Artificial Intelligence has the potential to greatly improve clinical care and support the transition to value-based care, but also comes with risks. So what role should AI play in value-based care models? The question was posed last week during a panel discussion at CES 2024 that was moderated by Dr. Jesse Ehrenfeld, president of the American Medical Association.

Chris Jagmin, vice president and chief medical officer of state programs and medical policy at CVS Health’s Aetna, said it’s important to use AI in areas where there is little risk. This includes clinical decision-making and approving something as an insurer. He added that AI presents opportunities in areas that “drive us crazy,” such as automating call centers or automating claims payment.

However, “denying something based upon AI without human touch I think is wrong for clinical,” Jagmin said.

Dr. Maria Ansari, CEO and executive director of The Permanente Medical Group, echoed Jagmin’s comments, stressing the importance of having a human touch when using AI.

“It’s the physician who’s treating the patient, not AI,” Ansari said. “AI [should be used] to augment our work in clinical decision support tools.”

She added that in a value-based care model, AI should be used responsibly to predict who needs an additional level of care, such as who is at risk of having a stroke or a heart attack. 

This is especially important coming back from the Covid-19 pandemic since so many people weren’t coming in for preventive care, Ansari noted. For example, she said that The Permanente Medical Group had a two-year backlog of women who didn’t come in for a mammogram during Covid-19.

“How do you prioritize that? … Instead of just using traditional risk factors for who might be at risk for breast cancer, we used AI to look at their old mammograms, which were read as normal, but picked up 1000 data points that would actually predict who’s going to go on to have cancer and prioritized those patients,” Ansari said. “That prediction model turned out to be 60% more accurate than traditional risk factors. 

“I think we can use AI when we talk about value-based care to improve health outcomes in the right way that’s responsible and not denying care, but actually enhancing our care,” she continued.

The conversation came after President Joe Biden signed an executive order in October on the safe development and use of AI. In December, the White House announced that 28 providers and payers made commitments for responsible use of AI in healthcare. These organizations include CVS Health, Mass General Brigham, Premera Blue Cross and Sanford Health.

Photo: metamorworks, Getty Images