How Hospitals Use AI To Triage The Triage

artificial intelligence AI chatbot

Unless one is having a medical emergency, going to an ER, in general, is often not the best way for a patient to receive care for more minor ailments. It is expensive, especially for the uninsured, but increasingly for the insured with a high deductible as well. It’s also incredibly time-consuming; ERs by nature of what they are triage patients in order of need, meaning if one is not profusely bleeding, unconscious, actively having a heart attack or a baby, the wait can be long.

And that is just the baseline. In a world where the COVID-19 virus is spreading rapidly, the government, health organizations and medical community have spoken very clearly and in a single voice on this: They do not want patients with non-emergency patients heading to ER, on the fear that the highly contagious virus will spread rapidly through hospitals and vulnerable patient populations.

The problem, GYANT Chief Operating Officer and Co-founder Stefan Behrens told Karen Webster, is that patients are often going to the ER, or calling into the triage nurse at medical centers for 25-minute check-ins — because they don’t know what else to do. Seeking healthcare is the ultimate better-safe-than-sorry decision of a person’s life. If it is hard to navigate, consumers will start congregating in places they associate with treatment, like emergency rooms.

It is what inspired GYANT into business four years ago, during a different global outbreak of the Zika virus.

“Our original vision was to use technology to make healthcare a little bit easier for patients to navigate,” Behrens told Webster, noting that their initial hypothesis was pretty simple: There has to be a way of making it easier to navigate the system to find information or treatment as necessary, with the same tools that were already working in consumer applications.

So they started with “a little AI virtual assistant” to help people understand their risk for Zika, offering up the right information on reducing exposure or finding healthcare options for women of childbearing age that could be affected. That little artificial intelligence (AI) assistant today has grown-up to pair with some very large healthcare providers — and to offer a far wider range of end-to-end services to help consumers get better outcomes via the power of digitizing their care. And, Behrens noted, as COVID-19 is becoming a bigger — and yet still an uncertain issue — it is looking increasingly likely that no-longer-little AI is about to find itself in much, much more extensive use.

The End-To-End Care Assistant

Though GYANT started its life as a consumer-facing program, they moved to a B2B model where their assistant comes to consumers via healthcare providers, insurers and medical centers, Behrens told Webster. By working through providers, he explained, they can integrate into the electronic health record (EHR) and thereby deliver patients to the right care provider as necessary. And then find out what the care outcome: Did their assistant diagnose the situation correctly and did it move the consumer along the right path?

“In order to really move the needle, you have to integrate with the people delivering care and paying for care because that’s the only way to get true adoption,” Behrens said.

Moreover, their assistant then has that connection to the patient so it can continue the care relationship all the way through to resolution. So, Behrens noted as an example, if the medical visit involves a prescription, GYANT’s AI can follow on with a text to make sure it has been filled, or that the customer correctly took it for the entire prescribed course.

But from the patient’s end, that doesn’t feel invasive because it is consistent with all the ways the AI chatbot functions as a triage tool that makes sure the patient ends up in the right place, with the right provider and the proper treatment. That, he noted, is accomplished via a quick Q&A with the AI that fills in the symptoms, finds the background, recommends where the patient should seek help from there and offers to make the appointment. Via that scheduling, Behrens noted, all the information is already provided to the system and moves on with the patient. And, whatever the outcome, he said, the system learns from it to perfect its future recommendations.

Recommendations, he noted, that often advise the customer to stay home. An issue with other telemedicine platforms that have scheduling as a component is that they are overly directive toward that end. GYANT’s AI is set-up to allow patients to “play around with it” and ask it questions about general medical issues, or look for suggestions. Sometimes what a patient needs more than a care provider, Behrens told Webster, is information to assess their conditions better. Sometimes, they need the recommendation to stay home, drink fluid and monitor for improvement or deterioration.

In the age of COVID-19, he noted, they are going to need both.

The Telemedicine Tipping Point 

Their data over the last few years indicates that when consumers use telemedicine — when they talk to the AI — they tend to come back to it. The experience itself has much appeal, he noted, but often getting consumers into the flow of trying it proves to be a hurdle. Or, he noted, it has been, though COVID-19 has started changing that. In the last week, they’ve started to see the usage numbers related to coronavirus and respiratory infections climb along with their use numbers.

“We expect that number to go up significantly in the coming weeks. And we’ve actually been working with our partners to develop a more specific response,” he said.

“We’ve deployed additional questions to our protocols, so people with fever or cough now get questions on their travel history, whether they’ve had any exposure and so on, to make sure that we follow the CDC guidelines. The goal is to make [sure] we can guide people into the appropriate place and in most cases, keep them out of an ER and toward a safe medical encounter or to talk to a provider in case they’re concerned about exposure.”

The “looming epidemic,” he noted, will keep them busy for 2020 — and in a way that he admitted was quite unexpected. And it isn’t’ the only big change on the deck — the firm is rolling out with a new slate of provider partners throughout the year and expanding into a pediatric platform.

However, while COVID-19’s effects on human health remain a mystery still unfolding, he suspects its effects on telemedicine will be salutary. Which, he said, believes will be a boost to patient health all-around.

“This combination of something that is very, very visible that causes people to be afraid … is a  good combination to actually … try these tools out for the very first time. I think if patients have a good experience, it will significantly drive adoption across the board.”