Epic Cogito: Chronicles, Clarity, and Caboodle Architecture, SlicerDicer, Nebula, and Aura Explained

Clinical data engineers, biomedical informaticists, and biostatisticians searching for Epic Cogito and the Epic data model need to understand how operational clinical documentation is separated from analytical reporting and cloud machine learning. Within an Epic Systems enterprise environment, Epic Cogito is the overarching analytics and business intelligence suite that unifies real-time operational reporting tools (Radar and Reporting Workbench), self-service cohort exploration (SlicerDicer), and the three-tier database pipeline spanning Epic Chronicles, Epic Clarity, and the Epic Caboodle data warehouse.

How Do Epic Chronicles, Clarity, and Caboodle Differ?

To prevent heavy SQL joins, population registries, and clinical trial feasibility queries from degrading sub-second charting latency at the bedside, Epic structures its persistence layer into three distinct database environments:

Architectural Dimension	Epic Chronicles	Epic Clarity Database	Epic Caboodle Data Warehouse
Primary Role	Real-time transactional operational database (OLTP)	Normalized relational reporting database	Dimensional Enterprise Data Warehouse (EDW)
Underlying Engine & Schema	Hierarchical tree structure built on MUMPS (InterSystems Cache / IRIS)	Relational SQL tables (Microsoft SQL Server, Oracle, or Teradata) with 18,000+ tables	Dimensional star schema (*Fact and *Dim tables on SQL Server)
Data Freshness	Real-time (immediate read/write from Hyperspace, Hyperdrive, and MyChart)	T-1 day (populated via nightly Extract, Transform, and Load [ETL])	T-1 day (populated nightly after Clarity ETL completes)
External Non-Epic Data	Clinical interface messages (HL7 v2, HL7 FHIR R4) ingested into patient records	Epic Chronicles data only	Combines Epic Clarity data with non-Epic external datasets (claims, ERP, cost accounting, external registries)
Metadata Dictionary Tool	Record Viewer / Item Information (Ctrl + Click)	Clarity Compass	Caboodle Console (Data Dictionary)
Primary Cogito Reporting Tools	Reporting Workbench, Radar operational components	Custom SQL extracts, SAP BusinessObjects / Crystal, Power BI / Tableau	SlicerDicer, Radar analytical metrics, enterprise BI dashboards, OMOP CDM ETL

1. Epic Chronicles (Real-Time Operational Hierarchy)

Epic Chronicles is the single transactional database where clinicians enter orders, chart flowsheet vitals, verify barcoded medications, and sign notes. Rather than using relational tables, Chronicles stores data in a four-level hierarchy optimized for rapid single-patient retrieval:

• Master Files (INI): Three-letter functional categories such as EPT (Patient), ORD (Order), SER (Provider/Resource), HNO (Note), and FLO (Flowsheet).
• Records: An individual entity within a Master File (for example, a single patient’s record in EPT identified by a unique internal ID).
• Contacts (CSN): Date-stamped or encounter-specific versions of a record, tracked by a Contact Serial Number (PAT_ENC_CSN_ID) so historical state changes remain auditable.
• Items and Lines: Discrete data elements identified by item numbers (single-response or multiple-response lines) within a contact.

2. Epic Clarity (Normalized Relational Reporting Database)

Because traversing hierarchical trees across hundreds of thousands of patients is computationally expensive, Epic extracts Chronicles items nightly into the Epic Clarity database. Clarity is a normalized relational SQL database containing thousands of tables (such as PATIENT, PAT_ENC, ORDER_PROC, ORDER_MED, and IP_FLWSHT_MEAS). Research data engineers use the Clarity Compass metadata tool to map Chronicles INI and Item numbers to their corresponding Clarity table and column names, including awareness of deprecated or replaced columns.

3. Epic Caboodle (Dimensional Star-Schema Data Warehouse)

While Clarity mirrors Chronicles’ granular structure—often requiring 8 to 15 table joins to assemble a single clinical encounter summary—Epic Caboodle (formerly known as Cogito Data Warehouse) reorganizes data into a simplified star schema consisting of central Fact tables (e.g., EncounterFact, MedicationOrderFact, LabComponentResultFact) surrounded by descriptive Dimension tables (e.g., PatientDim, ProviderDim, DateDim, DiagnosisDim). Caboodle uses surrogate keys (*Key columns, with negative keys representing special states like -1 Unspecified, -2 Not Applicable, and -3 Deleted) and supports Slowly Changing Dimensions (Type 1 and Type 2 historical tracking via IsCurrent flags) alongside external non-Epic data sources.

How Does Data Flow from Epic Chronicles into Clarity and Caboodle for Clinical Reporting?

The nightly data pipeline and front-end delivery layer operate under the Epic Cogito analytics framework:

1. Stage 1 — Chronicles to Clarity ETL: During off-peak overnight hours, the Clarity ETL server reads modified Chronicles records (tracked since the prior extraction timestamp), flattens hierarchical contacts and multi-response lines into relational rows, and loads them into Clarity staging and production tables.
2. Stage 2 — Clarity and External Sources to Caboodle ETL: Once Clarity packages finish, Caboodle execution packages transform normalized Clarity tables—along with any custom non-Epic SQL sources—into star-schema staging tables and finally into conformed *Fact and *Dim tables documented in the Caboodle Console.
3. Stage 3 — Cogito Front-End Delivery:
   • Radar: Role-based dashboards embedded directly in the clinician or researcher workspace that display key performance indicators, Metric Framework (IDN) summaries, and links to workqueues.
   • Reporting Workbench (RWB): Real-time operational reporting that queries Chronicles directly (or hybrid Clarity caching) to produce actionable patient lists—for example, identifying currently admitted inpatients who meet inclusion criteria for an acute clinical trial.
   • SlicerDicer: Epic’s self-service population exploration and cohort discovery tool built on top of Caboodle data models. Principal investigators use SlicerDicer to interactively filter patient populations by ICD-10 diagnoses, LOINC lab thresholds, medications, and social determinants of health without writing SQL, and can link local findings to national de-identified benchmarks in Epic Cosmos.

What Are Epic Cogito, Nebula, and Aura in Epic’s Analytics and Specialty Architecture?

Beyond the core three-tier database stack, health systems frequently encounter two specialized Epic platform layers that feed into or consume Cogito analytics:

Epic Nebula (Cloud Compute and Machine Learning Fabric)

Epic Nebula is Epic’s cloud-hosted platform (architected on Microsoft Azure) that provides managed compute infrastructure for executing predictive machine learning models, natural language processing pipelines, and cloud-scale telemetry without requiring each health system to provision and maintain local on-premises GPU clusters. Predictive models running on Nebula—such as deterioration indices, no-show risk models, or sepsis early-warning classifiers—ingest real-time Chronicles features and return risk scores directly into Hyperspace/Hyperdrive clinical workflows and Cogito monitoring dashboards.

Epic Aura (Specialty Diagnostics and Device Network)

Epic Aura (spanning Aura Labs and Aura Devices) is Epic’s network-level interoperability framework for specialty diagnostics, reference laboratories, genetic testing vendors, and remote medical devices. Historically, integrating an external molecular diagnostics laboratory into an health system required building and maintaining custom point-to-point HL7 v2 order/result interfaces for each hospital. Through Aura, participating diagnostic laboratories and device manufacturers connect once to the Aura network; ordering clinicians can place specialty orders natively inside Epic, and discrete structured results (including genomic variants and PDF reports) return directly into Chronicles, Clarity, and Caboodle for immediate clinical decision support and SlicerDicer research queries.