Which Clinical Documentation AI Tools Are Health Systems Using?

Internal medicine physician Dr. Missy Scalise can recall the darkest day of her career vividly.

“I was taking my son to his five year-old annual physical and the doctor was asking him the typical sort of questions like, ‘What are you doing in school?’ or ‘What are you excited about?’ Then the doctor asked him, ‘What does your mom play with when she’s with you?’ and my son answered, ‘She works on her computer.’ That just about broke me — I’m still sad about it,” Dr. Scalise said.

That “work on her computer” involved clinical documentation required after each patient visit. 

Dr. Scalise began searching for ways to be able to spend more time with her family after that appointment. She realized the only way she could do that while still completing all her clinical documentation tasks was to get up at 5 a.m.so she could complete her notes before her family woke up. For years, she spent her early morning hours doing administrative work — until nine months ago, when her health system began leveraging an AI assistant to reduce clinicians’ documentation burden.

Not surprisingly, Dr. Scalise — who serves as program director of the internal medicine residency program and chair of the clinician wellbeing committee at Ascension Saint Thomas in Tennessee — believes that these AI-powered documentation tools will become ubiquitous among healthcare providers in the next five years. Three other health system leaders who were interviewed this month echoed her feelings.

To them, clinician documentation assistants are one of the most valuable AI tools that hospitals can adopt because they help both clinicians and patients by alleviating burnout and allowing for more meaningful interactions.

Ascension Saint Thomas & Suki

Dr. Scalise began using an AI documentation tool earlier this year as part of her health system’s pilot program with Suki. This month, Ascension Saint Thomas announced its plans to integrate Suki’s AI assistant into its residency program as part of a greater system-wide rollout.

Redwood City, California-based Suki seeks to lighten the documentation burden through its AI-powered voice assistant for physicians. By calling out to Suki’s AI assistant, a physician can quickly access key information about their patient, such as their medications, vital signs, allergies or surgical histories. Physicians can also use Suki’s tool to do things like dictate clinical notes, pull up their weekly schedule and assist with ICD-10 coding. 

Suki’s assistant can also generate clinical notes from ambiently listening to a conversation between a physician and their patient. These notes are automatically sent back to the patient’s EHR so that the physician can review and make any necessary changes before finalizing it.

As a result of using Suki, Dr. Scalise has won back precious hours in the day, which she can now spend with her family. Before the technology was introduced, it used to take hours to do clinical notes manually, she said.

“I’Il ask the patient if it’s okay that I use [the tool] during a clinic visit at the very beginning. I use the app on my phone, put it to the side, turn it on, and it listens to our whole conversation. At the end of the visit, I just tap to create a note, and it spends a couple minutes calculating the note. I’ve often moved on to the next patient, so at the end of the day, I review all those notes — and they’re usually accurate. I probably spend 15 minutes editing,” Dr. Scalise explained.

She said that Suki’s tool stood out to her health system because of how easily the technology can integrate into its Athenahealth EHR. It typically takes about five business days to integrate the AI assistant into a health system’s Athenahealth EHR.

Dr. Scalise pointed out that some AI tools require clinicians to copy and paste generated notes into the EHR. With Suki, clinicians are able to edit and finalize notes directly in the EHR, which reduces their number of overall clicks.

Suki also has integrations with other EHRs, including Epic, Cerner and Elation Health, according to the company’s website.

In an interview last year, Suki CEO Punit Soni said that his company’s EHR integrations are a major part of its overall strategy.

“In healthcare, it is not AI that makes you successful. It’s EHR integrations that make you successful.” he told MedCity News. “The number of engineers at Suki that work on [EHR integration] is probably twice the number of engineers who work on AI.”

Ochsner Health & DeepScribe

Last month, Ochsner Health announced a system-wide deployment of a different documentation tool. The New Orleans-based health system is rolling out DeepScribe’s AI assistant to its 4,700 employed and affiliated physicians.

Like Suki, San Francisco-based DeepScribe is integrated with several EHRs, including Epic and Athenahealth. The company’s tool ambiently listens to the patient-provider interaction and provides a transcript that captures the dialogue. It also produces clinical note drafts that clinicians can edit and submit in the EHR.

Ochsner knew it had to do something to alleviate the massive burden associated with clinician documentation, said Dr. Jason Hill, MD, the health system’s innovation officer.

“In most industries, data entry is generally relegated to individuals that are on the lowest-skill part of the food chain. Healthcare is unique. The most highly skilled people do the data entry — and it’s a lot of data entry,” he remarked.

He added that most physicians will candidly tell you that documentation is their least favorite part of the day. 

While most doctors loathe the clinical documentation process, they also recognize that it’s important, Dr. Hill noted. He said that many performance and quality ratings are based on what is found in clinical notes.

