Participatory Digital Twins for Chronic Care: From Predictive Models to Shared Sensemaking (Preprint)

UNSTRUCTURED Health digital twins, computational models that integrate longitudinal data, simulation, and forecasting, are increasingly proposed as tools for chronic care management. Most current implementations, however, are expert-oriented, prioritizing technical optimization and clinical prediction while offering limited support for patient understanding, engagement, or participation. This orientation is particularly misaligned with chronic care, which unfolds largely outside clinical settings and depends on patients’ daily decisions, social context, and sustained engagement over time. In this Viewpoint, we argue for reframing digital twins as participatory systems that support shared sensemaking among patients, caregivers, and clinicians, rather than functioning solely as directive, expert-facing tools. We propose a conceptual framework that positions participatory digital twins as boundary objects capable of bridging computational models, clinical reasoning, and lived experience. Within this framework, generative artificial intelligence serves as a translation and interaction layer, enabling plain-language dialogue, exploration of uncertainty, and “what-if” reasoning that allows users to interpret model outputs in relation to their own contexts, goals, and constraints. We outline key design principles for participatory digital twins, including visible uncertainty, negotiated rather than prescriptive care, mechanisms for incorporating patient context and social drivers of health, and governance structures that support accountability and recourse. By shifting the focus from optimization alone to understanding, interaction, and trust, participatory digital twins offer a pathway toward more equitable, human-centered, and sustainable models of AI-enabled chronic care.