Analysis of Ambient AI Transcription in Outpatient Palliative Care

Key Message : This study highlights the emerging adoption of ambient AI scribes in palliative care outpatient clinics. Analysis of utilization trends and survey data offers a first-in-field perspective on how ambient AI scribes can enhance clinician presence and reduce documentation load, while underscoring challenges in fully capturing complex goals-of-care discussions.

Abstract:

Background:
Clinical documentation remains a leading contributor to physician burnout in palliative medicine, where encounters rely on nuanced communication, empathy, and shared decision making. Emerging data from Kaiser Permanente Northern California suggests that ambient artificial intelligence (AI) scribes can reduce documentation burden and improve clinical experience (1). A Stanford qualitative study similarly found reduced cognitive load, temporal demand, and after-hours work among physicians using ambient AI, despite accuracy and note-style challenges (2). In 2024, ambient AI scribes were implemented across 16 hospitals in Southern California to support real-time transcription and note generation.

Methods:
This descriptive, retrospective cohort study included full-time physicians in palliative medicine and other specialties using ambient AI scribes in clinic visits from April 2024 to April 2025. Structured surveys assessed satisfaction, accuracy, workflow efficiency, and trust in the AI tool using a 10-point Likert scale and open-ended questions. Two researchers independently conducted a thematic analysis of qualitative responses.

Results:
Twenty-eight palliative care physicians completed the survey; 80% used ambient AI for more than six months. On average, AI scribes were used in 56% of encounters. Most respondents (79%) agreed or strongly agreed that AI improved documentation quality and reduced administrative burden. Respondents reported saving 3-5 minutes per encounter, improved patient engagement, and reduced after-hours charting. Qualitative themes emphasized the need for better capture of goals-of-care discussions, emotional and social nuances, and enhanced EHR integration to optimize use in palliative settings.

Conclusion:
Integration of ambient AI transcription into outpatient palliative workflows was associated with high satisfaction and meaningful improvements in efficiency and physician well-being. These findings align with prior studies and underscore the need for continued refinement of AI models to better capture complex, valued-based conversations in palliative medicine. This study offers early insights into clinicians' experience and directions for future AI optimization.

References: 1. Tierney AA, Gayre G, Hoberman B, et al. Ambient Artificial Intelligence Scribes to Alleviate the Burden of Clinical Documentation. NEJM Catalyst. 2024;5(3). doi:10.1056/cat.23.0404

2. Shah SJ, Crowell T, Jeong Y, et al. Physician Perspectives on Ambient AI Scribes. JAMA Netw Open. 2025;8(3):e251904. doi:10.1001/jamanetworkopen.2025.1904