“We live in an age where we have giant health records that can tell you everything about patients. But for some reason, everybody wants to look at the doctor’s note to know how a patient is doing,” he explained. “[Documentation] has become a huge thing that serves a lot of masters. As a doctor, you want to show that you’re doing your work, but some doctors are better at showing their work than others.”

After a long day, it’s usually difficult for a physician to recall specific interactions with one of the dozens of patients they saw that day, Dr. Hill added.

When searching for a tool to solve this problem, Ochsner piloted DeepScribe’s tool and another one that Dr. Hill declined to name. 

He said DeepScribe’s technology saw better adoption and satisfaction rates across the health system’s clinicians. Dr. Hill also noted that DeepScribe did a better job of collaborating with Ochsner and responding to clinicians’ feedback during the pilot process.

Clinicians tell him often about how much of a positive difference DeepScribe’s tool has had on their jobs.

“In my history of being a clinical informaticist, I’ve done a lot of projects in it. I have never, in any of my projects, had doctors take a film themselves telling me how amazing something was and then send it to my inbox unsolicited. We had physicians in our pilot take selfie videos saying, ‘This is the best piece of technology I’ve ever had at my disposal.’ Usually it’s the opposite experience — usually doctors are asking, ‘What are you doing to me?,’” Dr. Hill remarked, referring to other technology that providers may perceive to be burdensome.

Kaiser Permanente & Abridge

Just last week, Kaiser Permanente unveiled plans for a big documentation AI rollout as well. The health system will be deploying Abridge’s clinical documentation tool across its 40 hospitals and more than 600 medical offices in eight states.

Abridge was founded in 2018 by a UPMC cardiologist and two Carnegie Mellon University researchers. Its tool, which is integrated with Epic, listens in on appointments, creates a transcript and generates a clinical note for physicians to edit in the EHR.

Desiree Gandrup-Dupre — senior vice president of care delivery technology services at Kaiser Permanente — said that the health system gathered feedback from both physicians and patients during its Abridge pilot. 

“A lot of physicians have made comments saying that for the first time in their careers, they were finishing their charting by the end of the clinic, versus having to complete it after hours,” she noted.

As for patients, they reported “feeling so much more engaged” with their providers, Gandrup-Dupre added. Before adding ambient listening to visits, patients often felt like their physician wasn’t listening to them because they spent the whole appointment sitting in front of a computer typing, she pointed out.

Abridge’s technology stood out as the best tool for Kaiser to adopt for three main reasons. The first is that the tool was scalable and easy to integrate across the health system’s enterprise, and the second was that clinicians gave overwhelmingly positive feedback during the pilot, Gandrup-Dupre said. 

Additionally, Abridge’s AI did well on Kaiser Permanente’s verification testing, she noted. The health system had its quality assurance team perform a series of tests on the AI so it could independently assess if its clinicians felt comfortable with the quality of the transcripts and generated notes.

The Ohio State University Wexner Medical Center & Microsoft DAX

An enterprise-wide deployment may also soon be in the cards for The Ohio State University Wexner Medical Center. This week, the health system announced that it has started piloting Microsoft Dragon Ambient eXperience (DAX).

Microsoft’s DAX tool came with its 2022 acquisition of Nuance. The AI-powered tool, which is integrated with EHRs like Epic and Cerner, creates clinical notes by listening to patient-provider conversations and generates structured documentation in real-time.

Internal audits show that the tool currently saves Ohio State clinicians about 30 minutes to an hour per day, said Dr. Harrison Jackson, one of the doctors participating in the program. 

To him, this is a welcome benefit but not the main goal of adopting Microsoft DAX. The goal for Ohio State’s pilot is less about saving clinicians’ time and more about improving the quality of the physician-patient relationship, he declared.

“We’re making more eye contact, and our body posture shifts away from the computer and faces the patient. We’re able to pick up on more nonverbal cues,” Dr. Jackson explained. “It’s much more clear that we are completely focused on our patients, and that pays such benefits in the quality of information we get.”

He said Microsoft DAX stood out as the right tool for Ohio State due to its ease of use. In his opinion, most clinicians should be able to use the tool after a two-minute instructional video.

Ohio State has already run 5,000 patient encounters through DAX since July 1, Dr. Jackson pointed out. He added that the health system has also quickly expanded the program from 25 initial users to 500.

In Dr. Jackson’s view, all hospitals should deploy clinical documentation assistants as a service to their patients.

“And anytime we’re serving our patients better, our job satisfaction goes up,” he declared.

AI tools are proliferating in multiple areas within healthcare, but when it comes to clinical documentation, the verdict is in: AI makes doctors more efficient and satisfied, which ultimately will have a positive effect on patients.

Photo: megaflopp, Getty Images