AI, Chatbots & Natural Language Processing: The Present & Future of Digital Healthcare – with Greg Fitzgerald

For this episode of The Innovation Engine podcast, we take a look at what the future of digital healthcare may hold for both patients and providers. We cover the role AI and chatbots will have in the future (and are already having, in some cases), some of the inherent challenges in innovating in the healthcare space, and how one leader at a Fortune 100 company in the healthcare space views the intersection of digital and healthcare.

Joining us for this episode to talk about those topics and more is Greg Fitzgerald, the Staff Vice President of Digital Product Engineering at Centene. Greg is responsible for the strategic direction and development of the consumer technology products emerging from the Digital Product Engineering team at Centene. His career has included leadership positions in a number of successful healthcare startups, including running engineering teams for LiveHealthier, a digital wellness company acquired by Centene, and HealthCentral, a social community of patients and experts acquired by Remedy Health Media.

LISTEN TO THE EPISODE

You can tune in to the full episode using the SoundCloud embed below.

EPISODE HIGHLIGHTS

Artificial Intelligence & Conversational Agents

  • We discuss a whitepaper Greg co-authored on Conversational AI and Healthcare along with Amy Troop: “Artificial Intelligence and Conversational Agents in Healthcare”
  • The first question they wanted to answer, Greg says, was do people want to even interact with something like this in the healthcare space? It’s one thing to ask Alexa to turn on your lights. It’s another thing to ask Alexa to book a doctor’s appointment for you or ask, “Hey, what’s this thing on my foot?” What they found is that, yes, “people are very interested in interacting with this type of experience in the healthcare space.”
  • From there, they looked at what types of things are people interested in asking about? What they found is they got a lot of queries; everything from “How do I add a dependent to my plan?” to “Is my prescription for Xeljans covered by my insurance?”
  • Finally, can a conversational experience actually communicate information better than a traditional web experience, or a paper packet that might be sent in the mail? “And, again, I think the answer turned out to be yes – to an extent,” Greg says. “And so, one of the learnings that we took away was conversational experience is really good for communicating information. But as soon as you get to anything more complex, or maybe the things that require complex data collection, you actually need to drop down into a mixed user interaction,” or something that has more traditional non-conversational aspects to it.

AI & Natural Language Processing in action

  • Colloquium is a platform at Centene that serves as the foundation for all of their conversational experiences.
  • It is a HIPAA-compliant natural language framework that allows Centene to train domain-specific models within the healthcare space. So, there are models for things like answering questions about insurance and doing triaging and diagnosis.
  • There is also a framework and administrative system for authoring the responses to the queries that come in via NLU.
  • “One of the things that we’ve really focused on at Centene and with this product is how do we make sure that we have the ability to understand what the user is saying but, also, the ability to make sure we’re providing the correct answer?” Greg says.

But is implementing these natural language systems actually helping anyone?

In a word, yes.

  • There was an 18% improvement in binder payment rates, which is very significant because it means “18% more people are going to be able to use their insurance when they first go to in the new year than had Amber not been present.”
  • There was a 40% increase in PCP selection rates as compared to non-Amber users.
    And when they did some qualitative surveying of their members, they saw a significant increase from Amber users, almost a full point on a five-point Likert scale, on topics like, “I understand the basics of how my coverage work.”
  • Finally, they saw a 20% increase in respondents’ likelihood to be satisfied with their choice of Ambetter.

Resources:

Leave a Reply

Related Posts

CES 2019 Recap: AI for AI’s Sake, Mobility as Transpor... For this episode of The Innovation Engine podcast, we take a look back at the year’s biggest consumer technology show, CES, which just wrapped in Las ...
What’s Next in AI? The One Thing I’m Looking for... Upon returning from CES last year, I declared IoT dead. When smart razors and connected breast pumps litter the showroom floor, I think it’s safe to s...
AI & Machine Learning Will See You Now, and Other Takea... A 3Pillar team and I spent a few days in Santa Clara recently for the 12th annual Health 2.0 Conference. As usual, we spent some time after the confer...
For Whom the Bell Tolls in Healthcare? Well, Everyone. Do you know for whom the the bell tolls? I'll cut to the answer: it's healthcare companies. Yes, all of them.The long-overdue announcement from Be...
The Pulse of Digital Health at CES 2018 I got the chance to attend CES a few weeks ago and wow… as a first time CES-er, I’ve never experienced anything like it. Between the stellar speakers,...