CommonSpirit Health leverages data and AI to reimagine health care after the pandemic

The pandemic is forcing us to reimagine what a “good” health care system looks like. For CommonSpirit Health, that includes AI and personalized care. …

Last Chance: Register for Transform, VB’s AI event of the year, hosted online July 15-17.


The pandemic has taxed the health care system in the U.S., and CommonSpirit Health is among the organizations that has experienced a particularly broad array of challenges because of its sizable reach and the diversity of the constituents it serves. “CommonSpirit Health is the largest not-for-profit health care provider in the country,” said Rich Roth, SVP and chief strategic innovation officer, in a keynote at Transform 2020. “We service 21 states, and throughout those states, we are doctors, nurses, hospitals, clinics, health plans, and a number of other things.” The group serves both incredibly rural areas and densely populated urban centers.

Although the burden and urgency that COVID-19 has placed on the health care system has only exacerbated the problems that were already there, Roth allows for some optimism — or at least, can see some redemptive outcomes. “I think what we’re seeing by this ‘fast forward’ button is an incredible blending of different parts of the health care ecosystem,” he said. He explained that the way health care systems are designed, there are these small “worlds” — insurance, pharmacology, social services, and so on — that tend to be distinct and siloed from one another. But now, they’re being forced together, because they need one another in order to “succeed” against the pandemic.

That’s something of a structural shift, but there are other practical ways that technology can play a role in improving patient care, relieve some burden on frontline health care workers, and increase overall efficiency.

Some of it simply requires thinking more holistically about patients and personalizing care. Roth pointed to the case of a patient who had an oncological disorder. The patient needed special medication and couldn’t afford a delay in refilling that prescription, but they were afraid to leave their house, let alone go into a medical facility, because of COVID-19. The company had to understand who in that patient’s life — in this case, maybe a neighbor — could help them navigate the risks and access the care they needed. “There are behavioral health situations there, there are medical health situations there, there’s access to care and transportation — all these things blending together are what’s going to make a difference,” he said.

“AI’s role is critical in that,” Roth added, “because AI can cross so many different elements of care and offer up something that could be best received by patients, blending those different sections together so that the data helps inform the caregivers, or the social workers, or the friends and family in order to get better outcomes.”

Roth offered some examples. For instance, when a patient has to check into the hospital, “we want to make sure to take care of the patient and get them home and return [them] to wellness as soon as possible.” When doctors and nurses are performing rounds, it’s common to go into rooms in numerical order, one after the other. But perhaps there’s a patient with particular needs in room seven, and they need a specific referral in a timely manner to avoid staying an extra day or two in the hospital. “One of the great opportunities for us is to use our AI, and to say [to hospital staff making rounds], ‘You shouldn’t go [to rooms] one-two-three-four, you should actually go to room seven first.’ Because […] based upon our thousands of analysis of patients, [if] we don’t address the patient in room seven first, they’re going to have to stay longer than they would need to otherwise.”

“Using AI that way, really to accelerate our workflow, and to clearly show to our caregivers the clinical benefit of why that data is important, is a great example of how technology can help enhance care,” he said.

Take the example of a connected inhaler that has sensor data to track where, when, and how often a patient uses it. It can be difficult even for an adult to remember and record each of those uses. But, Roth said, “It’s almost impossible to ask a kid, ‘How many times did you use your inhaler today?’ ‘Five, 12, two — I don’t know, maybe it was 35,’ but [with a connected inhaler] you actually have the real data.” From that, he said, you have rich information that can help determine issues as specific as whether a certain room in that patient’s house is causing them a respiratory problem.

Good data can certainly improve patient outcomes and streamline care. But data is hard, and it’s notoriously hard within health care systems. “There’s different systems that hospitals have, that health insurance providers have, that other organizations across the continuum have. It’s very hard to blend that data together,” Roth said.

That’s a big data problem. Though noting that the big data problems in health care need to be solved, and that many people are working on it, Roth said that there are use cases for which small data is ideal. “I think big data and all these streams can work really well for life sciences research, second indication research, working on rare diseases, developing algorithms, etc., but there’s also small data that can be really meaningful in an AI way.” Case in point is connected inhalers. That data may be useful as an anonymized contribution to a larger dataset, but it’s more impactful for that particular patient’s care needs.

But returning to the idea of extracting something redemptive from COVID-19, Roth can envision a better health care system than the one we have today. “I think if what comes out of this horrible pandemic is a focus on vulnerable patients, a focus on the issues in communities that result in these terrible health outcomes on a physiological level — such as heart disease and diabetes, cancer, etc. — then we’re going to be so much better as a country.”

Live Updates for COVID-19 CASES