Trump Administration Backs Clinical AI Agents With Three-Year FDA Approval Pathway

The initiative, called Agentic AI-Enabled Cardiovascular Care Transformation (ADVOCATE), aims to create the first FDA-authorized agentic AI technology capable of supporting clinical decision-making in a high-risk setting.

The Trump administration is moving to establish a regulatory pathway for generative clinical AI, with the Department of Health and Human Services’ Advanced Research Projects Agency for Health (ARPA-H) soliciting proposals to develop agentic AI assistants designed for cardiovascular care and intended to receive US Food and Drug Administration authorization within about three years.

The initiative, called Agentic AI-Enabled Cardiovascular Care Transformation (ADVOCATE), aims to create the first FDA-authorized agentic AI technology capable of supporting clinical decision-making in a high-risk setting. To date, the FDA has approved only predictive AI tools, not generative or autonomous systems that can plan and execute clinical tasks.

Under ADVOCATE, ARPA-H plans to fund the development of two AI systems. The first is a patient-facing agent focused on cardiovascular disease, which accounts for roughly 200,000 deaths annually in the US, despite the availability of low-cost and widely used treatments. The second is a supervisory agent designed to monitor continuously learning AI systems, addressing a key gap in healthcare AI governance.

The cardiovascular AI agent is expected to support clinicians by performing tasks such as scheduling appointments, adjusting medications, and providing diet and exercise guidance. It will connect to electronic health records, integrate data from wearable devices, and be available to patients around the clock. According to ARPA-H, the agent should be capable of performing certain clinical assessments that a cardiologist might conduct over the phone, such as identifying signs of heart failure or modifying medication doses.

Haider Warraich, M.D., program manager at ARPA-H, said the agency intentionally chose a clearly defined, high-risk clinical area to ensure the technology falls squarely under the FDA’s regulatory authority. The goal is for ARPA-H and the FDA to jointly determine what evidence is required for generative AI authorization and how such systems should be monitored over time.

The full ADVOCATE program, including development, deployment, and FDA review, is scheduled to span approximately 39 months. AI agents will be co-developed with health systems and deployed locally after two years to evaluate clinical and operational impact.

ARPA-H plans to select participating innovation teams by June 2026, followed by a competitive down-selection process after the first year of development.

Industry groups have offered cautious support. The American Heart Association said it backs rigorous testing of AI approaches for heart and stroke care, while the American Medical Association emphasized that autonomous AI in cardiovascular care raises patient safety concerns if deployed without clear physician oversight.

Stay tuned for more such updates on Digital Health News