Data Marts

When you connect to the Bulk Data API, you will have access to one or more data marts.

What is a Data Mart?

A data mart is a set of tables in a database schema in an Amazon Redshift database. Many applications and frameworks can connect and authenticate to a data mart for a broad set of use cases.

See the data model documentation to understand the core concepts that will be made available as data tables within a data mart. A data mart will likely expose a subset of these core concepts. The mart may also extend these core concepts with additional tables that have been developed for customized use cases.

A data mart is based on a set of patients, members, study participants, or administrative concepts (e.g. terminology, or audit events) that serve as the root data concept for the mart. All data in the data mart will relate to and describe this base set.

Data marts are refreshed at a configurable interval by services that obtain updated data from HIEBus and recompute enriched concepts. For realtime marts, this interval is commonly nightly or several times throughout the day. If appropriate for the use case, data marts may preserve a static data set for any time interval in order to serve as a baseline, archive or validation set.

Using Data Marts

Data marts serve as a connection point between applications that consume bulk data and the CareEvolution health data platform that integrates and enriches source data into analytics-ready concepts. The following video demonstrates how a data mart appears in Tableau, DBeaver SQL client, a Juptyer notebook, and as a file in AWS S3.

Data Mart Demonstration

Examples of Data Marts

Commonly implemented data marts are described below:

  • Participants in an RKStudio project, their survey data, and all health data associated with the participants.
  • All patients in a specified “active population” for a health system. And their health data from all sources as well as enriched concepts required for the use case.
  • Patients attributed to a population health program, region or provider. And their health data from all sources as well as enriched concepts required for the use case.
  • Members of a health plan, ACO, or assigned to a specified set of attributed providers. And their health data from all sources as well as enriched concepts required for the reporting use case.
  • Patients or members with a specified recent utilization history and their associated data.
  • A disease or condition registry based on historical diagnoses, services, or treatments.
  • A research set of patients or members (and all their associated data) with a specified utilization and data history that selects these individuals for data analysis or a model development use case.
  • All terminology concepts in reference namespaces, mappings to local namespaces, value set membership, and usage statistics to validate, analyze and improve terminology administration.

Data Mart Authorization

Data marts are the most granular securable construct within the Bulk Data API. Once granted access to a mart, a credentialed user will have access to all data within that mart. CareEvolution may present your organization’s data in several marts in order to provide appropriately limited sets of data to each user, application or use case.

Data marts are commonly blinded (fully or partially) to obscure or limit the PHI revealed in the data tables if it is not required for a specific reporting or research use case. Use cases that require PHI for drill-down workflows, traceability, or integration with other resources will contain the appropriate PHI in the data mart.