One flattened schema across all of your data.
Every source and every patient lands in the same normalized clinical tables — so a query is written once, not once per source.
Join the teams running population queries on one unified schema.
Read customer stories
Data Warehouse
The schema is the product. The pipeline is your choice.
Normalized clinical tables, not raw payloads
The warehouse holds the Harmonization Engine's output — the same consolidated FHIR R4 data the API returns, flattened for analytical queries.
Managed: Snowflake
Metriport operates the transformation and shares normalized tables through Snowflake Secure Data Sharing.
Manual: the open-source Tuva path
Standardized into the Tuva data model for Snowflake, BigQuery, Redshift, and other documented warehouse environments — customer-operated, fully open.
Write the query once
Population questions stop being one query per source format. The schema is the same for every patient and every origin.
Benefits
Population questions, without an ETL project first.The transformation is the product — your analysts start at the query.
The pipeline your data team was going to build.
Clinical data arrives analytics-ready: converted, coded, deduplicated, and flattened. The months of transformation work between raw records and a queryable table are already spent.
See the Harmonization EngineQuery soon after processing, not quarterly.
Managed Snowflake data becomes available after processing runs, so analysis reflects the population as it is — not as it was at the last batch export.
Read the analytics docsOpen standards underneath, no lock-in on top.
The manual path runs on Tuva, an open-source healthcare data model, in the warehouse you already operate. The managed path is a convenience, not a wall.
See the APIHow it works
Pick the path, connect the warehouse, write the query.Managed Snowflake, or open-source Tuva in your own environment.
Choose the operating model
Two documented paths to the same schema — the docs’ own comparison, verbatim.
Or run the open-source model yourself
The Tuva path is customer-operated and warehouse-agnostic: Snowflake, BigQuery, Redshift, and other documented environments.
Support
FAQs
Normalized clinical tables derived from the same consolidated FHIR R4 data the Medical API returns — the output of the Harmonization Engine, not raw source payloads.
Metriport operates the clinical-data transformation and shares normalized tables through Snowflake Secure Data Sharing.
Tuva supports Snowflake, BigQuery, Redshift, and other documented warehouse environments. This is a customer-operated open-source path, not a native managed connection to each warehouse.
Managed Snowflake data becomes available after processing. Contact Metriport for the current delivery behavior rather than relying on a fixed refresh interval.
Explore our platform
All our products
Where does the data come from?
What shape is it in once we have it?
Where does my team actually use it?