Clinician Perspectives on Ambient AI Scribes in the Intensive Care Unit: Qualitative Interview Study

Area of Science:

• Medical Informatics
• Clinical Communication
• Artificial Intelligence in Healthcare

Background:

• Structured team-based communication is vital in intensive care units (ICUs).
• Accurate documentation of complex ICU discussions is challenging due to time pressures and clinical demands.
• Ambient artificial intelligence (AI) scribes offer a potential solution for documentation assistance.

Purpose of the Study:

• To explore ICU clinicians' perceptions of integrating ambient AI scribes into team-based discussions.
• To inform the real-world implementation of AI scribes in ICU clinical workflows.

Main Methods:

• Conducted interviews and focus groups with diverse ICU clinicians (n=52).
• Utilized grounded theory analysis of transcripts to identify documentation needs and implementation considerations.
• Included nurses, attendings, trainees, respiratory therapists, and advanced practice practitioners.

Main Results:

• Clinicians highlighted persistent documentation barriers like time constraints and competing responsibilities.
• Participants expressed enthusiasm for AI scribes' potential to reduce burden and improve quality.
• Key requests included output personalization, robust consent, and data use transparency.

Conclusions:

• ICU clinicians are optimistic about ambient AI scribes easing documentation and improving discussion capture.
• Concerns exist regarding transparency in data use.
• Successful implementation hinges on clinician training, customization, and clear institutional policies on data use and consent.