NAM Outlines Vision for AI-Enabled National Health Data Infrastructure

The National Academy of Medicine (NAM) has released a perspective paper outlining how a unified, adaptive digital and data architecture could address persistent challenges across the health-care ecosystem. The proposed infrastructure would integrate AI tools and other emerging technologies to enhance patient-centered care coordination and accelerate disease discovery.

The discussion paper is part of a series from the NAM’s Commission on Investment Imperatives for a Healthy Nation. In the perspective paper, an expert working group representing academic, research, and industry stakeholders (and including ASCO AI in Oncology's Editor-in-Chief, Ravi Parikh, MD, MPP, FACP) outlined a cohesive framework to strengthen and unify the U.S. digital health ecosystem.

“Architecture is more than a technical specification—it is the foundation for breaking down silos and unlocking system-wide impact,” stated co-author Peter Lee, PhD, President of Microsoft Research. “It signals a commitment to a clear, adaptable roadmap for the future.” 

The co-authors argue that such infrastructure is essential to advance innovation in disease prevention, diagnosis, treatment, and more. A unified system would enable seamless information exchange and interoperability across the entire health-care ecosystem, replacing siloed systems that limit access to critical health data for both providers and patients.

The authors emphasize that the U.S. health-care system must move beyond electronic health records (EHRs) toward comprehensive data digitization. A fully integrated data infrastructure, they argue, could reduce costs and improve the efficiency and quality of care delivery. To achieve this, they call for a more deregulatory approach focused on interoperability standards rather than just EHR functionality.

“Overcoming the inertia of an industry with limited direct consumer pressure may require the use of financial, regulatory, legal, and cultural interventions to gain the necessary leverage for greater system alignment,” the study authors wrote in the NAM paper. “The authors have identified potential levers—meant to be implemented cohesively rather than as individual interventions— that could mobilize collective efforts and define long-term strategies to realize the envisioned digital and data architecture described in this discussion paper.”

At the core of the proposed framework is a Learning Health System, defined by the NAM as one “in which science, informatics, incentives, and culture are aligned for continuous improvement, innovation, and equity—with best practices and discovery seamlessly embedded in the delivery process, individuals and families active participants in all elements, and new knowledge generated as an integral by-product of the delivery experience.” The NAM has also outlined a set of shared commitments, or a trust framework, to guide the development of a safer, more affordable, effective, and equitable health system.

Building a modern digital health ecosystem capable of effective data exchange also requires several key elements, including the Trusted Exchange Framework and Common Agreement for network exchange, application programming interface standards, data security frameworks, integration protocols, private-public collaboration, compliance with security and privacy regulations, a patient-centered data model, and EHR certification.

The authors also highlight the implementation of AI tools and systems as critical to a continuously learning and improving health system. They call for appropriate regulatory and monitoring frameworks, as well as federated approaches, to guide the development and oversight of AI tools and technologies. At the same time, they emphasize that AI and other emerging technologies are essential to drive innovation, address scalability challenges, and improve care coordination, disease discovery, risk prediction, and clinical decision-making. Restricting their development, rather than supporting safe and ethical integration, could hinder progress. The authors also propose incentives to encourage the responsible deployment of AI tools and expand access to clinical decision support tools for providers.

The authors identify four areas of system misalignment—regulatory complexity, industry fragmentation, misaligned financial incentives, and resistance to innovation—to illustrate how a modern data architecture could drive progress. They also highlight four use cases—cardiovascular disease, maternal/fetal health and maternal mortality, non–small cell lung cancer therapies, and diabetes mellitus—to demonstrate how a comprehensive digital and data infrastructure could improve clinical outcomes and reduce costs.

DISCLOSURES: The perspective paper, which reflects the views of the authors, not their organizations or the NAM, was completed with support from Healing Works Foundation, Doris Duke Foundation, and the Gordon and Betty Moore Foundation. For full disclosures of the study authors, visit nam.edu.  

ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO^®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.