Skip to main content
Version: 3.1.3

Core Capabilities

  • Data Ingestion & Flattening: Automates conversion of FHIR® resources into analytics-ready tables.

  • Common Data Model(CDM) Compatibility: Supports Sentinel (v6.0.2), PCORnet (v4.0), i2b2/ACT (v1.4), and OMOP (v5.2) out of the box for regulatory and research needs.

  • Analytics Enablement: Integrates natively with Tableau, Power BI, Looker, etc. and supports SQL-based analytics workflows.

  • Security & Compliance: Full audit, lineage tracking, encryption, and user/role management built in.

  • Management Portal: Review data models, create custom marts, monitor flows, configure integrations.

  • Cloud-Agnostic Operations: Leverages cloud storage for scalability and cost reduction.

  • Observability: Real-time monitoring of data flows, jobs, and performance.

  • Third-Party Integrations: Connect to downstream data warehouses (Databricks, Snowflake), applications and AI/ML platforms.

High-Level Architecture

Hyperion®'s architecture is structured into three primary layers, adhering to the Medallion architecture principles:

Hyperion architecture

1. Bronze Layer (Raw Data Ingestion)​

  • Data Source: FHIR® Server
  • Process: Extracts FHIR®-formatted clinical and claims data, converting them into structured files.
  • Storage: Deposits these files into a designated landing zone, such as ADLS or Amazon S3.

2. Silver Layer (Data Transformation and Structuring)​

  • Process: Performs data flattening by parsing nested FHIR® resource structures and converting them into normalized, relational data models.
  • Storage: Stores the transformed data in a core schema within a relational database, facilitating efficient querying and integration.

3. Gold Layer (Semantic Views and Analytics)​

  • Process: Develops specialized database views that reference the Silver layer's relational tables.
  • Features:
    • Industry Standard Schemas: Supports schemas like Sentinel, PCORnet, i2b2/ACT, and OMOP.
    • Custom Data Marts: Allows creation of tailored data marts for specific analytical requirements.

Cloud-Native & Containerized:​

  • Hyperion® is cloud agnostic, deploys on Kubernetes, and integrates seamlessly on AWS, Azure, S3, ADLS, and on-premises environments.
  • On-Premises: For organizations requiring local data hosting.
  • SaaS (Planned): Managed service option for future release.

Data Flow Overview

  1. Data Extraction: Hyperion® connects to the FHIR® server to extract clinical and claims data.
  2. Data Conversion: The extracted data is converted into structured files and stored in the Bronze layer's landing zone.
  3. Data Flattening: In the Silver layer, the system flattens complex FHIR® resource structures into normalized relational tables.
  4. Data Modeling: The Gold layer utilizes these relational tables to create semantic views.
  5. Data Consumption: End-users access these views and data marts for reporting, analytics, and decision-making purposes.