AI-powered scribe tool enables more patients, less burnout

Dr. Kendell Cannon describes how her primary care practice sifted through the hype of artificial intelligence and found a scribe that does what it promises.

Dr. Kendell Cannon, chief medical officer at Herself Health, a primary care practice for women 65 and older

Photo: Dr. Kendell Cannon

A major issue with primary care, especially value-based care, is how much documentation burden there is for providers: It means hours of "pajama time" with the electronic health record after dinner, and hours of more charting at home on weekends. 

THE PROBLEM

It's something that most healthcare providers have come to accept as being just part of the job. However, it is a leading cause of burnout among clinicians.

"The other difficulty is that time spent on documentation is time that could otherwise be spent on delivering patient care," said Dr. Kendell Cannon, chief medical officer at Herself Health, a primary care practice for women 65 and older. 

"A lot of our team at Herself Health learned from hard-earned experience that having scribes to alleviate documentation burden can be a triple win," she said. "Patients have more access to care – more slots available to them – and they have a fully focused provider during their visit. 

"Providers no longer have hours of documentation homework," she added. "Organizations benefit from the economic uplift of provider productivity."

The issue with hiring scribes is that it's operationally expensive, difficult to scale and hard to train, she added.

"Often the turnover of scribes is very high and most leave within the first year; there's also significant variance in quality," she said. "The Herself Health team was trying to overcome these challenges by piloting AI-powered scribing technologies.

"When I joined, they were trying to get DeepScribe to work for their providers," she recalled. "They ran into the same issues I had seen with other AI solutions: They struggled to get providers to adopt the product, because they didn't work as described and did not improve clinicians' quality of life."

Over the past five to 10 years, Cannon had become accustomed to marketing hype that overpromises what the technology can actually do. Herself Health providers weren't happy with the quality of DeepScribe notes, and there was a significant delay in turnaround time (often hours, sometimes the next day), which ultimately interrupted their workflow.

"We actually had to simply stop using it and start hiring scribes, despite the fact we still were paying for DeepScribe," she noted.

Cannon had seen the AI-powered scribe technology from vendor Ambience Healthcare take off like wildfire with providers at her previous practice, Pine Park. Providers loved that it was real-time, embedded deeply in their EHR workflow, and produced high-quality notes – often better than what a scribe would write, she said.

"I had been very skeptical of another AI solution; however, I let them try to work with one of my clinicians who was on the verge of quitting because the documentation burden outweighed her enjoyment of practicing medicine," she recalled.

"I also saw some of my providers who were weeks behind on their charts completely turn it around after onboarding onto Ambience," she continued. "One clinician actually told me she'd thought about quitting because it is hard to do home-based care, but that she can't ever go back to not using Ambience."

So Cannon made it one of her top priorities to bring Ambience to Herself Health as soon as she joined.

PROPOSAL

When Cannon first met with the Ambience team at Pine Park Health, they offered to start with providers who were having the most trouble with documentation. She sent them to one of the providers who was really struggling.

This person wasn't closing out their charts, and staff would have to spend hours on weekends helping the person catch up. Cannon was skeptical the vendor could help, because she had seen so many "AI-powered scribes" that just never worked as advertised.

"That first day after onboarding this provider, we had an all docs meeting," Cannon said. "She could not stop talking about how amazing the program was. She was showing everyone the notes the software had written for her, and within days, she was starting to close her charts on time again. That's when I realized that the vendor team had really cracked it."

MEETING THE CHALLENGE

The workflow for Herself Health providers is simple. Ambience lives on top of the Elation EHR as an extension. It automatically syncs with Elation to know which patient a provider is seeing, and it's just a single click to start capturing the visit.

Once the visit is over, the software generates the documentation pretty much instantly. The provider reviews the note in Ambience's interface, makes any edits they need to, and then clicks a single button to have all of it exported into the right parts of the chart in Elation.

"Because it's so simple, it's drastically reduced how much time it takes to sign off on a chart," Cannon said.

RESULTS

"The speed at which we were able to deploy the product at Herself Health was unprecedented," Cannon reported. "We signed the contract on a Wednesday, everybody was onboarded by Friday, and, by the end of Monday, every provider was using Ambience for every single one of their visits.

"At Pine Park, we were saving so much time on multitasking and post-visit documentation that we could actually open up two more slots each day per provider," she noted. "Providers typically could see up to five patients per day given the complexity and the senior-living, home-based care setting, so adding two slots was a significant increase."

ADVICE FOR OTHERS

"This space has been full of marketing hype for a very long time, with companies claiming they have technology that works or promising that it will be here soon," Cannon advised. "When evaluating AI technologies, I would recommend you try it for yourself before you believe anything. If a vendor is unable to do live, unscripted demos, they're not ready for production.

"You should be able to bring your team to pressure test it across a range of realistic scenarios before deciding to invest any resources," she concluded. "I've seen too many examples of AI companies making promises and wasting a lot of our time and resources because they never ended up delivering on those promises."

Follow Bill's HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.

Women In Health ITResource Center

Stay Informed

Subscribe today to receive our FREE monthly e-newsletter