The Future of Health Information Exchanges: From Data Pipelines to Data Refineries

By Leo Pak, CEO – InterstellaSeptember 3, 2025

Health Information Exchanges (HIEs) have been the backbone of healthcare data sharing, connecting electronic medical records (EMRs) to ensure continuity of care across providers. Each day, a regional HIE processes millions of messages such as lab results, prescriptions, discharge summaries, and referrals, totaling nearly a billion transactions annually in some systems. For years, this defined success: HIEs as pipelines, moving data efficiently.

But as healthcare pivots to predictive, equitable, and value-based models, pipelines alone are insufficient. They transport raw data but do not address its flaws. Up to 40 percent of patient records contain inaccuracies, and batch processing introduces delays incompatible with real-time care needs. To thrive, HIEs must evolve into data refineries, AI-native systems that transform raw data into clean, context-rich Intelligent Assets to power better outcomes, lower costs, and health equity.

The Limits of Legacy Pipelines

Traditional HIEs were built for a simpler era, designed to shuttle data between providers. They have excelled at connectivity, but their reliance on batch processing, delivering data hours or days late, falls short of modern demands. Siloed systems miss critical social and behavioral factors, and data errors undermine trust.

The rise of national frameworks like the Trusted Exchange Framework and Common Agreement (TEFCA), Carequality, and CommonWell heightens the urgency. These systems offer low-cost, scalable connectivity, threatening to commoditize HIEs that only move data. Meanwhile, AI-driven care requires zero-latency, high-quality data to predict risks such as heart failure or behavioral health crises before they escalate. Batch processing is the language of the past. Streaming intelligence is the future.

A New Vision: HIEs as Data Refineries

The future of HIEs lies in becoming data refineries, dynamic systems that do not just route data but refine it into high-grade fuel for AI, predictive analytics, and equitable care. Unlike pipelines that transport raw crude, refineries quarantine errors, normalize codes, integrate diverse data sources, and produce actionable insights.

This shift is existential. HIEs that remain pipelines risk becoming utilities, squeezed by cost pressures. Those that become refineries will be indispensable partners, empowering clinicians, patients, payers, and communities.

At Interstella, we are pioneering this transformation with our Lynqsys platform, a Data Refinery as a Service model that processes millions of messages daily into Intelligent Assets. These are self-updating, context-rich datasets ready for AI copilots, patient portals, and decision-support systems. By combining clinical data with social determinants such as housing instability, Lynqsys can flag at-risk patients and trigger timely interventions, reducing preventable emergency visits and improving outcomes.

The Pillars of Transformation

1. Real-Time, AI-Native Processing

Batch pipelines are relics in an era where AI-driven care demands immediacy. HIEs must leverage AI and standards like FHIR (Fast Healthcare Interoperability Resources), set to be enforceable by 2027, to process data in real time.

AI tools such as natural language processing (NLP) and machine learning can clean errors, standardize unstructured clinical notes, and address the 40 percent inaccuracy rate in patient records. This enables predictive models to act before crises occur.

• Example: AI-driven tools can flag subtle patterns across vitals, medications, and social factors to predict heart failure weeks in advance.
  

• Impact: Real-time analytics reduce redundancies, enhance clinical decisions, and align with value-based care’s focus on outcomes.
  

2. Whole-Person Intelligence

Health is more than medical records. HIEs must integrate clinical data with mental health, social determinants of health (SDOH), behavioral insights, and even genomic information to create holistic patient profiles. At InterStella we call this Whole Person Intelligence. Legacy systems, blind to factors like housing instability or untreated anxiety, miss the drivers of health outcomes. By partnering with community organizations, HIEs can fuse these data domains, ensuring no patient’s story is half-written.

• Example: Our partnership with NinePatch shows how integrating clinical and SDOH data can alert navigators to eviction risks, preventing crises like emergency room visits for patients whose anxiety and financial stress exacerbate asthma.
  

• Impact: Whole-person care fosters health equity, addressing disparities and improving outcomes for underserved populations.
  

3. Sustainable Innovation

Grant dependency has led to failures like California’s Cal eConnect, which collapsed after $38 million in funding dried up. With only 34 percent of HIEs currently self-sustaining, the future demands revenue-generating models.

HIEs can create reusable data assets, curated cohorts for conditions such as congestive heart failure, that serve hospitals, payers, public health agencies, and community organizations. By offering Data Refinery as a Service, HIEs can build recurring value while fostering innovation through technology partnerships.

• Example: A single curated cohort can reduce hospital readmissions, manage payer costs, track public health trends, and align community services.
  

• Impact: Sustainability ensures HIEs thrive in a competitive landscape, moving beyond the grant trap to deliver lasting value.
  

A Roadmap for All HIEs

This vision is ambitious but achievable, even for smaller or resource-constrained HIEs.

• Immediate Steps (1–2 Years): Pilot AI-driven data refinement using NLP and FHIR APIs. Begin integrating mental health and SDOH data through community partnerships. Monetize data assets for specific use cases such as reducing readmissions.
  

• Mid-Term Goals (3–5 Years): Scale to Health Data Utilities with nationwide interoperability via TEFCA. Use blockchain for data security and expand revenue models to diverse stakeholders.
  

• Long-Term Vision (Beyond 5 Years): Build AI-orchestrated ecosystems that anticipate global health needs, seamlessly integrating clinical and non-clinical data for holistic insights.
  

For rural or underserved HIEs, transitional support through grants, shared infrastructure, or partnerships with technology leaders like InterStella can bridge the gap and ensure inclusivity.

Overcoming Challenges

The path to data refineries faces hurdles that must be addressed.

• Data Quality: Persistent silos and errors require ongoing AI and standardization efforts.
  

• Ethical AI: Addressing bias, transparency, and privacy concerns is critical to maintain trust.
  

• Inclusivity: Smaller HIEs need tailored support to adopt refinery models without sacrificing current connectivity.
  

• Funding: Shifting to revenue-based models demands upfront investment, which collaborative ecosystems and public-private partnerships can help offset.
  

The Call to Action

As the healthcare industry gathers at Civitas in Anaheim this September, the conversation will focus on sustainability, innovation, and equity. The message is clear: HIEs must evolve or risk obsolescence.

Data refineries are not a distant dream. They are the present, as platforms like Lynqsys demonstrate. By embracing AI-native processing, Whole Person Intelligence, and sustainable models, HIEs can lead healthcare into a data-advised era where data does not just move. It anticipates, empowers, and heals.

The choice is stark: cling to pipelines and face commoditization, or embrace refineries and shape the future. At InterStella, we are building that future today, solving the threat of dirty data and unlocking the promise of intelligent, equitable care.

Let’s talk at Civitas 2025 in Anaheim. Contact me at leo.pak@interstella.us to schedule a conversation.