👋 Welcome

Tuva is an open-source dbt project that transforms healthcare data into a common model for analytics, applications, and AI inside your data warehouse. Connectors map payer claims and clinical data into the Input Layer; Tuva Core standardizes and enriches it, and optional Data Marts add focused analytical capabilities.

Tuva is trusted by more than 100 U.S. healthcare organizations. Start with synthetic data or your own data, or explore the Tuva 1.0 features and upgrade guide.

Tuva Core

Tuva Core transforms connector-mapped Input Layer data into a normalized and enriched common data model for analytics, applications, and AI. Optional Data Quality checks help teams validate their inputs.

1Data Contract

Input Layer

A standardized set of tables that defines how payer and provider source data must be mapped before Tuva Core runs.

Payer ClaimsEHRHIEADTLabs

2Optional Validation

Data Quality

Structural and logical checks that help identify missing, invalid, duplicate, and inconsistent source data before analytics.

CompletenessValidityIntegrityPlausibility

3Claims Groupers

Claims Preprocessing

Reusable claims transformations that create enrollment flags, encounter groupings, service categories, and provider attribution.

Service CategoriesEncountersMember MonthsProvider Attribution

4Analytics-ready Model

Core Data Model

A normalized, enriched, patient-centered data model that unifies claims and clinical data for analytics, applications, and AI.

ClaimsClinicalShared

5Reference Data

Data Assets

Versioned terminology, value sets, provider data, and synthetic data that dbt seed or dbt build loads for Tuva Core to use to validate source data, enrich standard codes, and map granular codes to higher-level analytic concepts.

TerminologyValue SetsProvider DataSynthetic Data

Data Marts

Standalone Data Mart packages further transform Tuva Core for specific analytical use cases and release independently.

CCSRAHRQ CCSRs

Standalone package implementing AHRQ's Clinical Classification Software Refined diagnosis and procedure groupers on top of Tuva Core.

QIAHRQ Quality Indicators

Standalone package implementing AHRQ's Quality Indicators, covering the retained Prevention Quality Indicators (PQIs).

CCWCMS Chronic Conditions

Standalone package identifying chronic-condition evidence in Core diagnoses and encounters using CMS condition code sets.

HCCCMS HCCs

Standalone package for CMS-HCC risk adjustment models, recapture calculations, and suspecting workflows.

EDNYU ED Classification

Standalone package assigning ED visits the highest-probability category from the NYU ED Classification algorithm.

QMQuality Measures

Standalone package with open-source quality measures and readmissions logic, excluding HEDIS measures.

Tools & Extensions

Tools & Extensions add source connectivity, dashboards, FHIR interoperability, Medicare ACO workflows, and reusable semantic outputs to Tuva.

CConnectors

Transform raw healthcare source data into the Tuva Input Layer using dbt connector projects.

DDashboards

Explore BI templates for healthcare analytics and check their Tuva version requirements before connecting data.

FHIRFHIR

Flatten incoming FHIR JSON with FHIR Inferno, or prepare Core outputs for FHIR resources with FHIR Preprocessing.

ACOMedicare ACOs

Work with CMS MSSP data pipelines and connectors, with migration guidance for legacy ACO dashboards.

SLSemantic Layer

Reshape Tuva outputs into reusable dimensions and facts for BI tools and downstream analytics.