Artificial intelligence and its potential to change healthcare

A panel of physicians and leaders in the field expressed enthusiasm for AI’s possible benefits for patients. They also said solutions must be designed with health equity in mind.

Many have hailed the potential of artificial intelligence to transform healthcare.

Michael Howell, Google’s chief clinical officer and deputy chief health officer, says, “It’s hard to imagine a technology that is more hyped than AI.”

Even so, Stephen Parodi, executive vice president of The Permanente Federation, says, “Widespread AI use in healthcare is still in its infancy.”

Still, many are projecting significant growth in the prevalence of AI in medicine in the near future.

During a one-hour forum hosted by The Permanente Federation Monday, healthcare leaders, all physicians, assessed the possibilities of AI, the keys to success, and expectations on its future uses.

Even in a forum where leaders talked about potential challenges, including designing technology with patients in mind and the urgent need to focus on equity, the participants spoke with enthusiasm, even excitement, about the growing role of artificial intelligence in medicine.

It’s appropriate to bring some healthy skepticism and ask questions about the potential of artificial intelligence in healthcare, Howell said.

However, Howell also said he expected, “AI will do things we didn’t think were possible.”

Earlier interventions

Edward Lee, executive vice president and chief information officer of The Permanente Federation, talked about how AI is being used across the Kaiser Permanente system.

At Kaiser Permanente, researchers have used AI to examine retinal images of patients with diabetes, to possibly determine if patients are more likely to lose their vision, Lee said.

In addition, Kaiser Permanente is using AI-powered models to analyze which patients in hospitals may be at higher risk of deteriorating or could require intensive care.  "This gives us a chance to intervene before patients get sicker,” Lee said.

Hundreds of patients have likely been saved, he said, “and that’s a conservative estimate.”

The system is using AI to analyze emails to make sure they are getting to the right member of the care team. “This helps our patients get timely responses to their health concerns,” Lee said.

John Halamka, president of Mayo Clinic Platform, said he expected that within the next six quarters, artificial intelligence is going to be brought into the workflow of electronic health records.

The Mayo Clinic has been increasingly using AI in research. Mayo Clinic researchers have been studying the use of artificial intelligence to identify pregnant patients who may be at risk for complications, as well as patients who could have greater likelihood of suffering a stroke.

When asked about when AI would gain greater prevalence, Halamka cited the author William Gibson, who once said, “The future is already here, it’s just not evenly distributed.”

“I believe the perfect storm for innovation requires technology that’s good enough, policy that’s enabling and cultural change that creates a sense of urgency,” Halamka said.

Patients have greater expectations of healthcare, and that will help expand the use of AI in medicine, panelists said. “The cultural demands of our patients will drive us forward,” Halamka added.

Google Health is using artificial intelligence to bring better technology to care teams, and also in reaching out to consumers when they’re searching for health information online, steering them to relevant and accurate results and away from misinformation, Howell said. The tech giant is also using AI in community context, he said, such as better projections of flood threats.

Vivian Lee, president of health platforms at Verily, a sister company of Google, talked about the use of AI algorithms to identify patients at higher risk of hypertension, substance use, or a longer hospital stay. She said the goal is getting that information “to the clinicians to make that data more actionable.”

Artificial intelligence also presents opportunities to engage patients in different ways, and that goes beyond just personalized medicine, Vivian Lee said. With AI, she said the question becomes, “How do we move to precision health and precision engagement?”

“I really believe the advances we are making now will enable us to do personalized care at scale,” Vivian Lee said.

During the forum, participants, including the audience weighed in on where AI would have the most potential to improve healthcare. Most said it would be the use of artificial intelligence to predict potential health risks.

“I think the thing about risk prediction is it can affect not only individual patients … it can affect entire populations, entire communities,” Edward Lee said. “We can positively contribute to the health of many, many patients.”

Focusing on health equity

Even as the panelists touted AI’s promise, they also said health systems aiming to use artificial intelligence must focus on closing healthcare disparities.

“There is deep evidence that care that isn’t equitable just isn’t high quality,” Howell said.

“Everyone should have the opportunity to receive the full benefits of AI … We should work systematically to make sure that happens," he said.

Researchers are using artificial intelligence to predict risks in patients, but as Howell noted, the problem is some data is missing when it comes to patients from underrepresented communities. In a sense, disparities can be baked into the data being analyzed.

Vivian Lee shared similar concerns. “We need to be attentive to bias and health equity,” she said.

Fatima Paruk, chief health officer and senior vice president of Salesforce, said AI could be both an enabler or a barrier. But she said, “It leaves me thinking we can deliver more equitable care.”

The technology of AI in and of itself is only so useful, Edward Lee said.

“Combining with expertise is when you can really make a difference in the lives of the patients,” he said.

The panel’s members said they were hopeful in part because much of the research in AI and the new artificials intelligence are being developed by those in the healthcare industry.

Paruk touted AI’s potential, combined with remote patient monitoring, in helping older patients potentially live at home longer. Health systems could eventually use data to get a sense of when those older patients may need more assistance.

That would also be a boon to many in the “sandwich generation”, who are caring for both their children and aging parents. “There’s a huge amount of potential there,” she said.

While panel members noted similar predictions about electronic medical records reducing demands on physicians, Paruk and others said AI could reduce burnout among clinicians.

But ultimately, the panel members expressed the most enthusiasm for how artificial intelligence could transform patient care.

“I’m incredibly hopeful for the future,” Paruk said.