Ambient Artificial Intelligence Scribe Implementation in an Ambulatory Setting: Experience and Outcomes form a Single Medical Group (Preprint)

BACKGROUND Healthcare providers spend an excessive amount of time within electronic medical record (EMR) systems documenting patient encounters, often amounting to hours of time outside of regular office hours. This affects provider productivity and directly contributes to burnout. Artificial intelligence (AI) is becoming more integrated in medical care, including development of speech recognition and note writing algorithms. Limited studies exist on how these AI tools are impacting provider satisfaction, work-life balance, and patient satisfaction. OBJECTIVE The aims of this study were to assess the use of Ambient AI in medical documentation and the effects it has on note quality, time spent in EMR, provider burnout, and patient satisfaction. METHODS This is a prospective study with the Hawaii Pacific Health Medical Group (HPHMG) to pilot an AI note writer. Abridge was chosen as the AI platform and integrates with the Epic EMR. A goal of 75 providers for a 3-month pilot period was established from December 2024 through February 2025. Surveys were distributed to providers before and during the trial period. Epic Signal and Abridge data were used to correlate provider perceived outcomes with medical record recorded outcomes. Users were then divided into groups based on frequency of AI use (≥60% is high utilization). The primary outcome was time in documentation per appointment RESULTS A total of 80 providers were recruited with 79 completing the pilot. Over 25,000 notes were generated across 23 specialties. Providers who reported spending 8 hours or more per week on notes outside of clinic hours decreased by 75%. Signal metrics found a 21% decrease in time in notes per day and a 13% decrease in pajama time among high utilizers. There was an 8% decrease in the number of providers reporting burnout symptoms and a 50% decrease in provider perceived clinic note completion difficulty without a reported decrease in note quality. Providers reported 83.7% of notes required less than a quarter of the notes to be edited. Patient experience via “provider listened to me” scores, while improved (92.3% to 92.6%) was not significant. CONCLUSIONS This ambient AI note writer decreased time providers spent writing notes in clinic, decreased time in the EMR outside of work hours, and decreased symptoms of burnout without sacrificing note quality. Further study on cost effectiveness capabilities in relation to increasing patient census are ongoing as are long term studies regarding provider burnout.