AI alleviates burnout, reduces documentation time by 72% in primary care

Key takeaways:

• Voice-enabled technology helps physicians complete documentation more quickly.
• Other technology compiles data in a way that makes it easier for physicians to sift through, reducing burnout.

Artificial intelligence has had a major impact in reducing burnout in primary care by speeding up administrative duties, experts said during a webinar held by the American Academy of Family Physicians.

Steven Waldren, MD, vice president and chief medical informatics officer of the AAFP, said that, if implemented correctly, artificial intelligence (AI) could be “a great, disruptive innovation for primary care.”

Waldren discussed some of the key, existential problems facing family medicine, including administrative burnout, an undervaluing of primary care and threats to the scope of practice.

Electronic health records documentation

“Administrative burnout and burden is an epidemic that started before the pandemic. Over 50% of family physicians report some level of burnout in their practice today,” Waldren said. “Admin burden is the biggest problem in our specialty of family medicine.”

Physicians report that electronic health record documentation makes them feel constantly rushed and that they often work overtime to stay caught up, Waldren explained. So, the AAFP looked into an AI assistant for documentation, and found one in Suki.

The voice-enabled technology helps physicians complete their documentation more quickly, Waldren said, and they found that this could be “an essential innovation for family medicine.”

He added that 132 physicians tried out the technology and reported a 72% reduction in their documentation time per note — translating to a calculated time savings of 3.3 hours per week per clinician, according to the AAFP lab report. After the trial, 60% opted to become paying customers and use the solution moving forward.

“When we talk with our members who are adopting these types of solutions, they talked to us about it being kind of a breakthrough,” Waldren said. “We saw them actually improve in regard to their practice satisfaction and their burnout scores just by using one particular tool focused on one particular administrative burden.”

Jarrett Dodd, MD, a family medicine specialist with Central Virginia Family Physicians Medical Group, said that he has been using Suki in his practice.

“It was really almost like going back to how I used to document ... with a Dictaphone. We had a transcriptionist that would take care of all that for us. We would just speak our notes into the Dictaphone, and Suki just sort of brought that back except a whole different level.”

Now, he said, they have no transcriptionists, and the dictation automatically goes into the record, allowing him to “capture a much richer note” and “communicate our medical decision making in a much fuller way.”

“The nice thing about Suki is you can talk to your medical decision-making right into the chart,” he said.

Clinical review

The other major administrative burden that physicians face is chart review “and being able to report on those types of things if you’re looking to be part of some type of value-based payment arrangement,” Waldren said.

“How do you make sure that you got all the data for this particular patient? What’s new since the last time you saw this patient, especially if they’ve been in the hospital or they have seen a couple of subspecialists? Then you only have a few minutes with patients, so what’s the most pertinent thing for today?” Waldren said. “Just trying to find all that data can take up the entire amount of time that you would have scheduled for the visit.”

So, the AAFP looked into an AI assistant for clinical review that can “pull the data together in a problem-oriented manner and give you a snapshot of exactly what’s going on with your patient without having to search and click and find things.”

The AI assistant, from Navina, “allows the clinician to focus on providing care instead of data mining,” Yair Lewis, MD, PhD, senior vice president of medical at Navina, said during the webinar.

“Navina allows you to get to know your patient really well in a fraction of the time,” Lewis said. “Our algorithms are able to ingest multiple data sources, including different types of EHR data, structured data, unstructured information, scanned documents, data coming in from health information exchanges like [Consolidated Clinical Document Architecture documents], claims information, and then we use AI to create the clinical portrait of the patient.”

Though just 10 physicians participated in the initial lab looking at this solution, Waldren said that they saw a 60% reduction in the physician’s time to review these notes —this number didn't include the review time for nurses and other staff members — and a 25% increase in diagnoses that, while they may have been known to the clinician, were not captured in the EHR and therefore were not sent to the insurance company to make sure they had appropriate risk scores.

“Two of these practices actually had done a lot of value-based work, and even though they had done a lot of value-based work, they will see a 37% increase in the risk scores that they had,” he said.

Waldren additionally noted “a significant drop in burnout” because physicians no longer had the cognitive burden of pulling all of the data together and ensuring that everything is reported to the appropriate parties.

Dodd has also been using Navina since October and said his group recently signed a contract to roll it out to other physicians.

“What Navina I think really offers us is the opportunity to really streamline how we dig around in the chart,” he said.

Dodd described how his practice receives records from a variety of care centers, like hospitals and specialists, that are all in different formats and rarely labeled correctly.

“When you try to search for things, it is just a bear. Navina takes all of that and filters through all that, brings out the most recent and the most pertinent documents, puts it in a very easy to find location, and it just makes it that much simpler,” he said. “I had a hospital transition of care visit earlier today, opened Navina, and there's discharge summary right there. I didn't have to go looking for it, I didn't have to have my nurse go looking for it; it was right there, highlighted, ready to go. That’s a great benefit when we’re slammed.”

Dodd also mentioned the ability to search through unstructured data. He said that in his practice’s usual method, it is impossible to search faxed reports because they are scanned as images, but in Navina, he can search, for example, “aorta” and find every single document in the chart — structured or unstructured — that has “aorta” written there.

“Many of my partners have had their worlds changed by being able to get home sooner, be with their families and not have to spend so much time dictating or documenting,” he said of both technologies.

References:

• AI assistant for clinical review to reduce burden and improve quality and value-based care outcomes. https://www.aafp.org/dam/AAFP/documents/practice_management/innovation_lab/report-navina-ai-clinical-review-phase-1.pdf. Accessed March 13, 2023.
• Using an AI assistant to reduce documentation burden in family medicine. https://www.aafp.org/dam/AAFP/documents/practice_management/innovation_lab/report-suki-assistant-documentation-burden.pdf. Accessed March 13, 2023.
• Waldren S, et al. How AI is Transforming Primary Care from Automation to Insight and Beyond. March 14, 2023; (webinar).