Written and maintained by CASRAI Editorial Board
Last updated
Epic Systems (Epic Systems Corporation, headquartered in Verona, Wisconsin) is an enterprise Electronic Health Record (EHR) software platform used by academic medical centers, integrated health systems, and community hospitals. Unlike departmental software, what clinicians and informaticists call Epic EHR (or Epic EMR) is a unified clinical, revenue-cycle, and research data architecture built on a three-tier database stack—Chronicles, Clarity, and Caboodle—coupled with standardized HL7 FHIR Release 4 (R4) APIs, Care Everywhere health information exchange, and the multi-system Epic Cosmos research network.
What Is Epic EHR and Who Runs It?
Founded in 1979 by Judith Faulkner as Human Services Computing and later renamed Epic Systems Corporation, Epic develops and maintains a single, integrated codebase spanning inpatient acute care, ambulatory clinics, emergency medicine, oncology (Beacon), radiology (Radiant), laboratory (Beaker), pharmacy (Willow), cardiology (Cupid), perioperative care (OpTime), revenue cycle (Resolute), and patient engagement (MyChart).
When academic medical centers evaluate Epic versus Oracle Health (formerly Cerner), a defining characteristic of Epic Healthcare is that all clinical and administrative modules share a single underlying patient record in Chronicles rather than bridging acquired third-party departmental databases.
How Is the Epic EHR System Architected?
To prevent heavy analytical queries or population-level research extracts from slowing down bedside charting, medication barcode scanning, or order entry, Epic separates its data persistence into three distinct architectural tiers:
- Chronicles (Operational Transactional Database): Chronicles is Epic’s proprietary, hierarchical, non-relational database management system built on InterSystems Cache / IRIS technology. All real-time clinical interactions inside the clinician workspace (Hyperspace and the web-based Hyperdrive client) and the MyChart patient portal read from and write to Chronicles. Data inside Chronicles is organized into a strict four-level hierarchy:
- Master Files (INI): Three-letter categories representing core entity types—such as
EPT(Patient),ORD(Orders),SER(Provider/Resource),HNO(Clinical Notes), andLWS(Workstations). - Records: A single entity instance within a Master File (for example, one specific patient’s record in
EPT). - Contacts: Time-stamped or encounter-specific snapshots within a record (such as an outpatient clinic visit or hospital admission).
- Items and Lines: Individual data fields (identified by item numbers) and multi-response values recorded during that contact.
- Master Files (INI): Three-letter categories representing core entity types—such as
- Clarity (Relational Reporting Database): Because hierarchical tree structures are inefficient for large multi-patient aggregation queries, an automated nightly Extract, Transform, and Load (ETL) pipeline extracts Chronicles Master File items into Clarity, a normalized relational database hosted on Microsoft SQL Server, Oracle, or cloud relational infrastructure. Clarity contains thousands of relational tables used by biomedical informaticists, biostatisticians, and compliance analysts to write complex SQL extracts.
- Caboodle (Dimensional Enterprise Data Warehouse): Formerly known as Cogito Data Warehouse, Caboodle sits downstream of Clarity and reorganizes clinical and operational data into a simplified star-schema dimensional model (Fact tables and Dimension tables). Caboodle can also ingest non-Epic external datasets (such as payer claims, external registries, or state death indices) and serves as the high-speed engine behind SlicerDicer, Epic’s self-service visual cohort discovery tool.
How Do Epic’s Operational, Reporting, and Interoperability Layers Fit Together?
In a live health system or academic research center, data flows continuously across Epic’s operational, reporting, and standards-based interoperability interfaces:
- Real-Time Analytics (Reporting Workbench / Radar): When clinicians or trial coordinators need same-day operational lists (such as active inpatient screening lists or overdue lab checks) that cannot wait for the nightly Clarity ETL, Epic’s Reporting Workbench queries Chronicles directly using indexed search templates.
- HL7 FHIR R4 and SMART on FHIR (`open.epic.com` / `fhir.epic.com`): To comply with the ONC 21st Century Cures Act certification rule (45 CFR § 170.315(g)(10)), Epic exposes standardized RESTful HL7 FHIR R4 APIs profiled to the HL7 US Core Implementation Guide. Third-party clinical decision support tools, digital therapeutics, and research Electronic Data Capture (EDC) connectors launch directly inside Hyperspace/Hyperdrive or MyChart via the SMART App Launch Framework (OAuth 2.0 / OpenID Connect) and register through the Epic Showroom (formerly App Orchard).
- Care Everywhere and National Frameworks: Epic’s native health information exchange module, Care Everywhere, exchanges longitudinal clinical summaries (HL7 C-CDA and FHIR documents) both across Epic organizations and with external non-Epic vendors participating in the Carequality interoperability framework and TEFCA.
How Does Epic Support Clinical Trials and Research Data Extraction?
Academic medical centers use Epic’s dedicated research functionality to connect clinical care with regulated clinical trial operations across the study lifecycle:
- Cohort Feasibility via SlicerDicer and Epic Cosmos: Principal investigators use SlicerDicer (querying local Caboodle de-identified or identified populations under IRB oversight) to test protocol inclusion and exclusion criteria—such as ICD-10-CM diagnoses, LOINC lab thresholds, RxNorm medication histories, and genomics indicators—before opening a trial. For multi-institutional real-world evidence (RWE) and rare-disease epidemiology, participating institutions query Epic Cosmos, a HIPAA-de-identified dataset aggregating longitudinal records across hundreds of health systems.
