Interfaces and Agents for Medical Digital Twins: Design, Integration, and Evaluation

This scoping review charts how user interfaces (UI/UX, dashboards/visualizations, workflow UI, API/SDK/portals) and AI agents (LLMs, multi-agent systems, reinforcement learning, orchestration/automation, data mapping/ETL alignment, clinical decision-support agents) are designed, integrated, and evaluated within health/medical digital twins (HDTs). We adopt a tiered inclusion (Interface+Agent primary; Interface-only / Agent-only secondary) and a synchronization-level coding (S2/S1/S0) to capture both deployed and proto-twin ecosystems while preserving alignment to our research questions. Objectives. - Catalog interface components and interaction patterns used in HDTs. - Characterize AI-agent roles (simulation, decision support, orchestration, data mapping) and integration patterns. - Summarize evaluation methods (usability/HCI, human-in-the-loop, clinical/technical metrics). - Produce a taxonomy + codebook spanning UI features, agent roles, data modalities, settings/users, equity and governance. - Identify gaps/risks and propose reporting/design guidelines for future HDT interfaces and agentic workflows.