How Ambient AI is Helping One Health System Make Life Easier for its Doctors

St. Mary’s Healthcare launched Suki’s AI tool through the MEDITECH EHR earlier this year. Officials say the technology is ‘life-changing.’

Small health systems and hospitals face the same problems as their larger counterparts, yet they often don’t have the same resources to address them. That’s why an ambient AI tool can be a literal lifeline to sustainability.

At St. Mary’s Healthcare in upstate New York, physicians began using Suki Assistant earlier this year to capture the doctor-patient encounter. The results so far, according to Julie Demaree, executive director of clinical innovation and transformation, have been “life-changing.”

“The big thing I can see is when the doctor says, ‘I can go to the gym after work,’” she says. “I can see that they’re relaxed. They can go in and just focus on the patient. They say, ‘I can just talk to the patient.’”

“We’re accountable to so many stakeholders,” Demaree continues. “We’re trying to document medical decision-making. We’re trying to document to communicate with the patient who’s going to go home and read it in the portal. We’re trying to document with our colleagues. We’re trying to document in case we ever get sued. We’ve just got all this stuff we’re trying to fit into a note. And especially in this rural area, we are their everything. We’re their cardiologist and pulmonologist and their dermatologist. So that visit – it’s complex. We’re covering the gamut. … It is life-changing for them.”

Ambient AI is by far the most popular new tool being embraced by healthcare organizations these days, a statement supported not only by the number of health systems and hospitals using the technology but also the number of vendors coming out with new products. It promises to not only reduce the time spent by clinicians in taking notes and transcribing the relevant data to the medical record, but also improving the accuracy of that interaction and its revenue cycle value by applying the right codes.

Demaree says she demonstrated the AI tool, accessible through the health system’s MEDITECH EHR platform, last fall, at which time she pretended to be an obnoxious parent with a sick child. Roughly 75 physicians were present.

A sometimes overlooked benefit of this tool is that it can be configured to recognize particular words or phrases (and identify the appropriate CPT codes for billing), enabling health systems and hospitals to program the technology for certain departments, like OB/GYN, ENT or the ED, as well as for specialties like oncology. Physicians can also adjust the platform to their specific needs.

“Physicians really have a lot of pride in what they put in their notes and how it looks,” Demaree says. “Everybody has their own style.”

“All they really want is the history and the plan, because that’s what takes them forever to do,” she adds. “Because the EMR has a kind of pre-configured physical [checklist] that they can just load and change what’s abnormal, that’s not really something that they need Suki for. But the story, that patient story, this long, rambling thing, and the plan, where they’ve already told the patient the instructions, but now they have to go back to their office and write, ‘Here was your diagnosis, and here’s what I’m ordering, here’s the plan, here’s what I want you to do. That’s now all captured, and they can just review it and put it in the note.”

Just as important, she says, the technology captures everything in one instance.

 “You’ve already signed your note and now someone tells us you forgot [something], or you have already finished everything and you go to put the diagnosis in and it says that’s not specific enough and now you have [to go back and check your notes].”

And that’s what reduces stress and saves time, reducing the time spent in front of a computer and enabling clinicians to spend more of it in front of the patient.

Demaree says the ROI was almost immediate. Time to document completion dropped by 50%, which reduces the time patients spend waiting for referrals and payers spend waiting for the claim. The percentage of open notes also dropped, while E&M codes and RVUs have gone up—good for the bottom line, though annoying for patients.

That’s also where generative—and, eventually, predictive—AI could make a significant impact. Imagine a tool that not only captures the doctor-patient encounter and applies the correct billing codes, but one that also helps the doctor to diagnose and treat the patient’s medical needs. It would also outline preventive care and wellness options, alternatives to expensive treatments and drugs and links to other resources.

“It can all happen in one big bundle,” says Demaree. “That would be my dream.”

For now, the tool is a critical factor in St. Mary’s efforts to retain and attract physicians.

 “I see this as step one in a huge change for us,” she says. “It’s hard for us to recruit physicians to a small town. It’s also really expensive when we lose physicians. So we’re trying to grow our way to profitability, and to do that you need doctors. And this community needs us.”

“I feel like we’ve done nothing but add things to doctors for years, and now we can we we’re giving something back,” Demaree adds. “I really think this is where things are going to get better.”