FHIR at a Crossroads for Real-World Evidence

The Electronic Health Record Association’s June 23 comment letter arrives just weeks after the Food and Drug Administration opened Docket No. FDA-2025-N-0287 to explore whether HL7 FHIR can carry real-world data into formal submissions. Representing almost thirty vendors, the EHR Association applauds the inquiry yet warns that inconsistent coding practices, limited lineage metadata, and opaque de-identification workflows still undermine research-grade reliability. FDA’s docket signals a strategic shift from viewing FHIR solely as a messaging format to considering it a regulatory substrate that could align clinical care documentation with pharmaceutical evidence pathways. If successful, the initiative would compress the feedback loop between bedside observations and safety analytics. If unresolved technical gaps persist, sponsors may be forced to maintain parallel data infrastructures that inflate cost and delay innovation. By placing the burden of proof on pilot demonstrations, the agency has effectively turned standards theory into an operational stress test. (federalregister.gov, ehrablog.org)

Lineage gaps and semantic drift
EHR developers underscore that a transport standard cannot compensate for data entered without shared semantics. A 2022 JAMIA study found a forty-one percent mismatch between manufacturer-recommended and locally used LOINC codes, underscoring entrenched variability even inside high-performing institutions. Aggregating those discordant elements across registries and claims systems magnifies bias when tokenization obscures provenance. The association therefore urges FDA to ground its effort in tightly scoped pilots that embed minimum metadata payloads, actor roles, and audit trails, drawing on profiles already drafted by the HL7 Vulcan Accelerator. Without such scaffolding, sponsors risk submitting analytic tables that lack verifiable ties to source systems, a deficiency that would erode confidence in real-world evidence. Trade analysts speaking to RAPS Regulatory Focus describe the docket as a necessary first step toward modernizing FDA’s antiquated file-format catalog, yet they acknowledge the technical debt that must be retired before routine FHIR submissions become feasible. (academic.oup.com, confluence.hl7.org, raps.org)

TEFCA offers a governance on-ramp
The association supports adding “research” to the list of authorized exchange purposes under the Trusted Exchange Framework and Common Agreement, arguing that a national governance layer can unify consent, security, and audit requirements now splintered across institutional review boards. ONC’s recent standard operating procedure describes how new purposes will be phased in, suggesting a clear policy path. Integrating research into TEFCA would let sponsors query de-identified cohorts through Qualified Health Information Networks while preserving patient preferences at network scale. The approach, however, multiplies operational complexity because QHINs would have to validate tokenization services that reconcile identities across payers, EHRs, and registries. Smaller hospitals could struggle to meet those thresholds, risking further concentration of study populations in large academic systems. State Medicaid agencies that rely on TEFCA for claims adjudication might also hesitate to add research traffic without additional federal funding. The timing of TEFCA expansion will therefore influence how quickly FHIR-based real-world evidence gains traction. (healthit.gov, healthit.gov)

USCDI scope and standards pressure
USCDI Version 3 defines the statutory floor for nationwide interoperability, yet its data classes omit pharmacogenomics, imaging metadata, and device identifiers that regulators consider vital for high-resolution safety monitoring. ONC’s USCDI portal and accompanying Standards Bulletin invite public comment on future expansions, but the EHR Association cautions that overloading USCDI could destabilize certification timelines. Instead, the group proposes a purpose-built implementation guide that maps required research elements to existing classes, allowing incremental adoption without overwhelming smaller vendors. This strategy aligns with ONC’s risk-based HTI-1 rule, which presumes documentation quality will improve only when clinicians perceive direct value. Coupling pilot incentives, such as expedited regulatory review, with measurable provenance requirements could create that value, motivating structured data entry at the point of care. The alternative is a proliferation of bespoke extensions that fragment the very evidence ecosystem the FDA hopes to streamline. (healthit.gov, healthit.gov, healthit.gov)

Strategic stakes for evidence stakeholders
FDA’s docket has moved the FHIR conversation from aspirational interoperability to operational accountability. If pilots demonstrate that provenance-rich FHIR pipelines can meet regulator expectations, pharmaceutical sponsors will gain a faster route from observational insight to labeling change, and health-system data assets will appreciate in strategic value. If the effort stalls, the divide between care documentation and research evidence will widen, forcing parallel infrastructures that inflate cost and slow innovation. Legal experts at Covington Digital Health note that the agency has signaled willingness to expand its data standards catalog if industry proves the model, a call to action for developers who wish to shape rather than follow federal policy. The next year will therefore test whether a standard forged for clinic-to-clinic messaging can evolve into a regulator-grade evidence platform. Stakeholders investing now in lineage capture, vocabulary harmonization, and TEFCA readiness will help decide that outcome, turning policy exploration into pragmatic data stewardship. (covingtondigitalhealth.com)