TEFCA and the interoperability imperative: what health systems need to prepare
TEFCA will connect every major health network in the US through a common trust framework, but connectivity without data integrity creates new risks. Health systems that prepare only for technical compliance will find that TEFCA interoperability exposes every upstream data quality problem they have been ignoring.
TEFCA connects networks. It does not fix the data flowing through them.
The Trusted Exchange Framework and Common Agreement, finalized by ONC and managed by the Sequoia Project as the Recognized Coordinating Entity, establishes a national floor for health data exchange. Seven QHINs were live at the start of 2025. eClinicalWorks became the eighth that January, Oracle Health has since been designated, and new candidates keep entering: Metriport was accepted as a candidate QHIN in September 2026. The network of networks is no longer a pilot. It is the floor. The promise is straightforward: any provider, payer, or public health agency connected to a QHIN can query and receive patient data from any other QHIN participant without a custom point-to-point interface.
That promise is real. But it carries a problem that most implementation teams are not discussing openly.
What TEFCA actually requires
TEFCA defines six permitted exchange purposes: treatment, payment, health care operations, public health, government benefits determination, and individual access. Each QHIN must enforce the Common Agreement's privacy, security, and technical requirements on every participant in its network.
The technical layer runs on FHIR R4 and HL7 standards. The trust layer runs on organizational agreements, identity proofing, and audit obligations. QHINs must verify that participants meet baseline security and privacy requirements before granting access.
Notice what is missing from that list: any requirement that the data itself is accurate, complete, timely, or fit for the purpose it will be used for.
The interoperability gap nobody is staffing for
TEFCA solves the connectivity problem. It does not solve the data trust problem.
When a health system receives a TEFCA query response containing a patient's medication list from three different networks, it gets three lists. Those lists may conflict. Medications may be coded differently. Timestamps may reflect when data was entered, not when the medication was prescribed. Provenance metadata, if it exists at all, varies by source system.
Before TEFCA, health systems could blame data silos for incomplete records. After TEFCA, those silos dissolve, and every inconsistency becomes visible. A system deploying clinical AI on TEFCA-sourced data now trains or infers on records assembled from dozens of source systems with no standardized integrity scoring.
This is the gap. TEFCA interoperability gives you more data. It does not give you better data.
Key statistics
{{chart:250}} DTI dimension weights: what matters most for TEFCA-sourced records
Five numbers that frame the TEFCA data trust challenge:
Where TEFCA breaks without a trust layer
Three specific failure modes will emerge as TEFCA adoption scales:
Consent fragmentation. A patient may have consented to data sharing at one institution under one set of terms and opted out of research use at another. TEFCA does not reconcile consent semantics across networks. When that patient's aggregated record lands in an AI training pipeline, consent governance breaks unless each record carries machine-readable consent metadata.
Provenance collapse. TEFCA-sourced records arrive with transport-layer metadata, but chain of custody from the original source system is often incomplete. A lab result that passed through an HIE, a QHIN, and an EHR integration engine has been transformed at least three times. Each transformation is a provenance gap.
Recency mismatch. Exchange latency varies by QHIN and participant. A query may return a medication list that was current when it was cached but is now 90 days stale. Clinical decision support tools that treat TEFCA responses as current state will make errors.
TEFCA in 2026: documents first, FHIR next
TEFCA's first exchange runs on documents, not APIs. The founding QHINs stood up exchange on C-CDA documents; FHIR arrives in stages under the RCE's FHIR Roadmap, from FHIR support inside a single QHIN's network, to QHIN-facilitated FHIR exchange, to direct QHIN-to-QHIN FHIR connectivity, and finally end-to-end FHIR exchange. QHIN-to-QHIN pilots were scheduled for 2025; full end-to-end exchange is targeted for 2026 into 2027.
For data trust, the mixed-format years are the hardest ones. A document and a FHIR resource describe the same encounter in different shapes, with different provenance metadata and different gaps. Validation tooling has to cover both, plus the seams where a document becomes a resource. Anyone building only for the end state will misread the transition. (The version gap inside FHIR itself is a second problem: FHIR R4 vs R5: what changes for data trust scoring.)
Individual access is the purpose to watch
Of TEFCA's six exchange purposes, two carry a required response today: treatment and Individual Access Services, the purpose that lets patients request their own records electronically. Adoption of individual access has lagged, and the community has converged on FHIR-based exchange as the way to enable it.
Watch this purpose because it changes the data mix. Patient-mediated requests pull records across networks at consumer scale, including through apps whose data handling the health system never sees. Every record that arrives this way still needs the same five answers: source, consent, currency, corroboration, validation. Volume makes the trust layer more urgent, not less.
Payers are coming onto the network
The next wave of TEFCA traffic is payer-shaped. Payers told CMS they want document-based exchange phased out in favor of FHIR APIs, with prior authorization as the lead use case, and the 2026 roadmaps name the pieces: prior auth, clinical data exchange, payer data exchange, bulk FHIR, and the Da Vinci accelerator profiles.
Payer data arrives at claims scale with billing semantics: codes chosen for reimbursement, timestamps from adjudication, results often missing. It is the lowest-trust feed most systems will connect, and TEFCA is about to make it a firehose. (Claims data lags 30 to 90 days. Here is what it costs AI models measures one dimension of the problem.)
What health systems should build now
{{chart:27}}%7Breturn%20value%20%2B%20'%25'%7D%22%7D%7D%7D%7D%7D) Data standardization time: before and after DTI scoring (imaware case study)
TEFCA readiness is not just a network onboarding project. It is a data trust infrastructure project.
Health systems preparing for TEFCA-sourced data should implement three capabilities before scaling AI on interoperable records:
The hospital systems already building this infrastructure will be the ones that turn TEFCA from a compliance checkbox into a clinical advantage.
The regulatory trajectory
ONC has signaled that TEFCA participation will eventually become a condition of certification for health IT products. CMS is watching TEFCA adoption as a potential lever for value-based care program requirements. The FDA's emerging AI guidance already expects provenance documentation for training data.
These three regulatory vectors converge on the same requirement: data that moves through interoperable networks must carry verifiable trust metadata. TEFCA provides the pipes. Health systems must provide the trust layer.
SuperTruth scores incoming EHR data at the point of ingestion, before it reaches a model. The DTI Engine evaluates every TEFCA-sourced record across 8 trust dimensions, flagging provenance gaps, consent conflicts, and recency failures before they contaminate clinical AI or trigger audit findings. If your system is deploying clinical AI on interoperable data and needs to answer an auditor's questions about what you trained on, schedule a conversation with the SuperTruth commercial team or (215) 918-4140.
Further reading:
Jason Alan Snyder
Co-founder of SuperTruth and Artists & Robots, and an inventor on the Data Trust Index patents. Twenty-plus years building technology inside Interpublic Group. He writes here nearly every day on data trust, provenance, and what AI should be allowed to act on, and publishes essays on his Substack.
About SuperTruth · LinkedIn · Substack · jasonalansnyder.com
See it in practice
EHR data scored before any AI model sees it.
DTI integrates with Epic, Oracle Health, and all major EHR systems.