- Participant Recruitment and MyChart Research Workflows: Once a study record (
RSH) is built in Epic, recruitment teams configure automated BestPractice Advisories (BPAs) that alert clinicians during an encounter when a patient meets study screening criteria, or send direct study invitations through MyChart allowing patients to review study summaries, complete pre-screening questionnaires, sign electronic informed consent (eConsent), and submit electronic Patient-Reported Outcomes (ePROs). - Research Billing Coverage Analysis and Charge Routing (Resolute): Enrolling a participant in an Epic study record associates their encounters and orders with a protocol billing calendar derived from the site’s Medicare Coverage Analysis (MCA). When a research-associated order is placed, Epic routes the charge to a dedicated research billing review workqueue so billers can verify whether each service is routine qualifying clinical care (billed to the patient’s insurer with appropriate clinical trial modifiers and NCT identifiers) or a research-only item billed to the sponsor grant account.
- EHR-to-EDC Data Extraction (FHIR & SQL): Clinical data managers extract source-verified vital signs, concomitant medications, adverse events, and laboratory panels from Epic into external EDC platforms (such as REDCap, Medidata Rave, or Veeva Vault) using either patient-authorized/system-scoped HL7 FHIR R4 APIs or scheduled Clarity/Caboodle SQL extracts mapped to CDISC ODM/SDTM standards.
Strengths and Operational Limits
- Strengths: Unified single-database clinical chart across inpatient, outpatient, and specialty modules; clear separation between transactional speed (Chronicles) and analytical SQL warehousing (Clarity/Caboodle); mature ONC-certified FHIR R4 endpoints and native clinical research billing workflows.
- Limits: Chronicles’ nightly ETL latency means Clarity and Caboodle are typically 24 hours behind live bedside charting; local customizations to flowsheets (`FLO`) or SmartForms across different hospitals require careful LOINC/SNOMED CT mapping before multi-site SQL queries can be federated outside of standardized US Core FHIR profiles.
Enterprise EHR and Research Data Platforms Compared
| Platform | Primary Role & Architecture | Analytical / Research Data Layer | Interoperability Standards |
|---|---|---|---|
| Epic Systems | Enterprise integrated inpatient and ambulatory EHR (Chronicles hierarchical operational database) | Clarity (relational SQL), Caboodle (dimensional EDW), SlicerDicer, and Epic Cosmos network | HL7 FHIR R4 (US Core), SMART on FHIR, Care Everywhere (Carequality), HL7 v2, C-CDA |
| Oracle Health (Cerner Millennium) | Enterprise hospital and health-system EHR built on a unified relational person-centric architecture | PowerInsight, HealtheIntent population health platform, and Oracle Real-World Data (RWD) | HL7 FHIR R4 (Ignite APIs), SMART on FHIR, CommonWell & Carequality, HL7 v2, C-CDA |
| MEDITECH Expanse | Web-based enterprise EHR widely deployed in community, regional, and critical-access hospitals | MEDITECH Data Repository (DR) relational SQL database and Business and Clinical Analytics (BCA) | ONC § 170.315(g)(10) FHIR R4 APIs, Traverse Exchange, Carequality, HL7 v2 |
| athenahealth (athenaOne) | Multi-tenant cloud ambulatory EHR and revenue-cycle platform (with inpatient athenaOne for Hospitals) | Snowflake-backed data view extracts and population health reporting feeds | HL7 FHIR R4 APIs, athenahealth Marketplace REST APIs, Carequality, CommonWell |
Frequently Asked Questions
How is the Epic EHR system architected?
The Epic EHR system is architected around a three-tier database pipeline that separates real-time clinical operations from analytical querying. Live clinical charting in Hyperspace/Hyperdrive writes directly to Chronicles, a hierarchical non-relational database organized by Master Files (INI), records, contacts, and items. Every night, an Extract, Transform, and Load (ETL) process copies Chronicles data into Clarity, an enterprise relational SQL database (Microsoft SQL Server or Oracle), which in turn feeds Caboodle, Epic’s dimensional star-schema enterprise data warehouse.
How do Epic’s operational, reporting, and interoperability layers fit together?
Clinicians and patients interact with the operational layer (Hyperspace/Hyperdrive and MyChart) backed by Chronicles, which ensures sub-second transaction speed during inpatient and ambulatory care. Real-time interoperability—including ONC 170.315(g)(10) HL7 FHIR R4 APIs, SMART on FHIR app launches, Care Everywhere C-CDA exchange, and HL7 v2 lab/device feeds—reads and writes against Chronicles. Meanwhile, retrospective quality reporting, SlicerDicer cohort exploration, and longitudinal research extracts run against Caboodle and Clarity so complex queries never degrade live clinical performance.
How does Epic support clinical trials and research data extraction?
Epic supports clinical research across four workflows: (1) study feasibility and cohort discovery using SlicerDicer (on Caboodle) and the multi-institutional Epic Cosmos network; (2) participant recruitment and eConsent via MyChart research invitations and clinician BestPractice Advisories (BPAs); (3) clinical trial billing compliance by linking enrolled participants and research orders to a protocol billing calendar that splits routine care from sponsor-billed items; and (4) EHR-to-EDC data extraction via HL7 FHIR R4 APIs and Clarity/Caboodle SQL pipelines.
Sources
- Epic Systems Corporation — Epic on FHIR Developer Portal and HL7 FHIR R4 API Specifications
- Epic Systems Corporation — open.epic Interoperability Standards, SMART on FHIR, and Interface Documentation
- Epic Systems Corporation — Epic Cosmos Multi-Institutional Research Data Platform
- Epic Research — Dual-Validated Clinical Research Methodology on Epic Cosmos
