Related Experiment Videos
How to design digital health interventions with artificial intelligence: A scoping review
1School of Juridical Science, China University of Political Science and Law, Beijing, China.
Objectives:
Digital health interventions often struggle to scale due to rigid, resource-intensive design methods. A shift is underway toward evaluating AI-as-Agent, a dynamic partner in the design process, rather than a static clinical product. This systematic scoping review maps the landscape of generative digital health design and its impact on healthcare transformation.
Methods:
Following PRISMA-ScR guidelines, we searched PubMed, IEEE Xplore, and ACM Digital Library (2020-2025) for studies utilizing AI tools to actively support the design, development, or evaluation of digital health artifacts with human-in-the-loop validation.
Results:
Twenty-one studies met inclusion criteria. We identified distinct design roles for AI: Creator, Facilitator, Co-Designer, Tester, and Evaluator. Results indicate AI shows potential to mitigate critical design risks: synthetic data enables pre-clinical risk abatement; generative co-design fosters epistemic agency and patient inclusivity; and automated workflow analysis reduces clinician cognitive load. Based on these findings, we propose the AGENT Framework (Algorithmic Simulation, Generative Co-Design, Embedded Guardrails, Next-Best-Action, Tracked Evolution).
Conclusions:
AI is evolving into a collaborative design partner capable of enhancing the equity, scalability, and safety of digital health interventions. The AGENT framework offers a structured roadmap for stakeholders to integrate AI into the innovation lifecycle.
