How AI, Cloud & Smart Curation Are Transforming Federal Health Records

June 10, 2025

The Federal Electronic Health Record Modernization (FEHRM) program is undergoing a transformative shift in how it handles health data across the Department of Defense and the Department of Veterans Affairs. FEHRM Chief Technology Officer Lance Scott outlines the three pillars of this change: deduplication, data provenance, and auto-ingestion. Each plays a critical role in building a more efficient, responsive, and clinically useful health record system.

Deduplication is a silent hero, Scott explains. Pre-deployment testing reveals that up to 96% of duplicate data can be removed—far more than initially expected. This has profound implications, not only reducing storage and hardware costs, but also streamlining clinician access to clean, relevant data. Less noise, more signal.

Data provenance addresses a long-standing challenge in health data management: tracing the origin of medical records. Scott highlights how the FEHRM program can now identify the source of data entries—for example, pinpointing that a brain scan originated from a specific clinician in South Dakota. This capability ensures accountability and helps clinicians trust the data they use to make critical decisions.

The third—and Scott’s personal favorite—pillar is auto-ingestion. When a clinician accesses a patient’s record, the system pulls data not just from VA and DoD sources, but from over 250,000 external partners. If the source and domain are trusted, the system automatically ingests the data into the record, removing another burden from clinicians and putting more useful information at their fingertips.

The success of a pilot at the VA facility in Walla Walla last November has prompted a broader rollout, scheduled to go enterprise-wide this fall. The DoD is preparing to follow suit. However, Scott notes that a key ongoing challenge will be collaborating with clinicians—“the functionals”—to define what data is trustworthy enough for automatic inclusion.

Cloud migration is another critical enabler. Drawing from his experience with the Defense Medical Information Systems program, Scott recalls how moving to the cloud eliminated costly, static hosting environments in favor of dynamic, scalable infrastructure. The upcoming transition to Oracle Cloud will move 3–4 petabytes of active health data—cutting costs and improving flexibility.

However, the real evolution lies not in collecting more data, but in using it smarter. With 96% of the U.S. health market already integrated, the focus shifts to data curation. Scott discusses how the program uses natural language processing to identify “exclusionary concepts” like asthma, helping programs such as Military Entrance Processing (MEPCOM) quickly screen candidates and begin waiver processes. This alone has reduced their workload by 80%.

Artificial intelligence is playing an increasing role in these innovations. Scott previews Oracle’s next-generation EHR, featuring an AI assistant capable of ambient dictation and intelligent querying—tools that can drastically reduce clinicians’ after-hours “pajama time.”

In an environment filled with shiny new tech, Scott emphasizes a grounded approach: focus on the problem, not the solution. By listening closely to users and staying anchored in real needs, FEHRM is engineering a smarter, more responsive future for federal health care.