Metriport provides tools to transform your data for analytics and pipe it into warehouses like Snowflake, Redshift, BigQuery, and more - enabling you to do population level analytics on your patient data. There are two primary ways to perform analytics on your patient population data: Managed and Manual.

Managed: Snowflake Data Warehouse

Metriport offers direct integration with Snowflake, allowing you to analyze your patient data in a powerful, scalable cloud data warehouse. This approach provides:
  • Managed data transformation: Metriport does the heavy lifting to get your data into an analytics ready data warehouse 🎉
  • Near Real-time Analytics: Query your patient data right after it’s processed
  • Scalable Performance: Leverage Snowflake’s elastic compute resources for complex analytics
  • Secure Data Sharing: Use Snowflake’s Secure Data Sharing feature for secure data collaboration

Setting Up Snowflake Integration

  1. Create a Snowflake Account: Sign up for a Business Critical Edition Snowflake account and share your account identification with us:
    • View account details:
    • Copy these fields and share them with your Metriport account manager:
      • Organization Name
      • Account Name
      • Account Locator
      • Cloud Provider (AWS, Azure, GCP)
      • Region
  2. Access the data: Metriport will share your patients’ medical data with your Snowflake instance:
    • It uses Snowflake’s Secure Data Sharing for secure, compliant, and zero-copy data sharing between Metriport and your Snowflake account.
    • Data will be visible through Secure Views, one for each resource type.

Manual: Tuva Data Model

To accomplish this, we standardize our data into the Tuva data model - another open source project that focuses on making healthcare data more standardized for analytics. The Tuva approach provides:
  • Open Source: Transparent, community-driven data modeling
  • Healthcare-Specific Metrics: Built-in calculations for common healthcare KPIs
  • Multi-Warehouse Support: Works with Snowflake, BigQuery, Redshift, and other warehouses
To get started with Tuva, check out our data warehouse connector documentation on the Tuva site.

Schema Documentation

The Metriport data model consists of several core tables that represent different aspects of healthcare data. Each table is designed to capture specific information in a standardized format for analytics:

Core Tables

  • Condition - Symptoms, problems, complaints, and diagnoses reported during encounters
  • Encounter - Unique patient interactions with the healthcare system
  • Immunization - Vaccines administered to patients
  • Lab Result - Laboratory test results with LOINC codes
  • Location - Practice and facility locations where care is provided
  • Medication - Medications ordered or administered during encounters
  • Observation - Clinical measurements like blood pressure, height, and weight
  • Patient - Patient demographic information and core patient data
  • Practitioner - Healthcare providers who deliver care
  • Procedure - Procedures performed on patients
Each schema page provides detailed information about the table structure, including primary keys, foreign keys, and column definitions.