Health insurance provider Humana is hoping to utilize AI technologies as it works to advance its integrated care delivery model with digital tools, Heather Cox, the company’s chief digital health and analytics officer, said during the HIMSS conference on Feb. 13 in Orlando.
Last year, the company launched its digital health and analytics initiative with a goal of using data to create new tools for its services. In an interview with AI in Healthcare, Cox discusses the new initiative, the AI focus, and improving digital experiences for its members.
“I think we’re just really excited about the combination of us, the financier of today, but (also) what we can become as the health partner. That’s an element of what you’re hearing hear at HIMSS. It’s not going to be just about the plan information. It’s going to be about that consumer-mediated exchange,” Cox said.
“That opportunity to use external sources of data—as well as our traditional sources and the EHR data that’s sitting out there today—and combining all of those things with the consumer at the middle. I think we’re super excited about the potential for what that means for, again, the experiences we can create.”
Can you explain your role with Humana and what that entails?
I joined Humana in August 2018 to build this new organization called Digital Health and Analytics. What we’re focused on is really accelerating our journey around integrated care delivery and really hoping to take care outside of the traditional institutional arenas, and helping to enable our providers to provide better integrated care outside of the institution and to connect our members and patients to their providers (and caregivers) more effectively. So, utilizing data to deliver better digital experiences.
What are some upcoming projects?
The first 90 days has really been about learning and understanding the teams that were going to join this organization from within Humana. Also, working with my colleagues to create integrated strategies. (We asked) what can we do to really enable the businesses to deliver this integrated care more effectively, and what was missing from the current agenda that we needed to really double down on?
What I learned was there was not much missing. The good news is Humana has been an extremely innovative company over the course of history and certainly even at this moment. It was really more about how do we get to scale faster in some key arenas. Really, the primary focus is around accelerating machine learning and AI, and really getting tools to market faster for our members and providers.
What are some of ways, through AI and machine learning, you can do that?
We’ve had some really cool test-and-learns and some elements that have been brought into production. A great example would be the sentiment analysis we applied in the contact centers. When members would call in, we would apply machine learning algorithms to understand what the inflections were in (that member’s) voice to get them to the right call center representative to help them with their needs and match the right person to the right member to assess where they were. That’s a great application.
In addition of late, we’ve been testing Alexa’s skills, specifically around folks with congestive heart failure, diabetes and dementia. That has then led to a more recent test where we had an avatar on an iPad. His name is "Buddy," and (it’s) an AI-driven avatar. What’s special about this avatar is about 60 percent of the time it’s AI-driven and 40 percent of the time it’s human-driven. It’s a combination of person and technology, and it’s really there to combat loneliness and to make sure we could also triage care. So when someone needs a little bit of help with anxiety and to combat isolation, you have this avatar there as a partner.
What we found in the trial test we did for 90 days, we were actually able to prevent two ER visits through actually managing anxiety and helping someone talk their worries and getting them calmed down enough so they didn’t actually have to go to the ER, which was just a great outcome and result. We’re also finding that people fall in love with Buddy, and it’s like a family member. (We're) using technology to create real relationships and interactions that weren’t there before.
We’re also finding the value in AI-driven chatbots. Now for this one, we actually created tests around Medicare education and personalized shopping for Medicare plans. But what we’re finding is while simple in theory and concept, (people ask) much more nuanced sets of questions and they get very complicated very quickly. It started out simple, then we actually had to have humans come in and sort of intercede. Test-and-learns (have) failed, but that we’re learning from (them), so it’s actually been great.
AI and machine learning are huge topics in healthcare. What do you think AI’s role could be within the insurance side of the industry?
We see great hope. We can’t actually talk about specifics that we have coming, but we have some really cool offerings that we’ll see coming to market in the next year that from a plan perspective we think will be very exciting around subsegments of the Medicare population. I think this is really about creating those new experiences that we think are going to be the differentiator for us in the marketplace.
As we think plans will become much more of the commodity element—especially with some of the announcements that happened within the last 24 hours—I think the ability to use data and to apply those algorithms via AI to create the experiences that’s going to be our opportunity to really differentiate as a plan provider. We’re super excited about what we have on the agenda.