ARPA-H Commits $62.7 Million to Artificial Intelligence Bots for Heart Failure Treatment

The Advanced Research Projects Agency for Health (ARPA-H), a federal research agency modeled after the Defense Advanced Research Projects Agency (DARPA), has officially launched an ambitious initiative to revolutionize cardiac care. By committing $62.7 million to the development of artificial intelligence-driven "agents," the government aims to bridge the persistent gap in care for the 6.7 million Americans living with heart failure. The program, dubbed ADVOCATE, seeks to move beyond passive monitoring tools, pushing instead toward a future where autonomous AI systems are authorized by the Food and Drug Administration (FDA) to perform high-level clinical tasks, including symptom assessment, medication management, and the ordering of diagnostic laboratory tests.

This initiative marks a significant shift in how the federal government approaches health technology. By funding private companies and prestigious academic institutions to build these "autonomous agents," ARPA-H is effectively signaling that it views AI not merely as a diagnostic assistant, but as a potential surrogate for clinical decision-making, particularly in settings where access to specialized cardiology care is constrained.

The Scope and Strategy of the ADVOCATE Program

The ADVOCATE program is designed to address the "last mile" of medical care—the day-to-day management of chronic conditions that often results in hospital readmissions or health decline. Heart failure, a chronic condition where the heart is unable to pump blood as efficiently as required, is one of the most resource-intensive diagnoses in the American health care system.

Under the initial phase of the program, ARPA-H has awarded $33.7 million to a diverse cohort of entities. The recipients include technology-forward health firms Atman Health, UpDoc, and Tempus AI, as well as institutional research powerhouses such as Stanford University, Duke University, and the Kaiser Permanente health system. While $33.7 million has been allocated for the first year, the agency has built flexibility into the contracts, allowing for the remaining $29 million to be adjusted based on the development milestones achieved by the participating teams.

The technical goal is to move from "decision support"—where a doctor sees a recommendation and chooses whether to act on it—to "autonomous agents" that can perform clinical functions with minimal human intervention. These bots would be designed to ingest patient data, analyze it against established clinical guidelines, and provide actionable interventions, such as adjusting diuretic dosages or prompting for specific blood panels.

A Chronology of the Modern Health AI Push

The launch of ADVOCATE did not occur in a vacuum. It is the culmination of a multi-year trend toward digitizing and automating the clinical workflow.

• 2020-2021: During the height of the COVID-19 pandemic, the limitations of the traditional, visit-based care model became starkly evident. The rapid rise of telehealth and remote patient monitoring (RPM) normalized the use of digital interfaces for primary care, setting the stage for more complex automation.
• 2022: ARPA-H was officially established as an independent agency within the Department of Health and Human Services. Its mandate focused on "high-risk, high-reward" projects that have the potential to solve intractable health challenges.
• 2023: The FDA began releasing a steady stream of guidance documents regarding AI/ML-enabled medical devices, specifically focusing on "predetermined change control plans," which allow manufacturers to update software algorithms without needing a new, full-scale regulatory review for every minor iteration.
• 2024: ARPA-H solidified its focus on AI, culminating in the announcement of the ADVOCATE program. This reflects a strategic pivot from merely tracking health data to actively managing health outcomes via algorithmic intervention.

The Clinical and Socioeconomic Context

The urgency behind this investment is rooted in a sobering statistical reality. According to the American Heart Association, nearly half of the 6.7 million Americans with heart failure struggle to receive "optimal" care. The barriers are multifaceted: a shortage of cardiologists, the high cost of travel to specialized centers for rural patients, and the sheer volume of data generated by wearable sensors, which often overwhelms existing clinical staff.

In rural or underserved areas, a patient might wait months to see a heart failure specialist. In the interim, their condition could fluctuate, leading to emergency room visits that could have been prevented with timely medication adjustments. By deploying AI agents that can operate at the "top of their license" (or in this case, the top of their programming), ARPA-H hopes to democratize access to the same level of care currently available only at elite academic medical centers.

Regulatory and Ethical Implications

The transition toward autonomous clinical AI carries profound regulatory and ethical questions. Currently, the FDA classifies most AI-driven software as "Software as a Medical Device" (SaMD). The ADVOCATE program essentially tasks these companies with creating a new class of devices that are not just analytical, but procedural.

If an AI bot orders a lab test or recommends a dosage change that results in an adverse event, the legal and ethical liability remains a subject of intense debate. Who is responsible: the software developer, the health system that deployed it, or the physician who signed off on the system’s use? The FDA’s involvement in the ADVOCATE program suggests that the agency is looking to create a framework for "closed-loop" systems, where the device operates independently within pre-defined clinical parameters.

Furthermore, there is the issue of algorithmic bias. If the AI is trained on data sets that underrepresent minority populations or those with specific comorbidities, the recommendations it provides could exacerbate existing health disparities. ARPA-H, in its project briefings, has emphasized that the design phase must include rigorous validation against diverse patient populations to ensure equitable outcomes.

Analysis: The "Agentic" Future of Medicine

The shift from passive AI to "agentic" AI is the next major frontier in health technology. Traditional AI is a calculator; agentic AI is a coworker. If the ADVOCATE program succeeds, it could serve as a blueprint for the management of other chronic diseases, such as diabetes, hypertension, or chronic obstructive pulmonary disease (COPD).

However, the scale of this challenge should not be underestimated. Medicine is inherently probabilistic, and clinical decision-making often involves nuance that current Large Language Models (LLMs) and predictive algorithms struggle to capture. The integration of these bots into existing Electronic Health Record (EHR) systems—which are notorious for their lack of interoperability—will be a significant technical hurdle for companies like Tempus AI and Atman Health.

Moreover, the human element of care remains a critical variable. Patients with chronic heart failure often require emotional support and complex coordination of care that goes beyond medication management. The fear among some medical professionals is that an over-reliance on AI could lead to the depersonalization of care, potentially eroding the patient-provider relationship.

Looking Ahead

As the first year of the ADVOCATE program progresses, the industry will be watching the milestones closely. The success of these initial awards will likely determine whether Congress continues to increase funding for similar autonomous health initiatives.

For the participating organizations, the stakes are high. Stanford and Duke bring the academic rigor necessary to prove the efficacy of these systems, while Kaiser Permanente offers the large-scale, integrated delivery system required to test the real-world utility of the software. If these entities can demonstrate that AI agents can safely and effectively manage heart failure patients, it would fundamentally change the architecture of American health care, shifting the role of the human physician from a day-to-day manager of chronic conditions to a supervisor of increasingly capable digital systems.

The $62.7 million investment by ARPA-H is, in many ways, a down payment on a larger vision: a health system where the constraints of geography and human labor are bypassed by the precision and availability of high-level artificial intelligence. Whether the technology can overcome the inherent complexities of human biology and the rigid structures of clinical medicine remains the central question of the coming decade. For now, the ADVOCATE program stands as the most aggressive federal effort yet to automate the future of cardiology.