Predictive or generative AI: Which will change healthcare the most?

Health system digital leaders largely agree with the results of a recent Becker's poll that found predictive analytics is the technology that holds the most promise for healthcare.

In the July poll on LinkedIn, 41 percent of 1,027 respondents said predictive analytics had the most potential of the four options, followed by generative artificial intelligence (29 percent), wearables (16 percent) and digital health apps (16 percent).

"I find it interesting that predictive analytics, which is AI essentially, and then generative models, which are also AI, combined for 70 percent — which is not surprising at all," said Anthony Chang, MD, chief intelligence and innovation officer of Orange, Calif.-based Children's Hospital of Orange County. "But I also hope people feel that wearables and digital can all be embedded with AI. Essentially, AI is running through the entire thread. It's becoming more and more of a resource that enables other technologies."

He said the current wearables might be considered "dumb" and will only truly become "smart" once they are infused with AI.

"Predictive analytics, in many ways, has the lowest cost of entry and the most practical application for helping solve our biggest challenges right now," said Leah Miller, system senior vice president of divisional health technologies at Chicago-based CommonSpirit Health. "Predictive analytics can help us manage our expenses and gain better efficiencies and outcomes for our patients."

Ms. Miller said she was also "intrigued" that generative AI came in a "solid second."

"Health systems are still learning how to apply this most effectively," she said. "The one area I am confident it can make a difference is in patient experience — laid over our digital experience, it could make patient interaction much more friendly." 

CommonSpirit Health's marketing department, for instance, is using generative AI to draft marketing content, create images and personalize patient communications.

"Under the banner of 'hope is not a strategy,' digital engagement with our patients holds the most practical promise for improved healthcare," said Michael Restuccia, senior vice president and CIO of Philadelphia-based Penn Medicine. "Providing patients the ability to digitally engage in their care via patient portal capabilities, subtle nudges and reminders are scalable and available today with recognized tangible benefits."

Mr. Restuccia said the other technologies may not realize their potential for some time because their scalability and reliability leave much to be desired. "As we are well aware, unreliable data in the healthcare environment is a nonstarter," he said. "Time will tell."

Rochester, Minn.-based Mayo Clinic already has 184 predictive AI models, 18 of which are in clinical use and 35 are in the research stages, said John Halamka, MD, president of Mayo Clinic Platform. The health system is also using generative AI to draft inbox messages for providers and experimenting with it to summarize clinical visits, listen and document ambiently, and automate administrative tasks.

"The scope of digital health is hard to describe. Maybe we should just call it health," Dr. Halamka said. "Each patient's experience must meet the patient at their level of technological literacy and preference. That might include data stewardship of health records by patients and home-based sensors with home testing and home care."

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.


Featured Whitepapers

Featured Webinars