Can healthcare leaders bridge the gap between care management and actually influencing healthcare consumers with chronic illnesses, in their daily lives? That question was the subject of a fascinating educational session Thursday afternoon at HIMSS16 in Las Vegas. “Applying Cognitive Computing to Population Health,” with Pam Nicholson, senior vice president for strategic integration at Denver’s Centura Health, and Jeff Margolis, CEO of the Denver-based Welltok (and formerly the founder of Trizetto). Nicolson and Margolis described the unique multi-way partnership created by leaders of Centura Health, the x-based Welltok, and IBM, through its IBM Watson division.
“Our view is that there’s quite a difference between patients and consumers, particularly in how you provide information to them,” Margolis said early on in the presentation. “As we move towards a value-based model in healthcare, physicians are being paid for value or performance, and that’s considered acceptable. Should consumers be paid for performance, too?” he asked. In response to an instant survey, 96 percent of attendees said yes, while 4 percent said no.
The initiative that Margolis and Nicholson described was a project to create what is called CafeWell, a consumer-facing app that helps guide consumers who have chronic illnesses through lifestyle choices, including asking basic health questions and selecting healthy food and restaurant choices, among other choices. As Margolis emphasized, the decision was made at the outset not to try to provide actual clinical or medical advice. At the same time, the 60 consumers who participated in the pilot project around CafeWell are all active patients in the healthcare system.
Margolis contexted the CafeWell project by saying, “Most of what we call healthcare is really sick care. We return people to their green line”—the health status that a patient enjoyed before a major illness or injury cared for by the healthcare system—“and that’s important, and we have to do it” as a healthcare system. I was born with Crohn’s disease; it is genetic,” he disclosed. “I’ve dealt with it my whole life, and have spent 100 days as an inpatient.” Gesturing to a line on a projected slide, he said, “This is my morbidity curve; we all have a natural morbidity curve. But the whole idea of health optimization is to bend the morbidity curve of individuals.”
Indeed, Nicholson noted, “In a given population, somewhere between 15 and 20 percent of the population procures roughly 80 percent of the cost.” What’s more, she noted, while perhaps 20 percent of the morbidity involved is genetic and 20 percent is created or influenced by the larger environment, experts believe that fully 50 percent of any individual’s relative morbidity can be accounted for by specific lifestyle choices that that individual makes on a day-to-day basis, such as whether the person exercises and what kinds of foods he or she eats, etc.
Given that, Nicholson, Margolis, and their colleagues at the three organizations set out to leverage technology to address the gap between what healthcare providers can offer to patients with chronic illnesses through formal care management programs, and the countless choices that those patients need to make as consumers in their day-to-day lives.
“Through cognitive computing,” Margolis noted, “the consumer can ask a question, and NLP-facilitated technology can give them an answer, not a list. Cognitive computing can understand and answer questions. At Welltok, we’ve spent the last several years on the leading edge of teaching Watson how to learn to respond to questions intelligently. Watson learns the way a child might; you teach it patterns. You have it learn question-and-answer scenarios.”
“We wanted to impact the spiritual, physical, mental, and intellectual aspects of people’s health,” Nicholson added.
So what happened was that “We fed into Watson information on healthy dining,” Margolis noted. Now, Watson can give an individual information on heart-healthy restaurants wherever that person happens to be. “If you ask about heart health, it will give you low-sodium dining choices first,” he added. “Or you could ask about low-cholesterol or gluten-free choices,” too, he noted.
The presenters played a video for the audience of an interview with a young Colorado man and Centura patient named J.D. Schlapp, a very healthy-looking person who had had multiple cardiac stents implanted after it was discovered that one of his cardiac arteries was 99-percent blocked and the other was 97-percent blocked. One of the 60 patients who helped to test CafeWell, Schlapp described how delighted he was to be able to ask about nearby restaurants and instantly be given heart-healthy choices. As Nicholson noted, not only is J.D. eating far fewer greasy hamburgers and drinking far less sugary Mountain Dew soda as he was did; he is also deeply engaged in his own health, and mobile technology is facilitating that engagement.
The initial pilot project around this at Centura Health was purposely limited in terms of the number of participants; in fact, just 60 patients were a part of the pilot. But it is being expanded to 60,000 patients this year, and will be continually enlarged as an initiative.
What have been some of the organizational learnings at Centura around this? “The first learning,” Nicholson said, “is that we looked around and suddenly asked ourselves, where are the physicians? And so we created a physician council that includes 100 physicians, so that we could engage them to encourage them to use the app with their patients. Second,” she said, “hospitals do not own their own [healthcare consumer] content. I was certain that tons of content could be downloaded, but that turned out not to be true. Fortunately, we were able to bring the American Heart Association in as a content partner; and another company, emmi [a Chicago-based patient engagement solutions provider] , had lots of great videos to load. And Welltok itself had lots of knowledge and learning and content.”
Further, she said, “Finally, the biggest learning has been around timing. I thought that we could build this very quickly. We did develop a common vision early on, with all the different partners at the table. But it took longer than expected. The idea of a three-to-six month timeline”—which they had initially envisioned—“got blown. “What has helped us, though, is that we’ve kept coming back to the vision of technology that would be easy and that consumers would want to use.”
In addition, Nicholson said, “I can’t show you outcomes around reduced total cost of care, reduced readmissions, and so on. But when we look at J.D. and how he’s changing his health behaviors, it’s clear to me that this program is having an impact. And I think we need to change our mindset on what we’re looking to do here.”
Margolis noted that “We’re writing different software for Centura” than the company’s standard commercial software. “And it’s one software program, SAS-based and cloud-based. It’s been hard to do. But it’s configurable, so that the content, and recommended health itineraries as we call them, could be applied anywhere. This is an interesting use case for us to build together. Staying on the pure consumer side has far fewer challenges. Crossing over to the setting where you have heart disease with pre-acute and post-acute things going on, it was good to see that you could keep a good flow going on beyond the confines of formal medical care. We know that this is like a sliver in the whole idea of a health optimization platform. But what makes health optimization different? And what makes population segmentation different? If we do these things together, we’ll make a difference in healthcare.”
