ONC EHIgnite Challenge · Phase 1 Winner
A U.S. Department of Health and Human Services challenge, run by the Office of the National Coordinator for Health Information Technology (ONC)
SHIRLe Health: AI Interpretation of Electronic Health Information
From Raw EHI Exports to Readable, Usable, and Actionable Patient Intelligence
Precision Health Ventures — with its SHIRLe Health solution — is one of nine Phase 1 winners of the ONC EHIgnite Challenge, selected from nearly one hundred proposals. SHIRLe is an AI interpretation platform that transforms Electronic Health Information (EHI) exports into understandable patient health narratives. It ingests raw FHIR R4 exports from any certified EHR via SMART App Launch 2.0, no bilateral vendor agreements required. The integrated HealthDigits platform combines clinical records with patient-generated signals from wearables, home monitoring devices, and screenshot-derived data.
The work now advances to Phase 2: turning the concept into a real, usable tool that helps patients make sense of complex health records, concluding March 2027.
Read the ONC announcementHow SHIRLe Makes EHI Readable, Usable, and Actionable
Readable
SHIRLe’s AI ingestion pipeline normalizes raw FHIR bundles into a plain-language Health Summary. Clinical codes (SNOMED, LOINC, RxNorm) resolve to human-readable equivalents. Lab trends are visualized. Medication lists are de-duplicated and contextualized. The output reads like a letter from a knowledgeable care coordinator.
Usable
The SHIRLe interface is built around patient mental models, not clinical workflows. Patients navigate their record by condition, life chapter, or care team. An integrated natural language Q&A layer allows patients to ask questions about their own data and receive grounded, source-cited answers drawn exclusively from their own EHI export.
Actionable
SHIRLe surfaces care gaps, upcoming milestones, and medication adherence signals within the patient summary. When behavioral data reveals that glucose spikes correlate with specific dietary patterns, or that blood pressure improves with better sleep, SHIRLe surfaces those insights with suggested conversation starters for the patient’s next provider visit.
Patient-controlled
SHIRLe is patient-sovereign. Patients own their data fabric and grant granular consent for providers, payers, researchers, or wellness platforms to access specific slices of their record, and revoke that access at any time. This is not a provider-controlled portal; it is a patient-controlled health intelligence layer.
SHIRLe AI Health Timeline Engine
Data Input Layer
- EHI ExportsFHIR / JSON · C-CDA
- Wearable DataBP / CGM / HR · Sleep / Weight
- Screen CapturesConsumer apps · OCR extraction
Data Ingestion & Normalization Layer
Parsing · Validation · Standardization
Unified Patient Event Store
Diagnoses · Labs · Meds · Vitals · Signals
AI Interpretation Engine
Trend Detection · Correlations · Summary
Patient Experience Layer
Timeline View · AI Q&A · Visit Summary
- Secure & Compliant
- Interoperable by Design
- AI-Powered Insights
- Patient-Centric Experience
- Real-Time Value
Demonstrated Capability
HealthDigits has demonstrated extraction of structured, FHIR R4-compliant health observations from consumer device screenshots across two high-priority data domains, cardiovascular (heart rate with zone distribution) and metabolic (CGM glucose with meal context). In both cases, the platform recovers clinically meaningful signals unavailable from single-value extraction, including temporal patterns, time-in-range estimates, and dietary correlations, without requiring API contracts with device manufacturers.
Screenshot extraction functions as a universal interoperability layer that extends SHIRLe’s reach to virtually every health technology the patient uses, without requiring any vendor integration agreement.
Illustrative Use Cases
Diabetes Self-Management
SHIRLe extracts diagnosis records, A1c results, and medication history from the patient’s EHI export, then integrates glucose monitoring screenshots and wearable activity data via HealthDigits. The AI Interpretation Engine identifies that glucose spikes correlate with specific meal patterns and that stability improves on high-activity days, insights clinical EHI alone could not surface.
Cardiovascular Risk Monitoring
Home blood pressure readings, wearable sleep data, and EHI medication history combine across a 90-day timeline. SHIRLe surfaces that blood pressure improves consistently during weeks when sleep exceeds 7 hours and step count exceeds 8,000 daily, a behavioral health insight report the patient can share with their cardiologist.
Orthopedic Recovery Tracking
A patient recovering from knee replacement combines surgical records and follow-up summaries with rehabilitation logs, wearable activity, and physical therapy app screenshots. The Timeline Engine visualizes recovery trajectory against expected milestones, enabling earlier intervention between scheduled visits.
Built on Open Standards
- FHIR R4
- US Core IG v6
- SMART App Launch 2.0
- SNOMED / LOINC / RxNorm
- HIPAA-compliant by design
One integration pattern works across Epic, Oracle Cerner, athenahealth, and any SMART-enabled system. Client-side screenshot processing means raw images are never transmitted to SHIRLe servers.
Follow SHIRLe Health
As we move into Phase 2, we welcome health systems, EHR vendors, payers, and patient advocacy organizations who want to help turn patient-controlled health intelligence into a tool patients can use. Reach out.