Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Emerging Agent Communication Protocols for Healthcare AI Systems: A Scoping Review of Model Context Protocol (MCP) and Agent-to-Agent (A2A) Frameworks
ABSTRACT
Background:
As artificial intelligence systems evolve toward agentic architectures capable of autonomous reasoning and tool use, two interoperability protocols have emerged to govern how AI agents communicate with external systems and with each other: the Model Context Protocol (MCP) and the Agent-to-Agent (A2A) framework. Despite growing adoption in healthcare, no systematic synthesis of these implementations exists.
Objective:
This scoping review maps and characterizes healthcare implementations of MCP and A2A frameworks, examining their technical architectures, application domains, integration patterns, and evidence maturity.
Methods:
Following the PRISMA-ScR guidelines, we searched PubMed, IEEE Xplore, Web of Science, Scopus, and grey literature sources for studies implementing MCP or A2A in healthcare contexts. Two reviewers independently screened, extracted data, and appraised study quality using a modified Mixed Methods Appraisal Tool. The review protocol was registered with PROSPERO (CRD420251271558).
Results:
Thirteen studies met inclusion criteria, all published between November 2024 and early 2026. MCP or MCP-aligned implementations dominated (n=10, 76.9%), while two studies (15.4%) employed hybrid MCP/A2A architectures and one (7.7%) used A2A alone. Applications spanned clinical decision support (n=5), medical imaging and radiology (n=2), patient interaction and scheduling (n=1), oncology (n=1), federated digital health (n=1), cybersecurity (n=1), neurodevelopment (n=1), and fitness/well-being (n=1). Quality appraisal identified six high-quality studies (46.2%) with rigorous evaluations, one moderate study (7.7%), and six low-quality studies (46.2%) presenting primarily conceptual or prototype-stage work. No study reported prospective clinical validation with real patients.
Conclusions:
MCP has emerged as the dominant protocol for structuring agent-tool interactions in healthcare AI, enabling auditable, grounded workflows that reduce hallucinations in narrowly defined tasks. A2A remains relevant primarily for cross-institutional or federated scenarios. Current evidence supports deployment in assistive, backend applications rather than autonomous clinical decision-making. Critical gaps persist in agent identity verification, prospective clinical validation, and health equity considerations.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.