ARPA-H's ADVOCATE program has struck a nerve: roughly 300 research groups, start-ups, and clinics submitted solution proposals—more than any other program at the US health research agency founded in 2022. ARPA-H names the reason precisely: over 200,000 people die each year in the US from heart disease that would have been preventable with early detection. ADVOCATE, formally Agentic AI-EnableD CardioVascular CAre TransfOrmation, aims to become the first FDA-authorized autonomous AI system for cardiology.
The problem: 200,000 preventable deaths per year
Heart disease is the leading cause of death in the US. According to the American Heart Association's 2026 Heart and Stroke Statistics report, approximately 941,652 Americans died from cardiovascular disease in 2022. ARPA-H quantifies the preventable deaths at over 200,000 per year—cases where early detection and intervention would have significantly altered the outcome.
Rural areas in the US are particularly affected. ARPA-H explicitly notes on its program page that millions of Americans live in regions without adequate access to cardiologists. Heart rhythm disorders, early heart failure signs, and pre-infarction episodes often go undetected until the patient is already in emergency care. ADVOCATE aims to close this care gap through continuous AI-supported monitoring.
What ADVOCATE is meant to do: autonomous rather than reactive
The term 'agentic AI' differs fundamentally from conventional AI diagnostic systems. A classical AI tool in cardiology analyzes an EKG strip at a physician's request and delivers a result. An agentic system acts autonomously: it continuously monitors vital signs, assesses risks in real-time, independently decides whether a physician alert or direct intervention is needed, and carries it out without a human authorizing each individual decision.
ARPA-H describes core technologies as real-time EKG monitoring, automated risk prediction based on patient history and biomarker data, and autonomous recommendations for medication adjustments or emergency dispatch alerts. The program explicitly addresses the physician shortage in cardiology: where no specialist is reachable, the AI system should take the first response.
300 submissions: ARPA-H's most oversubscribed program
ARPA-H was founded in 2022 following the model of DARPA, the research agency of the US Department of Defense, and funds high-risk research with transformative potential in medicine. After the January 2026 announcement, ADVOCATE received approximately 300 solution proposals from research institutions, universities, and companies. According to ARPA-H, this is a record for the still-young agency.
Teams were scheduled to be selected by June 2026. After one year, an intermediate selection step will occur, continuing the most promising approaches. ADVOCATE is thus not the development of a single AI system but a competition: multiple teams develop different approaches in parallel before ARPA-H brings the best solution to FDA authorization.
39 months until FDA review: autonomy and liability as open questions
The total program duration is 39 months. This means: if teams start in summer 2026, an FDA decision is realistically possible by end of 2029. By then, several fundamental questions must be resolved that go beyond technical development.
The American Medical Association (AMA) has repeatedly demanded in its positioning on AI in medicine that autonomous systems need clear liability rules. If an AI system autonomously makes an incorrect diagnosis or triggers a wrong intervention, it remains unclear who is liable: the manufacturer, the hospital, or the physician deploying the system. The FDA, for its part, has clarified in its guidance for AI-based software as medical devices that systems making autonomous decisions with direct patient impact undergo stricter review than mere diagnostic support.
Add to this data protection: 24/7 monitoring of heart patients continuously generates highly sensitive health data whose storage and sharing are governed by different legal frameworks in the US and EU. In Germany, a comparable system would be bound by requirements of the Social Security Code and GDPR, which doesn't automatically permit importing a US-approved system. Regardless, ADVOCATE provides an important impetus: Germany's Association of Statutory Health Insurance Physicians has documented rural cardiology care gaps as a growing problem.