Normalized claim
Care gap chart review time: 50% decrease
reduced care gap chart review time by roughly 50%
UNC Health cuts care gap chart review time ~50% using Microsoft Fabric + AI (EASI) with Copilot for pipeline and notebook workflows
Use case typeDocument processing automationUpdated Jul 20, 2026
UNC Health standardized its data estate on Microsoft Fabric to create a single governed analytics platform supporting clinical AI, population health, and secure research. It built EASI to scan clinical documents, extract exam signals into standardized structures, and integrate findings into Epic with humans in the loop. Analytics teams also use Copilot in Microsoft Fabric to assist with pipeline development and notebook workflows.
- Organization
- UNC Health
- Industry
- Healthcare
- Location
- United States
- Published
- July 2026
Reported outcomes
83-93%
abstraction detection accuracyQuality & accuracy
−50%care gap chart review time9,700 care gapsdiabetic eye exam care gaps reviewed180,000 care gapsadditional care gaps reviewed25 studiesactive studies supported
Strategic outcomes
Innovation & cultureStandardized the enterprise data estate on a single governed analytics platformOther strategic outcomeEnabled a secure research environment for controlled AI and analytics useOther strategic outcomeExpanded care gap closure to additional screening measures
Catalog median for quality & accuracy deployments: +41% across 63 reported metrics. Compare benchmarks →
Why do we believe this?Outcome claims, sources, and evidence checks
Normalized claim
Abstraction detection accuracy: 83-93% increase
accuracy improved from 83% to 93% during the diabetic eye exam pilot
Normalized claim
Diabetic eye exam care gaps reviewed: 9,700 care gaps increase
reviewed approximately 9,700 diabetic eye exam care gaps
Normalized claim
Additional care gaps reviewed: 180,000 care gaps increase
nearly 180,000 additional care gaps across other quality screening measures
Normalized claim
Active studies supported: 25 studies increase
SHIRE supports 25 active studies
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- UNC Health
- Provider
- Microsoft
- Maturity
- Pilot
- Linked source
- Microsoft Customer Stories
Improved abstraction detection accuracy from 83% to 93% during the pilot
Customer identity supportedSource describes one deploymentMaturity supported
Primary read
Use case focus
Showing 3 of 3
- 1Document processing automation
- 2Healthcare analytics
- 3Data platform modernization
- Legacy data infrastructure could not keep up with 40% annual data growth.
- A 2023 warehouse outage disrupted operations.
- Manual chart review of unstructured clinical documents made it difficult to reliably surface care gaps in clinical workflows.
- Microsoft Fabric provided a single governed analytics platform.
- EASI extracted and standardized exam signals from documents for integration into Epic.
- Copilot in Microsoft Fabric assisted with pipeline and notebook development.
- Human abstractors remained the final checkpoint before findings entered the record.
- Reduced diabetic eye exam care gap chart review time by roughly 50%.
- Improved abstraction detection accuracy from 83% to 93% during the pilot.
- Reviewed approximately 9,700 diabetic eye exam care gaps and nearly 180,000 additional care gaps across screening measures.
- SHIRE supports 25 active studies.
Architecture
UNC Health standardized its data estate on Microsoft Fabric, using it as a single governed analytics platform. Copilot in Microsoft Fabric assists with pipeline and notebook workflows. The EASI framework scans clinical documents, extracts exam signals into standardized structures, and integrates findings into Epic while keeping humans in the loop. SHIRE is a secure research environment built on infrastructure within a trusted research enclave framework and enabled by the governed Fabric data foundation.
Sources & evidence1
Evidence: Medium65/100Evidence strength
- Customer explicitly identified
- Deployment status explicitly supported
- Primary source available
- Quantified outcome available
- Technical implementation details available
Type: Customer StoryPublished: Jul 20, 2026Publisher: MicrosoftEvidence: PrimaryConfidence: High
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
Was this useful?
Community
Comments
No published comments yet.