⚡EpilepsyLive ✉ Suggest a correction
Demonstration and educational project by Danielle Boyce. Not medical, legal, or regulatory advice. Sample and template language is provided for illustration and must be reviewed and adapted before any real use. Not affiliated with, endorsed by, or representing any advocacy group, registry, company, or institution named.
Learn Overview Patient Registries 101 Data Analysts Grant Writers Data Sharing in a Box
Module 9 of 19Part 3 · Data Collection

EHR Integration, CCDA & FHIR

Goal: Understand how to extract structured clinical data from EHR systems using industry standards, reducing participant burden and improving data quality.
▶ View slides ⇩ Slides (.pptx) ⇩ Full course (PDF)

Why EHR integration is important

Patient-reported data is invaluable, but it has limitations: participants may not know their exact lab values, may misremember medication doses, or may not be aware of diagnoses documented in their records. EHR-sourced data provides clinically precise, longitudinal records that complement what participants report.

The good news: federal law (the 21st Century Cures Act) now requires EHR vendors to provide patient access to their data, and your registry can leverage this.

C-CDA: The Legacy Standard

What is C-CDA?

The Consolidated Clinical Document Architecture (C-CDA) is an HL7 XML standard for clinical documents, the format EHRs have used for years to exchange participant records. When a participant "downloads" their health records from a patient portal, they typically receive a C-CDA document (or a PDF).

C-CDA documents include sections for:

  • Problems (diagnoses, coded in SNOMED CT or ICD-10)
  • Medications (RxNorm)
  • Allergies
  • Lab results (LOINC)
  • Procedures
  • Vital signs
  • Immunizations
  • Social history

C-CDA for registries

Participants can download their C-CDA from their patient portal and upload it to your registry. Your platform can then parse the XML to extract structured data, diagnoses, labs, medications, without the participant needing to manually enter it.

Limitations of C-CDA:

  • Snapshot in time (not continuously updated)
  • Inconsistent implementation across EHR vendors
  • XML parsing requires technical infrastructure
  • Requires participant action to download and upload

FHIR: The Modern Standard

What is FHIR?

HL7 FHIR (Fast Healthcare Interoperability Resources) is the modern standard for healthcare data exchange. Unlike C-CDA's document-based approach, FHIR uses a RESTful API model, meaning data can be queried, retrieved, and continuously synchronized, just like any web API.

FHIR represents data as Resources, discrete objects like Participant, Condition, Observation, Medication, and Procedure. Each resource has a defined structure, standard fields, and uses standard terminologies (SNOMED CT, LOINC, RxNorm).

FHIR R4 is the current stable version. FHIR R5 is available but not yet widely implemented.

hl7.org/fhir

FHIR for registries

With FHIR APIs, your registry can:

  • Pull a participant's diagnoses, labs, and medications directly from their EHR in real time
  • Receive automatic updates when new lab results are available
  • Query across multiple participants at clinical sites that grant API access
  • Submit data back to EHRs (e.g., registry-generated phenotype scores visible to the treating physician)

SMART on FHIR

SMART on FHIR is a framework that enables third party apps (like your registry) to securely access EHR data with patient authorization. It uses OAuth 2.0 for authorization, the same standard that powers "Login with Google."

A participant can:

  1. Visit your registry's patient portal
  2. Click "Connect my health records"
  3. Authenticate with their EHR's patient portal (e.g., MyChart)
  4. Authorize specific data types to be shared with your registry
  5. Your registry receives a FHIR access token and can pull their records

smarthealthit.org

Key FHIR Resources for registries

FHIR Resource What it contains Terminology
Patient Demographics, identifiers ,
Condition Diagnoses, problems SNOMED CT, ICD-10
Observation Lab results, vital signs, survey answers LOINC
MedicationRequest Prescribed medications RxNorm
Procedure Procedures performed SNOMED CT, CPT
DiagnosticReport Lab panels, imaging reports LOINC
DocumentReference Clinical notes, C-CDA documents LOINC
Questionnaire / QuestionnaireResponse Structured forms and participant responses LOINC

Implementation considerations

Site agreements

Pulling FHIR data from clinical sites (rather than directly from participants) requires:

  • Data Use Agreement with the health system
  • IRB approval covering EHR data access
  • Possible Business Associate Agreement (BAA) under HIPAA

Bulk FHIR

For population level data extraction (pulling data for many participants at once), use the FHIR Bulk Data Access specification, designed for exactly this use case.

hl7.org/fhir/uv/bulkdata

Checklist

  • [ ] Determined whether participant-initiated C-CDA upload or SMART on FHIR is the right approach
  • [ ] Selected platform with FHIR R4 support
  • [ ] Mapped registry data elements to FHIR resources
  • [ ] Drafted site agreements for clinical site FHIR access
  • [ ] Confirmed IRB protocol covers EHR data collection
  • [ ] Tested SMART on FHIR patient authorization flow

Key resources

← Module 8 | Module 10: Patient-Reported Data & 21st Century Cures →

Aligned with EpilepsyLive. The Registry Toolkit on this site frames record acquisition as five pathways a family or registry can use. This module covers the mechanics of the first two; the rest are summarized in Module 10 and demonstrated in the toolkit.
  • Patient portal login, SMART on FHIR (covered above).
  • Download & upload, a C-CDA file the person exports and re-uploads (covered above).
  • HIPAA authorization, a signed release directing a provider to send records.
  • Network exchange, TEFCA / Individual Access Services (see Module 10).
  • EEG & imaging studies, large binary studies requested separately (see Module 10).
See the interactive Registry Toolkit health records demo for each pathway.
RelatedRegistry Toolkit: health records demo·Patient-directed access & TEFCA/IAS
Something look wrong or unclear? Suggest a correction or send a note. EpilepsyLive is a demonstration project by Danielle Boyce, boycedatascience.com.