After success with physicians, Mount Sinai Medical Center expands ambient AI to nurses

Bringing ambient AI into the nursing workflow

Unlike a physician visit, where ambient technology captures an audible conversation between a clinician and patient, much of nursing care happens without being spoken aloud. To use ambient AI, nurses had to learn to narrate their work.

During an assessment, nurses bring the Chart with Art tool into the patient's room and explain to them how it will be used. They then activate Chart with Art and verbalize their findings into the Epic Rover -- a handheld device used to access the EHR -- while performing the assessment.

"That was probably one of the biggest learning curves for the staff, was to narrate what they were doing," Stewart said.

Mount Sinai started with head-to-toe assessments because they are relatively predictable, with plans to expand to pain assessments, reassessments and skin documentation.

Nurses were closely involved in designing the new workflow. Pilot participants developed scripts and practiced narrating care before using the technology with patients, Stewart shared. Nursing leaders and educators also practiced with the tool in Epic's training environment before it was introduced to staff.

Ambient documentation creates more room for patient interaction 

Mount Sinai's nursing pilot is still in its early stages, so the organization does not yet have the type of documentation metrics it has collected for physicians.

Still, nursing leaders are already seeing ways the technology could change the patient experience at the bedside.

Because nurses have to describe what they are assessing, patients hear more about what is happening during an examination. Stewart said that has prompted patients to ask questions about findings they might not otherwise have known the nurse was evaluating.

A nurse might tell a patient that their breath sounds are clear, for instance, prompting the patient to ask what that means. The nurse can then explain what they are listening for and why.

"This really has helped with the communication with nurses and their patients, for sure," Stewart said.

Mount Sinai plans to examine whether that experience translates into measurable changes in patient satisfaction. The health system will compare patient experience scores across units as Chart with Art expands, examining HCAHPS measures related to nursing, physician and discharge communication.

The tool also transcribes in English and Spanish, an important capability for Mount Sinai's patient population. More than 30% of the health system's patients speak Spanish as their primary language.

Addressing nursing concerns about AI 

Adapting nursing workflows was not the only challenge. Stewart said that, as with any new technology, nurses also wanted to know if the ambient AI would actually be effective, particularly in terms of reducing their workload.

"Nurses worried about, 'Oh gee, is it going to add to my day? Is it going to be more work for me?'" Stewart said.

As nurses began using Chart with Art, seeing that the technology could reduce rather than add to their documentation workload helped ease some of those concerns. This helped boost adoption.

"People are much more willing to embrace technology when they view it as a value add, as opposed to a real change to their workflow that may or may not decrease the amount of time," said Alon Weizer, M.D., chief medical officer and senior vice president at Mount Sinai Medical Center.

Ensuring AI-generated documentation accurately reflects the care nurses provide is another consideration.

According to Stewart, clinical oversight is built into the workflow to ensure this accuracy. Chart with Art generates documentation, but the nurse must review the information and make any necessary corrections before it becomes part of the patient's medical record. Mount Sinai has established a policy requiring that review, consistent with the health system's broader approach to keeping clinicians involved in oversight of AI-generated information.

Expanding ambient AI for nurses 

Mount Sinai began the pilot with about eight or nine day-shift nurses and plans to expand to roughly 20 nurses on the unit before continuing the rollout. Night-shift nurses are among the next group slated to begin using the technology, giving the organization an opportunity to see how it fits into a different nursing workflow.

Stewart also sees potential for ambient documentation in specialized nursing settings, such as intensive care and labor and delivery.

As other health systems consider extending ambient AI beyond physicians, Stewart said one of Mount Sinai's biggest lessons has been to involve nurses in the implementation rather than simply introducing the technology to them after decisions have been made.

"My best advice is to involve the staff early and definitely make sure that you have the support from whomever, whatever program you're using," Stewart said.

Elizabeth Stricker, BSN, RN, comes from a nursing and healthcare leadership background, and covers health technology and leadership trends for B2B audiences.