The Advanced Research Projects Agency for Health (ARPA-H), a federal agency established to catalyze high-risk, high-reward biomedical research, has officially launched its ambitious ADVOCATE program with a $62.7 million funding commitment. The initiative seeks to develop autonomous artificial intelligence systems capable of managing the complex, long-term treatment of heart failure. This move marks a significant shift in federal health policy, moving beyond research-focused AI toward the development of clinical-grade "agents" that could eventually operate with a degree of independence in prescribing medication and ordering diagnostic laboratory tests.
The program represents a direct response to a burgeoning public health crisis. Heart failure affects approximately 6.7 million Americans, a number projected to rise significantly as the population ages. Many of these patients, particularly those in rural or socioeconomically disadvantaged areas, face persistent barriers to care, including a severe shortage of cardiologists and the logistical burden of frequent clinic visits. By empowering AI to fill these gaps, ARPA-H aims to provide a scalable solution to a condition that remains one of the leading causes of hospitalization and mortality in the United States.
The ADVOCATE Program: Scope and Objectives
The ADVOCATE (Advancing Clinical Decision-Making and Optimized Treatment) initiative is designed to transition AI from a supportive tool used by physicians to a partially autonomous participant in patient management. According to the agency’s initial project brief, the core goal is to produce AI devices that obtain regulatory clearance from the Food and Drug Administration (FDA) to perform three critical functions: assessing the severity of patient symptoms, prescribing evidence-based medications, and ordering necessary lab tests to monitor patient health.
The program is structured in phases. ARPA-H has announced the first batch of awards totaling $33.7 million to a diverse group of stakeholders, including private health technology firms Atman Health, UpDoc, and Tempus AI, alongside academic and clinical powerhouses Stanford University, Duke University, and the Kaiser Permanente health system. The remaining $29 million of the total $62.7 million commitment remains subject to project performance benchmarks and may be renegotiated based on the progress of the initial prototypes.
Chronology and Strategic Development
The genesis of the ADVOCATE program follows a broader push within the Biden-Harris administration to modernize health care delivery through digital innovation. ARPA-H, which operates under the Department of Health and Human Services (HHS), was modeled after the Defense Advanced Research Projects Agency (DARPA). Its mandate is to cut through bureaucratic inertia to accelerate breakthroughs in medical technology.
In early 2023, ARPA-H began evaluating the feasibility of "autonomous clinical agents." Industry experts and health systems identified heart failure as an ideal pilot case because it is a "data-heavy" condition; it requires constant monitoring of weight, blood pressure, medication adherence, and biomarkers such as NT-proBNP levels. By late 2023, the agency issued a solicitation for the ADVOCATE program, seeking teams capable of building systems that do not merely suggest treatments to doctors, but can actively manage a patient’s routine.
The announcement of these awards this week signifies the start of the "Build Phase," where the selected consortia will begin training models on massive datasets of heart failure outcomes. The agencies anticipate a multi-year development cycle, with rigorous clinical validation required before any system can be submitted for FDA approval.
Supporting Data: The Heart Failure Crisis
The urgency behind the ARPA-H investment is underscored by sobering national health statistics. Heart failure is not a single disease but a syndrome resulting from any structural or functional cardiac disorder. According to the American Heart Association (AHA), the total cost of heart failure in the United States is expected to reach $70 billion by 2030.
Current treatment models rely heavily on the "guideline-directed medical therapy" (GDMT) approach, which involves a complex cocktail of beta-blockers, ACE inhibitors, and SGLT2 inhibitors. Adhering to these protocols is notoriously difficult. A 2022 study published in the Journal of the American College of Cardiology noted that fewer than 25% of eligible heart failure patients receive the full dose of recommended therapies, often due to a lack of frequent, expert titration of medications.
The ADVOCATE program hopes to rectify this "therapeutic inertia." By utilizing AI that can monitor a patient’s daily health data—often transmitted via wearables or home monitoring equipment—the system could trigger automatic medication adjustments, thereby keeping patients within therapeutic ranges and preventing the common cycle of "revolving door" hospital readmissions.
Stakeholder Perspectives and Institutional Involvement
The inclusion of major health systems like Kaiser Permanente and academic institutions like Duke and Stanford suggests a focus on real-world clinical implementation rather than theoretical model building. Kaiser Permanente, in particular, offers a unique environment for the development of these tools, as it possesses the longitudinal, integrated electronic health records (EHR) necessary to train sophisticated AI models.
Industry reactions have been cautiously optimistic. Tech-forward health companies like Tempus AI have long argued that clinical data, when combined with machine learning, can democratize specialized care. However, some clinical advocacy groups have raised questions regarding liability. If an AI agent prescribes an incorrect dose of a medication that leads to an adverse cardiac event, the legal and ethical framework for responsibility remains underdeveloped.
ARPA-H officials have signaled that the program will include a strong "safety and ethics" component. The design of the ADVOCATE agents will likely involve a "human-in-the-loop" architecture, where the AI manages routine adjustments but escalates complex or ambiguous cases to human cardiologists, ensuring that the technology acts as a force multiplier for human expertise rather than a replacement.
Broader Implications for AI in Medicine
The ARPA-H commitment is a bellwether for the future of the medical profession. If successful, the ADVOCATE program could set a precedent for how the FDA regulates autonomous clinical software. Traditionally, the FDA has focused on diagnostic AI—tools that read X-rays or identify skin lesions. Moving into the space of "therapeutic AI" (systems that actively treat patients) represents a paradigm shift.
There are significant implications for the healthcare workforce as well. If AI can effectively handle the titration of heart failure medications, the role of primary care physicians and nurse practitioners may shift toward more complex, empathetic, and holistic patient management. Conversely, critics warn that delegating medical decisions to algorithms could exacerbate existing biases in healthcare. If the training data for these models is derived from populations that do not reflect the diversity of the American heart failure patient base, the AI may perform poorly for women or people of color, potentially widening the existing health equity gap.
Furthermore, the integration of these systems into existing EHRs—which are often siloed and incompatible—remains a major technical hurdle. The success of the ADVOCATE program will depend not only on the efficacy of the algorithms themselves but on the ability of these systems to "speak" to the software already used by hospitals across the country.
Looking Ahead: The Path to FDA Clearance
As the program moves forward, the FDA will be tasked with developing new regulatory pathways for autonomous therapeutic agents. This will likely involve a combination of clinical trials and "real-world evidence" monitoring. The agency has already begun exploring the "Change Control Plan" framework, which allows software to update its algorithms while in use, provided those updates stay within a pre-approved scope.
For the patients currently navigating the difficult journey of heart failure management, the ADVOCATE program represents a beacon of hope for better, more consistent outcomes. While the journey from a government grant to a bedside clinical tool is long and fraught with regulatory and technical challenges, the $62.7 million investment marks a definitive step toward a future where artificial intelligence is a standard component of the cardiac care toolkit. The coming years will be critical in determining whether these digital agents can truly bridge the gap between clinical guidelines and real-world practice, ultimately helping millions of Americans live longer, healthier lives.
