Healthcare data model


A key feature of healthcare data solutions is the healthcare data model. The data model is comprehensive and well-known and can support many business domains in healthcare, such as clinical, administrative, financial, and social. The healthcare data model provides a common data language that helps data analysts, data scientists, and developers to collaborate and build data-driven solutions that improve patient outcomes and business performance. The healthcare data model provides a unified, validated, and enriched version of the healthcare data that can be used for downstream analytics.

The healthcare data model supports analyzing FHIR clinical data through the utilization of traditional SQL-based tools such as T-SQL and Spark SQL, by providing FHIR data in a relational form.

The healthcare data model is a composite model that uses standards where they exist, such as FHIR and DICOM. It offers a comprehensive schema for those domains where no standards exist, such as patient engagement data. In addition to supporting the breadth of FHIR data, the model goes further by supporting the following data:

  • Medical Imaging (DICOM)

  • Patient engagement data

  • Genomics data

  • Social Determinants of Health (SDOH)

However, the core is designed to persist FHIR data in a relational format by using delta parquet for enhanced efficiency. This is done by making FHIR resources available as tables and columns within the healthcare lakehouse. This relational format supports the industry SQL on FHIR initiative. For data domains that aren't covered by the FHIR standard, the schemas from the Azure Synapse database templates are used, and this supports bringing nonclinical information like patient engagement data to the core patient profile.

The data model supports use cases, such as multiple modal clinical research with OMOP analytics.

Diagram showing the healthcare data model.