More on Patient Engagement

How the 'A' in 'AI' can stand for 'Assistive,' and what that means for providers and patients

As AI and machine learning become a bigger component of care delivery, maintaining the human touch will be vitally important.

Jeff Lagasse, Associate Editor

Artificial intelligence and machine learning have the potential to change the healthcare industry a great deal, but as the technology develops and takes root, it's important to retain the human element in healthcare, using these new tools to assist, rather than replace, the human beings who are the engine of the industry.

So should the 'A' in 'AI' stand for "assistive" instead of "artificial"? Chris Gervais, chief technology officer at Kyruus, thinks it should.

In some ways it's simply a manner of terminology. Yet language is a big part of philosophy, and the philosophy here is that in order to serve patients better, AI tools need to augment the patient experience -- and make it easier for providers to do their jobs.

HIMSS20 Digital

Learn on-demand, earn credit, find products and solutions. Get Started >>

"For us, where we're at in healthcare, there's a lot of froth around, 'Well, we use AI to help automate when a patient calls into a call center so they don't have to talk to a human and just get what they want,'" said Gervais.

"Well, what we actually want is to have technologies … that make these humans be able to handle things in a much more human way, because we can make them better and smarter at what they're doing. We're just augmenting the human with extra strength."


On a surface level it make sense that the human element should be retained. But how to meld that human element with this emergent technology?

In Gervais' view, it's about assisting healthcare professionals with the data we already have, which can be done in a number of different ways, including incorporating simpler tools into their apps, and helping them glean the most accurate and relevant information from patients.

An example of what assistive technology looks like would be a patient who begins their clinical encounter interacting with a bot, answering questions, but then getting a "warm" transfer to an agent who can then handle the context, without the patient having to ask those same questions all over again.

The agent can then determine the best care option for the patient based on the need, whether it be with a chronic condition or something more acute.

"It's a low-impact way for patients to describe the patient stuff," Gervais said. "The provider can now focus in the problem and really focus on care quality. It makes the time they're spending so much better. It's not throwing a bot up, it's a handoff point to whee the patient is comfortable engaging."


For providers, the big benefit to a humanistic approach is operational efficiency. Human-centered technologies have the ability to provide data to providers in a rich manner, and also to present patient data in such a way that it's easier to determine which provider is best for them. It's about making it easier for the patient to access the care they need, and relieving staff of the administrative burden that can along with that.

"New tools can have a greater impact because we can get the right patient to the right provider, to have the right appointment and to get the right outcome," said Gervais. And it helps with the provider workload because they're spending more time actually seeing patients.

The human touch is important even for those individuals and entities who are immersed in the digital world, said Gervais.

"It's really important," he said. "I'm a full-on digital lifestyle guy, an early adopter of digital technology. But even me, I really want to talk to a human. It's less about me, and more that I'm coordinating care for my kids, and now also for my parents.

"For the care navigators, it's so important they provide that human touch ... What we're trying to get them to understand is that shouldn't be focusing on the screen. They're focusing on the patient at the other end of the call."

There are some early adopters of tech that's consciously humanizing, but there's some disagreement on how exactly to use it. One of the issues is the quality of data that's needed to train AI effectively. As they get into the nitty gritty of how to train these models, it becomes daunting task and lengthens the timelines for adoption.

But there are incremental steps the industry can take that can lead to broader transformation down the road. For instance, how so show the technology's ROI? One method is to increase customer satisfaction scores, conversion metrics and revenue/operational cost metrics.

Other effective means of implementation may yet come to the fore.

"The future's not written in terms of there being only one way to do this," Gervais said. "As we work with customers who want to prove into this topic really deeply, and as we look at how they want to empower their teams to do better … technology can help them provide that level of service."

Twitter: @JELagasse

Email the writer: