JMIR Publications

ABSTRACT

Background:

Structured team-based communication, such as multidisciplinary rounds, handoffs, and goals-of-care discussions, is foundational to high-quality care in intensive care unit (ICU) settings. However, accurately capturing these complex verbal interactions in the medical record remains a challenge due to time pressures, documentation burden, and competing clinical demands. Ambient artificial intelligence (AI) scribes, which passively transcribe and summarize spoken interactions, offer a potential solution to streamline documentation while preserving the fidelity of team communication. Yet, little is known about how ICU clinicians perceive the integration of these tools into their high-stakes, collaborative workflows.

Objective:

This study explores clinician perceptions of integrating ambient artificial intelligence (AI) scribes into structured team-based discussions, including multidisciplinary rounds, handoffs and transitions of care, and goals of care discussions, with a broader goal of informing implementation of these scribes into real-world ICU clinical workflows.

Methods:

We conducted interviews and focus groups with ICU clinicians, including nurses, attendings, residents/fellows, respiratory therapists, and advanced practice providers, who routinely participate in structured communication discussions. Transcripts were analyzed using grounded theory to identify documentation needs, barriers, and design considerations for ambient AI scribe implementation in an ICU setting.

Results:

A total of 49 individuals, including 18 ICU attendings, 2 APPs, 10 ICU trainees, 9 ICU nurses, and 10 ICU RTs, participated. Clinicians emphasized the importance of accurate documentation but noted persistent barriers such as time constraints, documentation burden, and competing teaching and patient care responsibilities. Clinicians expressed enthusiasm about ambient AI scribes’ potential to reduce documentation burden and improve quality but requested personalization of outputs, robust consent protocols, and transparency around data use. Participants viewed ambient AI scribes as a promising tool to enhance both documentation fidelity and communication quality in ICU settings. Successful implementation may be contingent upon specialty-specific customization, transparent data practices, and sustained efforts to build provider trust.

Conclusions:

ICU clinicians were optimistic about the potential of ambient AI scribes to ease documentation burden and improve the capture of critical clinical discussions. Success will depend on clinician training, customization of output, and transparent institutional policies on consent and data use.