Google releases toolset to aggregate, map data to FHIR standard


Google’s cloud arm on Thursday unveiled a new product designed to aggregate and standardize data for healthcare and life sciences organizations.

The tool, dubbed Google Cloud’s “healthcare data engine,” joins a growing market of tools that technology companies are releasing aligned with the Fast Healthcare Interoperability Resources, better known as FHIR, standard.

Amazon last week unveiled general availability of HealthLake, a tool that healthcare and life sciences organizations can use to aggregate and analyze data in the cloud; it also indexes unstructured data from clinical notes and medical images. Amazon Web Services, the company’s cloud arm, has partnered with outside companies that customers would work with to align data to the FHIR format before moving it to HealthLake.

FHIR got a boost in healthcare last year when two HHS agencies released regulations requiring some healthcare software developers and insurers to adopt application programming interfaces—protocols that let different applications to communicate and share data with one another—that align with FHIR.

The regulations, first proposed by CMS and HHS’ Office of the National Coordinator for Health Information Technology in 2019, had earned support from technology giants.

Microsoft Corp., which also released a tool that aggregates and stores health data using FHIR through its Azure cloud arm in 2019, wrote to CMS and ONC when their regulations were proposed to voice support. Apple, which offers its own tool that lets patients download health data from hospitals and clinics through FHIR-based APIs, also supported the rules.

Google’s latest cloud tool, which it released Thursday as a private preview, aggregates data from medical records, claims and clinical trials and maps it to the FHIR format. It builds on Google Cloud’s healthcare API, an application programming interface it publicly released last year that’s designed to link healthcare data with third-party applications.

Ideally, data will be used to help clinicians identify gaps in care or flag patients at high-risk for certain conditions by providing a “longitudinal” patient record, and provide a foundation for researchers to analyze data with artificial intelligence and other analytics tools in the cloud, according to Google.

Google said the healthcare data engine tool was informed by work it’s doing with Rochester, Minn.-based Mayo Clinic, which has a multiyear contract with Google Cloud.

As of last year, FHIR and APIs hadn’t been widely adopted in healthcare, according to a report market research firm Chilmark Research released in spring 2020.

Brian Murphy, the analyst who wrote the report and director of research at Chilmark, wrote that he expects that to change in future years. They’ll become “dominant” in the healthcare industry, as requirements from HHS’ interoperability regulations go into effect and software vendors continue to create new tools that build on the FHIR standards.

Earlier this month, CMS began enforcing requirements from its interoperability rule that mandate payers it regulates implement FHIR-based APIs that let patients access claims and encounter information. Beginning in December 2022, developers of health IT software that receive ONC certification will be required to make FHIR-based APIs available to customers.

Over the past year, FHIR adoption has picked up some steam, according to Murphy.

“The adoption’s steady but slow,” he said. Still, “there’s a lot of energy and excitement around it.”

Tech giants have keyed in on healthcare as an area that’s ripe for AI adoption, said Jeff Becker, a principal analyst at CB Insights, a firm that analyzes data on venture capital and startups.

Having broad development and adoption of AI tools in healthcare will require “a lot of small steps,” starting with figuring out how to store vast sums of health data—including medical records, claims and patient-generated health data from devices—in a standard way that algorithms can digest and analyze, he said.

“The data standardization piece … is really just the first step within a longer story,” Becker said.