Azure Health Data Services for scalable, secure, unified healthcare data management

Azure Health Data Services from Microsoft is a platform for unifying protected healthcare information (PHI) and healthcare data in one data store. It supports transactional and analytical workloads. It also enables cloud computing for AI applications and analytics to improve patient and research outcomes. Azure Health Data Services is central to the Microsoft Cloud for Healthcare.

Better, faster healthcare insights with protection

The health of people and populations depends on insights from medical and well-being data. However, disparate sources and systems with incompatible data formats can make it challenging to obtain these insights. Therefore, health data access, standardization, and trend information need a unified approach to make progress.

Azure Health Data Services is a platform-as-a-service (PaaS) solution designed to contribute to this unification. The platform ensures data privacy within compliance boundaries and de-identification of data for secondary use.

Platform APIs and features

Azure Health Data Services offers managed API services built according to open standards and frameworks.

• Fast Healthcare Interoperability Resources (FHIR?) service - includes FHIR APIs and endpoints in Azure for data access and storage in FHIR data format.
• Digital Imaging and Communications in Medicine (DICOM?) service - to acquire medical imaging data from any DICOMweb? enabled system.
• MedTech service: an IoT Connector service to bring high-frequency IoT device data into FHIR service.

Additional features include:

• Connectors to Azure Synapse Analytics, Azure Machine Learning, and Power BI for insights from healthcare data and Teams for collaboration and other Microsoft products.
• HITRUST CSF certification meets HIPAA and GDPR requirements and CCPA, ONC, CMS, and regional compliance mandates for handling protected health information.
• Control of access to health data within a compliance boundary with application monitoring (Azure Monitor) and role-based access controls (Azure RBAC).

FHIR service for interoperability

The emerging standard of FHIR for clinical data allows interoperation between FHIR systems in modern mobile, web, and other environments. It offers extensibility for integrating different types of data directly or through references. Converting data to FHIR allows rapid connection of data sources like electronic health record systems and research databases. Azure Health Data Services can transform data from legacy formats like HL7v2 or CDA or from proprietary formats like high-frequency IoT device data. FHIR can also simplify data ingestion and speed up the development of analytics and machine learning tools.

Microsoft makes three FHIR capabilities available:

1. FHIR service in Azure Health Data Services.
2. Azure API for FHIR (for FHIR data only).
3. FHIR Server for Azure, an optional open-source project within an Azure subscription.

FHIR service is an API and service designed for rapid deployment and use in data exchange. The consistent, RESTful, FHIR API is based on the HL7 FHIR specification. FHIR service provides scalable, secure management and storage in the native FHIR format of protected health information (PHI) data. A unique database isolates the data for each API instance. Security is ensured by a layered in-depth defense, advanced threat protection, and multi-region failover.

Azure Health Data Services is a superset of Azure API for FHIR with added services. Existing customers can continue using Azure API for FHIR without impact on service or pricing structure.

Developers who want to extend or customize FHIR Server or need access to underlying services like the database without using the FHIR APIs should use the open-source FHIR Server for Azure. On the other hand, they should choose the FHIR service for a production ready FHIR API and backend service with persisted data accessible only via the FHIR API.

DICOM service and imaging data

DICOM is an international standard for transmitting, processing, storing, retrieving, printing, and displaying medical imaging information. It is the primary medical imaging standard in healthcare. As a managed service within Azure Health Data Services, DICOM service can ingest and persist DICOM objects at a rate of multiple thousands of images per second. Communication and transmission of imaging data are available between any DICOMweb? enabled systems or applications through DICOMweb Standard APIs like Store (STOW-RS), Search (QIDO-RS), and Retrieve (WADO-RS).

DICOM service helps meet an essential need for clinical and imaging data integration. Not only can the imaging data persist securely in the Microsoft cloud, but it can also coexist with EHR and IoT data in the same Azure subscription. In addition, the DICOM service enables organizations to store FHIR references to imaging data and perform queries that combine clinical and imaging datasets. Different use cases include:

• Back-up of images
• Image exchange and collaboration
• Disaster recovery
• Studies over time (longitudinal studies) of individual patient diagnoses and cohorts
• Completing feedback loops, for example, with teleradiologists, and for AI/ML models

MedTech service for the Internet of Medical Things (IoMT)

Health data collected from patients and health care consumers can originate from various systems and devices in different formats. As an optional service of the Azure Health Data Service, MedTech service helps ingest health data from multiple and varied Internet of Medical Things (IoMT) devices. The devices first send their IoMT data into an event hub. Next, the MedTech service converts the device data into FHIR-based Observation resources. It then persists the converted messages into FHIR service.

Scalable and configurable, MedTech service can be combined with Microsoft open-source projects for acquiring IoMT device data from wearables. It may also be used with Azure Machine Learning Service, Power BI, and Teams for additional functionalities and insights. For finer-grained security and access control of its assets, MedTech service uses Azure Resource-based Access Control and Managed Identities.

Workspace creation and deployment

Workspaces for Azure Health Data Services can be set up with Azure Resource Manager (ARM) deployment templates as an infrastructure as code (IaC) process. Customization of the ARM templates and workspace and service instance creation can be done in a combined step. Azure Health Data Services deployment is possible via PowerShell, CLI, Terraform scripts or the .NET SDK.

By design, multiple data services that are provisioned within an Azure Health Data Services workspace will work together seamlessly. In addition, datasets and services in a workspace can also share specific configuration settings for easier organization and management.

To learn more, check out these resources:

https://docs.microsoft.com/en-us/azure/service-health/overview?WT.mc_id=AZ-MVP-4029307

https://docs.microsoft.com/en-us/azure/healthcare-apis/healthcare-apis-overview?WT.mc_id=AZ-MVP-4029307

https://docs.microsoft.com/en-us/azure/healthcare-apis/fhir/overview?WT.mc_id=AZ-MVP-4029307

https://docs.microsoft.com/en-us/azure/healthcare-apis/dicom/dicom-services-overview?WT.mc_id=AZ-MVP-4029307