Patient engagement has long been a tough nut for healthcare to crack, and it’s not for a lack of trying. New technologies that meet the patient in their home offer a clear opportunity to meet patients throughout the course of their daily lives, but the tools themselves are only half of the solution — making a real connection is just as much about addressing and responding to each individual’s mentality.

“We use a unique interface for approaching patients — we put a cute little robot in homes as opposed to a call center of nurses or sending someone out to patients’ homes,” Dr. Cory Kidd, CEO and founder of care engagement startup Catalia Health, told MobiHealthNews. “The reason behind that is not so much about the technology, but the psychology. … That psychology of human face-to-face interaction works really well there, and acts as an effective way to keep patients engaged over time.”

Catalia has come a long way since its first seed funding round back in 2015 and its most recent raise in late 2017. Just last month, its unique approach to in-home engagement yielded the startup a new partnership with Pfizer headlined by a 12-month pilot of its platform among patients receiving specialty drugs.

“One of the things we see in that specialty drugs market in general is there are typically care management programs that are associated with those and paid for by the drug manufacturer. The majority of those look like a call center,” Kidd said. “So at a high level we provide the same kind of program … but instead of a call center of nurses we put our robot, Mabu, into homes and use that to talk to patients every day, and then relay relevant information back to the clinician so they can provider better care. At a higher level it’s a care management program like would be delivered today, but it’s just something much more effective in being able to interact with each and every patient.”

There are a handful of key features and components to Catalia’s AI-driven engagement platform that Kidd will be breaking down next week during a talk at the Connected Health Conference in Boston. Some of these, like Mabu’s voice interface, are “almost expected” nowadays with the advent of personal assistants like Siri, Amazon Alexa and Google Home, he explained. Others that are arguably more important to achieving engagement are the presence of a physical interface toward which patients can direct their attention, he said, and the AI’s capability to tailor conversations to individual patients.

“That’s something that can not only be delivered not only through a robot interface like we use, but through other interfaces as well,” he said. “What’s happening in the background is we’re profiling each patient in a few different ways. Really, a couple of the critical ones are: the psychological model — how do we talk to that person, and how that conversation is different than the next person; and a lot about learning style — understanding how to talk to this particular persona about whatever we’re talking about.

“One comparison there you might think of is a doctor with good bedside manner. What we’re talking about that we don’t mean that she or he has more or less clinical knowledge than the next person, it’s about talking to a patient and connecting the right way.”

Similarly, Kidd stressed the importance of crafting each engagement as a traditional conversation rather than a series of information-based checklists and responses. Mabu, he said, follows the arc of a natural conversation with greetings, small talk and discussions that don’t overstay their welcome if the patient isn’t responding well to drawn out chats. That last point doesn’t lead to longer daily engagement metrics, but arguably can yield better long-term results.

“The metric that we care about isn’t so much on the day-to-day basis, how long they spent talking to Mabu. I don’t care if it’s a few minutes or a few hours, that’s not what’s important. What we care about when managing a chronic disease is what is the overall duration of that engagement. In other words, the first day they talk to Mabu to the last, how long is that engagement? … When you look at chronic engagement and say the apps that have been built, the challenge is people usually use those a few weeks. If we look at outcomes over time, that’s almost zero impact, so we’re looking at engaging people over a period of time.”

Kidd will be giving his keynote on these and other topics regarding psychology-driven AI design on Friday, October 18, at 8:35 a.m. on the Connected Health Conference main stage in Boston.