The Advanced Research Projects Agency for Health (ARPA-H) released Amendment 02 to its ADVOCATE solicitation (ARPA-H-SOL-26-142), refining technical expectations, evaluation criteria, and program execution details for one of its most ambitious artificial intelligence initiatives in healthcare. ADVOCATE is designed to accelerate the development, evaluation, and scalable deployment of clinical artificial intelligence (AI) agents that can safely support care teams in real-world environments. The overarching goal is to ensure that AI-enabled clinical decision support tools are accurate, reliable, interoperable, and capable of operating at scale across diverse health systems.
Amendment 02 provides additional clarity on TA1’s narrowed focus on heart failure care management, updated submission rules across tracks, ARPA-H’s limited use of LLMs in the review process, and the heightened emphasis on robust, scalable automation as a primary evaluation criterion for TA1 and TA2.
Overview of Key Updates
Technical Area 1
Technical Area 1 now aims to develop a clinical tool that provides care management to patients with heart failure. Requirements around care management tools surrounding chronic CVD and post-myocardial infraction have been removed from TA1.
Submission Requirements
Proposers must submit one unified solution summary. Teams intending to apply to both TA1 and TA2 must submit separate, technically independent solution summaries for each track. A proposer may submit one summary for TA1 and a separate summary for TA2. However, proposers submitting a solution summary for TA3 may not submit to TA1 or TA2 in any capacity.
Submission Review Process
ARPA-H will review and respond to all proposers who submit compliant solution summaries. Large Language Models (LLMs) may now be used to extract data elements, summarize content, and generate draft text to support the review process. However, LLMs will not replace human judgment, and all original documents will undergo human review.
Review Criteria: Overall Scientific and Technical Merit
ARPA-H still has an emphasis on the proposed technical approach being innovative, feasible, achievable, and comprehensive. However, For TA1 and TA2, a high degree of automation is a primary evaluation consideration. Proposals that demonstrate robust, scalable, and thoughtfully designed automation throughout the technical approach will be viewed more favorably. Automation should not be incidental, but rather integral to system design, performance, and operational scalability.
Key Dates
Key milestones include:
- February 27, 2026: Solution Summary due by 5:00 PM Eastern Standard Time
- April 1, 2026: Full Proposal due by 5:00 PM Eastern Standard Time
Awards are expected to support multi-year efforts aligned with phased technical milestones. Continuation is not automatic; performers must meet defined performance metrics to advance. ARPA-H’s phased funding model reinforces a milestone-driven approach tied directly to technical validation and scalability readiness.
Proposers should design budgets that reflect staged risk reduction, integration complexity, regulatory engagement, and commercialization planning.
If your company has considered applying for federal funding, your federal funding journey starts here. EverGlade is a national firm that helps organizations win and manage federal awards. We offer services ranging from Pursuit, Proposal and Post-Award support to comply with federal regulations at agencies including BARDA, ARPA-H, NIH, DTRA, JPEO, DOD, DIU, DOE, and DARPA.
