The Future of Health Information Exchanges: Transitioning from Data Movement to Intelligent Insights

Understanding the Evolution of Health Information Exchanges

Health Information Exchanges (HIEs) were born out of a simple promise: to make data move. For the past two decades, this promise has shaped everything from regional networks to national frameworks. TEFCA, the Trusted Exchange Framework and Common Agreement, was designed to hardwire the pipes, ensuring data could flow between systems regardless of vendor or geography. Versions 1.0 and 2.0 delivered that baseline — agreements were signed, QHINs were certified, and interoperability in its narrowest sense was achieved.

But the limits of this model are now impossible to ignore. TEFCA today guarantees transactions, not trust. It ensures that messages move, but it does not ensure they are accurate, complete, or usable. The numbers are sobering: nearly 40 percent of patient records contain errors or duplications. Batch processing still delays information by hours or days, undermining the promise of real-time care. While the industry celebrates transactions, the deeper forces that drive health — social determinants, mental health, and behavioral signals — remain fragmented, ungoverned, and often invisible.

The Consequences of Fragmented Data

The result is a national switchboard. Data moves, but meaning is lost. As artificial intelligence takes center stage in healthcare, this gap becomes dangerous. AI cannot thrive on dirty data; it magnifies every flaw. A duplicate record, a missing demographic, or a mismatched code — each becomes a crack in the foundation of predictive care. Dirty data is not just inefficient; it is a threat.

That is why TEFCA must evolve. The next phase — call it TEFCA 3.0 — must be about transformation, not transportation. If 1.0 and 2.0 were about building pipelines, 3.0 must be about building refineries.

The Role of Refineries in Healthcare Data

A refinery does more than move crude; it transforms it into high-grade fuel. In healthcare, that means quarantining errors before they poison the system. It means normalizing codes so data is interoperable in practice, not just in theory. It means assigning trust scores so contributors are accountable for the quality of their inputs. It means integrating clinical data with social and behavioral signals so a patient’s story is complete, not half-told. Above all, it means preparing data for the AI era, where algorithms depend on clean, context-rich inputs to anticipate rather than react.

Learning from Past Mistakes

The consequences of ignoring this shift are clear. History is littered with HIEs that mistook movement for value. They chased grants, stood up shiny new features, and collapsed as soon as the funding dried up. California’s Cal eConnect burned through $38 million before shutting down in less than two years. POLST registries were heralded as breakthroughs, but most were built as bolt-ons, disconnected from the broader data fabric, and quickly faded.

Today, many exchanges are making the same mistake with SDOH, treating it as a feature to be bolted on rather than a data domain that must be curated, governed, and integrated. These are iron horses — faster versions of what already exists — when what we need is the automobile.

Defining the Future of HIEs

The future HIE cannot be defined by how many messages it moves. Carequality and CommonWell can already do that, and at lower cost. If HIEs define themselves only by keeping pace, they will be replaced by utilities. Their only path forward is to create value above the pipe — intelligence that utilities cannot quickly replicate. That value comes from Intelligent Assets: datasets that have been refined until they are reusable, reliable, and revenue-generating.

Consider one curated cohort of congestive heart failure patients. Hospitals can use it to cut readmissions. Health plans can use it to manage costs. Public health agencies can use it to monitor population trends. Community organizations can use it to align services. One asset, four revenue streams. That is sustainability — not the fragile survival of chasing subsidies, but the durable growth of producing reusable, monetizable assets.

The Vision for TEFCA 3.0

TEFCA 3.0 is the opportunity to make this shift nationally. To set minimum data requirements so messages meet a quality threshold before entering the exchange. To adopt quarantine and replay logic so bad data is held back until corrected. To publish trust scores so contributors are accountable and transparency drives improvement. To expand the scope beyond clinical transactions, treating social and mental health data as first-class citizens in the data fabric. And to frame the entire enterprise as AI-readiness, not compliance.

The choice is stark. TEFCA can remain a switchboard — technically impressive, but clinically irrelevant — or it can become the refinery standard, turning raw transactions into assets that drive predictive care, health equity, and sustainability.

Conclusion: Embracing the Refinery Era

The pipeline era is ending. The refinery era has begun. At Interstella, we believe dirty data is the real threat to AI. TEFCA 3.0 is the chance to fix it — not by building faster horses, but by building the automobile healthcare truly needs.

In this new landscape, we must prioritize the quality of our data. Only then can we harness the full potential of technology to improve health outcomes and drive meaningful change. The future is bright, and the possibilities are endless. Let's embrace this transformation together.