Jagamisviis:

Clinical solution architecture of healthcare data solutions (preview)

  • Artikkel

[This article is prerelease documentation and is subject to change.]

Healthcare organizations face challenges in meeting the demand for real-time access to cloud-based data while ensuring secure sharing across teams. Healthcare data solutions in Microsoft Fabric (preview) help you address these challenges by standardizing how you manage data, and ensure consistency, accuracy, and reliability. You can unify, transform, and enrich data according to industry standards. You can also collaborate seamlessly with other teams and use advanced analytics to gain insights and build next-generation applications. For more information, see Overview of healthcare data solutions in Microsoft Fabric (preview).

Here's a clinical architecture diagram that helps you understand how the healthcare data solutions (preview) ecosystem integrates within the Microsoft Azure and Microsoft Fabric platforms. This diagram gives you a comprehensive view of the components that make up the clinical pipeline flow. Each component is depicted along with its orchestration and interaction with the designated lakehouse. When you deploy the capabilities, you can find these notebooks and lakehouses in your Fabric workspace, illustrating their collaborative functionality.

A diagram displaying the healthcare data solutions clinical ecosystem.

The solution combines ingestion, transformation, and enrichment activities with standards-based data models.

  • Export Fast Healthcare Interoperability Resources (FHIR) data from an Azure Health Data Services instance or land FHIR data into OneLake. You can use different ingestion methods such as the Fabric Data Factory or the OneLake file API.
  • Ingest activities to model FHIR data as bronze delta ingestion tables. This approach allows you to access data via SQL analytics endpoints and facilitates a consistent downstream experience.
  • Transform data using the Healthcare data foundations, Observational Medical Outcomes Partnership (OMOP), Dynamics 365 Customer Insights - Data preparation, and DICOM data ingestion pipelines.
  • Use Azure AI Language's Text Analytics for health to extract natural language processing entities and relationships from clinical notes and store the results back into the data model.

You can create Fabric Data Factory pipelines to orchestrate and schedule the activities while also adding custom activities. After the data is ingested and transformed along the various stages of the medallion architecture lakehouses, you can query and analyze it using Spark SQL or T-SQL. You can also create Power BI reports for visualizing the data and use Fabric data science capabilities (such as machine learning experiments).