Standards and APIs can move information. True healthcare interoperability requires the governance, trust and workflow design that make the information usable.
Healthcare has spent years treating interoperability as a problem of moving data. National exchange has expanded, standards have matured, and APIs have improved access to information. However, this availability alone doesn’t make information useful.
We have moved beyond the question of whether data can cross technical boundaries. The question now is whether dependable information reaches the right person, in the right workflow, in time to change what happens next.
E-prescribing remains the clearest example of healthcare interoperability at scale. It connects prescribers, electronic health records, pharmacies, pharmacy benefit managers and payers without leaving the patient or each organization to reconcile the process independently. A common standard was necessary: NCPDP SCRIPT became the foundation for prescription and related transactions under Medicare Part D. But standardization alone was insufficient. Success also required broad participation, regulatory reinforcement, conformance testing and integration into the daily work of prescribing and dispensing.[1]
The Trusted Exchange Framework and Common Agreement (TEFCA) represents important progress toward a nationwide US framework for trusted health information exchange. However, exchange volume does not, by itself, equal clinical or operational value. In 2023, 71% of US hospitals reported routinely having necessary outside clinical information electronically available at the point of care, while only 42% reported that clinicians often used it.[2] This gap provides a better measure of the unmet need than simply the number of records exchanged.
This is why debates that pit standards-based exchange against APIs are misplaced. Standards and APIs address different parts of the same problem. Standards establish common structure and meaning, while APIs provide practical methods for retrieval and use.
Mature standards such as HL7 v2, X12 and NCPDP should remain when they perform reliably. Replacing a functioning transaction standard merely because a newer technology exists creates cost without necessarily creating value. FHIR-based APIs are better suited to selective, timely access and newer use cases, including patient access, payer-to-payer exchange and electronic prior authorization. An API is not itself a healthcare interoperability strategy.
The harder work involves implementation guides, terminology, authentication, version control and conformance testing. An API cannot repair poor data or inconsistent definitions. It simply moves the ambiguity faster.
Data sharing succeeds only when the recipient can trust what arrives. A receiving organization must know that information belongs to the correct patient, comes from an identifiable source, is current and can safely support a decision. Structurally complete data may still be stale, duplicated, miscoded or clinically misleading.
Governance cannot be implemented after the interface is built. It requires clear ownership of data definitions, provenance, patient-matching controls, quality thresholds and a process for correcting errors after information has crossed organizational boundaries. The source and accountable steward of a laboratory value, medication list or authorization status must be identifiable before that information is used at scale.
These issues become more consequential in care coordination and value-based care. A clinician coordinating an episode needs timely, event-driven information: admission, discharge, result, medication change, referral or authorization decision. Population health teams need a broader longitudinal view combining clinical, claims, encounter and community-based information across an attributed population.
The blind spots remain substantial. In 2023, only 16% of hospitals reported sending summary-of-care records to most or all long-term or post-acute care providers, and only 17% did so for behavioral health providers.[2] Claims data may arrive too late for an immediate intervention, while clinical feeds may lack the completeness needed to understand utilization and risk. Closing those gaps requires an architecture matched to the use case, not simply another connection.
Patients belong inside this model as active participants, not couriers for a fragmented system. In 2024, 59% of individuals reported having multiple online medical records or patient portals, yet only 7% used an app that combined information from different portals into one place.[3] Access has expanded faster than integration.
Patients should be able to direct where information goes, authorize caregiver access and correct errors without navigating multiple portals. Secure patient-mediated exchange requires reliable identity verification, understandable permissions, auditability and revocation. Patient-generated information also needs a managed clinical destination. An unattended inbox receiving readings, documents or messages is just another broken handoff.
True interoperability reduces the reconciliation healthcare asks people to perform. It does not simply transfer that burden from one organization to another, from the system to the clinician, or from the clinician to the patient.
Healthcare has built much of the infrastructure required for information to move. The remaining work is operational: make the information dependable, put it where decisions are made, and assign accountability for what happens next. That is when exchange becomes interoperability.
Author Bio
