FHIR Server for Azure: An open source project for cloud-based health solutions
In the last decade, the healthcare industry has accomplished a significant transformation with the digitization of health data. As we look to the next ten years, our challenge will be to connect and leverage that digitized data to power innovation and AI.
Healthcare developers are tasked with the challenge to bring diverse data sets together and develop machine learning across those data sets. We believe the best way to support developers working with health data is to offer tools that allow them to come together – for collaboration, creation, sharing, and building on each other’s work.
Today I’m excited to announce the release of an open source project on GitHub from Microsoft Healthcare: FHIR Server for Azure. FHIR Server for Azure empowers developers with software that fully supports exchange and management of data in the cloud via the FHIR specification.
FHIR Server for Azure provides support infrastructure for immediate provisioning in the cloud, including mapping to Azure Active Directory (Azure AD), and the ability to enable role-based access controls (RBAC). Developers can save time when it’s required to integrate a FHIR server into an application or use it as a foundation to customize a unique FHIR service.
Microsoft is contributing this open source project to make it easier for all organizations working with healthcare data to leverage the power of the cloud for clinical data processing and machine learning workloads.
We believe in FHIR
Fast Healthcare Interoperability Resources (FHIR) is rapidly gaining support in the healthcare community as the next generation standards framework for interoperability, and it’s clear why. FHIR provides a simpler, easier-to-learn, and pragmatic framework. Building on the shared experience of healthcare standards communities, FHIR offers an extensible data model and a REST API to simplify the implementation and interoperability of health data.
In August 2018, Microsoft joined with Amazon, Google, IBM, and other companies in a commitment to remove barriers for the adoption of technologies that support healthcare interoperability, particularly those that are enabled through the cloud and AI. Supporting the FHIR standard and investing in open source to enable FHIR is core to that commitment.
We want to make it easier and more secure for organizations working with healthcare data. We’re starting with FHIR Server for Azure, and we plan to enable a broad set of core services in the Microsoft Cloud to support healthcare interoperability standards like FHIR.
AI in healthcare, starting with FHIR in the cloud environment
In almost every facet of healthcare, the ambition to create and deliver AI exceeds the tools available to deliver it. FHIR Server for Azure provides a foundation to address that problem. Working with data in the FHIR format, developers can use the server to quickly ingest and manage FHIR datasets in a cloud environment, track and manage data access, and begin to normalize data for machine learning workloads.
This open source release includes support infrastructure for easy deployment:
- Scripts and ARM templates are available for immediate provisioning in the Microsoft Cloud
- Scripts are available to map to Azure AD and enable RBAC
- The GitHub documentation includes an “easy button” that deploys FHIR Server for Azure into nearly any Azure public region.
As an open source project, FHIR Server for Azure is built with logical separation, enabling developers with flexibility to modify how it is implemented, and extend its capabilities as needed. The logical layers of the FHIR Server are:
- Hosting Layer – Supports hosting in different environments, with custom configuration of Inversion of Control (IoC) containers
- RESTful API Layer – The implementation of the APIs defined by the HL7 FHIR specification
- Core Logic Layer – The implementation of the core FHIR logic
- Persistence Layer – A pluggable persistence provider enabling the FHIR server to connect to virtually any data persistence utility. FHIR Server for Azure includes a ready-to-use data persistence provider for Azure Cosmos DB (a globally replicated database service that offers rich querying over data).
In our initial release of FHIR Server for Azure, we are supporting FHIR STU3, which is the current version of the FHIR API. We have been actively engaged with HL7 and the FHIR community to support the standards development process for FHIR R4 and are excited about the forthcoming publication of FHIR R4. We plan to support FHIR R4 in a future version of FHIR Server for Azure, once the R4 specification has been finalized and published by HL7.
Playing with FHIR
Lighting up AI and innovation with data normalized in FHIR is our passion, but we developed the FHIR Service for Azure with a focus on data security first. The requirements for HIPAA, HITRUST, GDPR and maintaining security between existing health systems can often stagnate innovation. That’s why Microsoft Healthcare has been working closely with healthcare organizations, government policymakers and other technology leaders to ensure we deliver trusted FHIR technology. When deployed in the Microsoft Cloud, all Azure services used to support the FHIR Server for Azure are ISO 27001 certified and meet all the compliance requirements for HIPAA and GDPR.
We want to help you play with FHIR! The future of cloud scenarios you can create with FHIR are unlimited, but some of the best cloud scenarios for implementation are:
- Interoperability: Health systems can facilitate interoperability and normalization of data between disparate systems through FHIR APIs and a FHIR service
- Internet of Medical Things (IoMT): Startups and new device/app developers can document and share data in the FHIR format, facilitating faster integration with provider and payer ecosystems (e.g. sending data to EHRs via FHIR APIs)
- Research: Leveraging data in FHIR formats can help to expedite normalization of data for data scientists and researchers and allow for more granular control of data that needs to be shared across groups. By enabling RBAC and audit logs, the use of a FHIR service allows data owners more control over data sharing.
Get started today
This open source project is fully backed by the Microsoft Healthcare engineering team, but we know that this project will only get better with your feedback and contributions. We are actively driving the development of this code base and test builds and deployments daily. You can learn more about the architecture of the FHIR server and how to contribute to the project at the FHIR server site.
HEATHER JORDAN CARTWRIGHT
General Manager, Microsoft Healthcare