iQHD 2.0 and FHIR·A: quality that flows in both directions.
Clinical documents flow through exchange networks at volume, but almost nobody can answer with precision: how complete are the documents from Hospital X? Are coded diagnoses conformant? Are medication lists populated? Are allergies documented consistently? Data quality is trusted blindly, and where it is measured, the measurement dies in a report.
Worse, the flow is one-directional. The source system that produced an incomplete document never hears about it, so the same defect arrives again tomorrow, and the network's quality never compounds.
iQHD 2.0 expands scoring from today's dimensions toward 120 quality dimensions across five families: presence, conformance, recency, semantic richness and provenance integrity. AI enrichment tools repair, reconcile and complete records in flight. And a proposed standard, FHIR·A, extends FHIR R4 with provenance-first, confidence-scored, agent-readable representations optimised for the AI systems that will increasingly consume this data.
Each CCD and FHIR bundle is assessed across the full iQHD dimension set as it traverses the Gateway, with no added latency budget for the requester.
Quality findings return to contributing systems as standardised FHIR resources: which sections were missing, which codes non-conformant, which timelines broken.
Where repair is safe and provenance-preserving, enrichment tools normalise versions, deduplicate entries and flag conflicts for human review.
Confidence-scored, provenance-first representations give AI agents what raw FHIR does not: an explicit statement of how much each element should be trusted, and why.
Source systems improve their own output based on Gateway-derived signals; every quarter's documents arrive better than the last.
Quality reporting, contributor benchmarking and trend analytics are delivered as assets the HIE owns and can take to its board and its participants.
Clinicians who query the HIE and receive high-quality data keep using it, reversing the disengagement spiral incomplete data creates.
When intelligent agents read records at scale, the network that ships confidence-scored data becomes the network they prefer.
Every document already flows through the Gateway's 47-dimension IQHD engine, and quality scorecards already ship in the HIE dashboard. Intelligent Quality is the natural next release: closing the loop back to sources, expanding the dimension set, and standardising the agent-facing representation.