Photo Credit: megaflopp

An ambient scribe tool reduced documentation time and after-hours work while improving clinician focus and patient engagement during outpatient visits.

An ambient scribe tool that “listened” to an outpatient visit and summarized the encounter using generative AI was associated with greater clinician efficiency, a lower mental burden of documentation, and a greater sense of engagement with patients. Researchers reported their findings in JAMA Network Open.

“Clinicians expressed differing opinions about the accuracy and completeness of the notes; however, on average, these clinicians viewed ambient scribing positively after using it for 7 weeks,” wrote corresponding author Kevin B. Johnson, MD, MS, and colleagues.

The study included 46 physicians, nurse practitioners, and physician assistants from a mix of medical specialties at 22 outpatient clinical sites at the University of Pennsylvania. Participants underwent a 2-week training period followed by 5 weeks of using the ambient scribe in practice.

“Clinicians with ambient scribe access were able to activate a recording button within the EHR [electronic health record] mobile application patient-visit interface to begin recording,” researchers explained. “Clinicians confirmed consent to audio recording with the patient and any other parties present in the encounter before starting passive, hands-free recording of the visit.”

After the recording was stopped, AI generally took less than 1 minute to produce documentation for the encounter.

According to the study, use of the scribe tool was associated with 20.4% less time working on notes per appointment (from 10.3 to 8.2 minutes), 9.3% increased same-day appointment closure (from 66.2% to 72.4%), and 30.0% less after-hours work time (from 50.6 to 35.4 minutes per workday). Note length was 20.6% greater with the scribe tool.

Targeted clinician perspective questions on a 7-point scale identified a lower sense of distraction during patient conversations and a lower likelihood of feeling mentally drained from documentation burden with use of the scribe.

Open-ended qualitative feedback revealed a range of clinician views on the tool.

“Some clinicians found them [the AI-generated notes] accurate and detailed, whereas others found them overly error-prone,” researchers wrote. “A recurring theme was the need for substantial editing and proofreading of the AI-generated notes, which sometimes offset the time saved.”

Usability scores averaged 76.6 out of 100, indicating the tool was easy to use. As to how likely clinicians were to recommend the tool to others, 35.1% were considered promoters, 35.1% detractors, and 29.7% passives.