The recent flurry of collaborations between healthcare organizations and Big Tech is a good sign that health systems are finding their footing in AI development.
Healthcare organizations are joining forces with some of technology's heaviest hitters to push AI projects out of the planning stages and into the hospital.
The announcements are, in part, an effort to get in front of the AI hype machine and demonstrate that health systems are putting this technology to work to improve critical issues like workforce stress and administrative overload. The industry doesn't want to repeat the missteps of the EMR rollout, when news stories about bad experiences overwhelmed talk of the positives and hindered EMR adoption and development.
In just the last month:
- Mass General Brigham announced the rollout of an AI algorithm for radiology "that will help increase operations' effectiveness and productivity." The technology was developed in a partnership with GE HealthCare, which agreed to a 10-year collaboration in 2017 "to explore the use of AI across a broad range of diagnostic and treatment paradigms."
- HCA Healthcare announced that its partnership with Google, forged in 2021, had led to the pilot of an AI platform to document emergency department conversations between doctors and patients, and that the two were now testing an AI tool to facilitate nurse handoff reports.
- Duke Health and Microsoft announced the launch of an AI Innovation Lab and Center of Excellence as part of a five-year partnership "aimed at responsibly and ethically harnessing the potential of generative artificial intelligence (AI) and cloud technology to redefine the healthcare landscape."
"Everyone's trying to get ahead of it," says Avishkar Sharma, MD, CIIP, director of AI at Jefferson Einstein, part of the Philadelphia-based Jefferson Health network, which has been working with Aidoc in the radiology space for several years and is considered a leader in that space. "It's an ever-present conversation [in every health system boardroom]."
At the AIMed Global Summit this past June in San Diego—as well as other healthcare conferences like ViVE and HIMSS—the focus on AI was around what many call "low-hanging fruit." To wit, healthcare organizations are looking to use the technology to handle administrative tasks that consume time and energy for staff, including doctors and nurses.
"That's the immediate benefit," says Stephen Motew, MD, MHA, FACS, executive vice president and chief of clinical enterprise at the Virginia-based Inova Health System. "Where are the small, value-added opportunities in our day-to-day operations … that can be made more efficient?"
Indeed, while questions remain around AI governance and policy, health system executives who want to get their foot in the door are launching small programs that use tightly controlled, non-PHI data, finding the benchmarks and the benefits, then moving on to more ambitious projects.
Sharma fits AI adoption into the Gartner Hype Cycle, which charts the maturity, adoption, and social application of technologies. The five stages of that cycle are Innovation Trigger, Peak of Inflated Expectations, Trough of Disillusionment, Slope of Enlightenment, and Plateau of Productivity. He says AI has moved beyond that first stage and sits between the second and third, with health systems looking to find meaningful value beyond the hype and potential.
"We're very much in that turbulent phase," he says.
And that's why these recent announcements are important. They show that health systems are putting skin in the game and moving forward with pilot projects.
Motew says these partnerships are also important at a time when operating margins are thin and health system leaders are hesitant to take on new ideas. Few health systems have the IT talent on hand to make these moves on their own or scale them out to the enterprise.
Furthermore, these partnerships support health systems who are moving their data into the cloud and need help with cloud management.
"This is what everyone is trying to figure out now," he says. "And we want a seat at that table."
Sharma says partnerships are essential to developing and scaling AI programs across the enterprise, but they also have to be nimble. Owing to the evolving nature of the technology, an AI program created now that will use a specific subset of data to address a specific pain point won't be the same program in, say, a year's time. The technology, the data, and the governance around it will mature dramatically.
"You have to build relationships that are ongoing," he says.
Lastly, AI programs coming down the pipeline need to be guided by clinicians. Both Motew and Sharma also say that while the C-suite needs to set safeguards and parameters for AI use, the true value of the technology will be found by those using it.
"We encourage our teams to play around with it," Motew says. "The best ideas are going to come from the people using it every day."
"Clinicians very much need to be in the conversation and in the driver's seat," adds Sharma.
“Where are the small, value-added opportunities in our day-to-day operations … that can be made more efficient?”
— Stephen Motew, MD, MHA, FACS, executive vice president and chief of clinical enterprise, Inova Health
Eric Wicklund is the associate content manager and senior editor for Innovation, Technology, Telehealth, Supply Chain and Pharma for HealthLeaders.
Healthcare organizations are joining forces with companies like Google, Microsoft, and GE to advance pilot projects that use AI in the hospital setting.
These partnerships are a sign that healthcare leaders are moving forward with projects that address key pain points, such as workforce stress and administrative overload.
Hospital executives need to move slowly on AI development, remaining nimble and making sure that clinicians are involved in the process.