What Google Cloud learned about interoperability from Mayo Clinic

A little over a year ago, Google Cloud and Mayo Clinic struck a 10-year strategic partnership, with the goal of building an “assembly line” of clinical AI tools. To support this, they’d need to have comprehensive and standardized data sets — a rarity in healthcare.

But this process of aggregating together healthcare data and bringing it to the cloud was time consuming, even for Mayo Clinic, which has more technical resources than most. Part of the challenge was that engineers needed to evaluate different options for cloud configuration, including best practices for networking and security.

Working with Mayo Clinic through this helped Google Cloud as it was building its Healthcare Data Engine, a cloud-based tool for healthcare companies to aggregate and standardize data from medical records, insurance claims, clinical trials and research.

“The learnings around that configuration are something that we’ve incorporated, as well as our learnings from other large organizations like (Mayo) who brought protected health information on the cloud,” said Marianne Slight, product manager for Google Cloud Healthcare Analytics, in a video interview.

The platform is intended to make it easier for health systems to get real-time information to identify high-risk patients, allocate resources, and other critical decisions. Before this, most health systems aggregated data manually using clinical data warehouses, which can be harder to scale and keep up-to-date.

Knowing what types of data sources and formats Mayo Clinic wanted to work with helped Google focus its attention on what to build. As a cloud provider, Slight emphasized that Google Cloud cannot look at customers’ health data or train models on that data.

Jim Buntrock, Mayo Clinic’s vice chair of information technology, said in a news release that the health system had been “hitting a wall” with its ability to innovate using on-premise software.

“There are so many applications of this. For example, building a ‘heads up display’ for the ICU — where moments matter — to help care teams direct their attention when and where it’s needed most,” he said. “From creating better ways to care for patients remotely even after they leave the hospital to making it easier for patients to interact with us via mobile app, we’re working alongside Google Cloud to build a platform for healthcare transformation.”

The Healthcare Data Engine is currently available as a pilot. Some of Google’s customers are looking to use it to run projects on population health and transitions of care. Another project is focused on health equity —  specifically, ensuring all patients were getting follow-up for screenings, Slight said.

Indiana University Health, one of Google’s named pilot customers, is working on using the platform to improve care delivery.

Google isn’t the only company working on this problem. Amazon also recently rolled out Amazon HealthLake, its own cloud-based system to help healthcare companies store and analyze health information.

Photo credit: ipopba, Getty Images