This past September, the Rochester, Minnesota-based Mayo Clinic announced a ten-year partnership with Google to advance the health system's cloud-based artificial intelligence and machine learning capabilities.  

Six months later, the shape of healthcare had changed dramatically in response to the COVID-19 pandemic – and so had many of Mayo's priorities.

"This was an opportunity for us to really test this model in an urgent setting," said Mayo Clinic Chief Information Officer Cris Ross in a complimentary HIMSS Learning Center presentation this week.

"What could we do to leverage even our earliest work to try to get insights into the course of the disease of [COVID19] and to augment what we are doing to advance cures and … treatment [of] the disease?" Ross continued.

In the presentation, The Investment from the Ground Up: What It Takes to Prime a Healthcare Organization for AI and ML, Ross joined Google Cloud Global Healthcare Solutions Director Aashima Gupta; Mayo Clinic Platform President Dr. John Halamka; Mayo Vice Chair of IT James Buntrock; and Google Cloud Healthcare and Life Sciences Engineering Director Ilia Tulchinsky to discuss the lessons learned from the first half-year of the partnership and what the organizations hope to accomplish in the future.

"Our vision here was to bring two great organizations together to advance healthcare," said Ross.

"In many ways, the pandemic has been a catalyst for removing barriers and encouraging collaboration – even among competitors," said Halamka, noting that a number of software giants have been driven to work together for the benefit of society.

"Technology has been a true enabler," he continued.

In general, Tulchinsky explained, priming an organization for AI and ML involves three major stages: integrating, harmonizing and analyzing data. When it comes to healthcare in particular, the data involve a high degree of complexity, with dissimilar data ontologies and modalities and large volumes of information. 

"Provenance and lineage are really important," said Tulchinsky. "What was the journey of the data up to this point?"

Tulchinsky also noted the importance of integrating AI and ML into a clinical workflow.

"We can have the best model in the world and it can fail to be useful if we didn't get the integration right," he said. "Bringing the right data, at the right time, in the right way, to a busy healthcare practitioner" is paramount to the success of AI and ML endeavors, he continued.

Buntrock demonstrated a few examples of AI and ML in use at Mayo – and noted that the system had relied on data to shape its COVID-19 response.

"We couldn't have predicted what would have transpired over these last ten months," he said. "We definitely needed data liquidity.

"We as an organization had to react very quickly [in] going after data that gives us better insight into bed management, into personal protective equipment, into staffing. We applied data to different types of contact tracing for employee health, and really started to predict different types of scenarios based on how [COVID-19] evolved as a virus amongst the nation itself," he added.

Still, the presenters stressed the importance of striking a balance between short-term and long-term needs.

"Yes, COVID-19 has created a new set of urgencies," said Gupta. "But the foundational work we're doing is positioning our joint work to move the needle forward" in three key areas. 

"At the end, it's about creating better AI-enabled tools to move clinical research [and] to enhance clinical and operational processes, and taking the patients along with us in their journey," Gupta continued.

For healthcare organizations looking to undertake their own endeavors, she advised, "Start with purpose and people. Start with thinking of building the foundation first, and thinking of security and privacy, not as an afterthought, but having those principles up front, shared broadly with your team members."

Kat Jercich is senior editor of Healthcare IT News.
Twitter: @kjercich
Healthcare IT News is a HIMSS Media publication.