked into 3 versions of 12 tasks each; (4) a three-stage analytical pipeline (conditional logit, mixed logit, latent class); (5) country-specific attribute level calibration to reflect local healthcare system characteristics; and (6) an interactive R Shiny dashboard for results dissemination and stakeholder engagement. The framework was validated using Monte Carlo simulation calibrated to published clinical trial evidence and the DCE literature. Nine attributes were identified through systematic review (12 relevant studies), expert consultation, and alignment with published health system preparedness assessments. Results: Simulation-based validation demonstrates the framework’s analytical capability. Under calibrated coefficients, the nine-attribute design achieves a D-error of 0.0023, adequate for mixed logit estimation with preference heterogeneity. Key illustrative findings include: (a) implementation-related attributes collectively account for 42.5% of total preference importance, exceeding treatment characteristics alone (35.9%); (b) geographic access to treatment centres and diagnostic pathway choice are the most highly valued implementation factors (WTP: EUR 20,571 and EUR 18,571 per year, respectively); (c) latent class analysis identifies three distinct preference segments (efficacy-driven, access-prioritising, and risk-averse) with implications for patient-centred implementation. The framework is designed for adaptation to different European contexts through country-specific level calibration: for example, cost levels calibrated to national pharmaceutical pricing, travel distances calibrated to geographic dispersion of specialist centres, and waiting times calibrated to local healthcare system capacity. Conclusions: PREFER-DMT provides a standardised, validated, and immediately deployable DCE framework for multi-country preference elicitation on Alzheimer’s DMT implementation in Europe. By covering implementation factors alongside treatment characteristics, the framework generates evidence directly relevant to the system-level decisions that European HTA bodies must make. The framework is operationalised as an open-access interactive platform and is available for adoption by national DMT working groups, HTA agencies, and research consortia.