Generative AI in preanesthetic consultations: Effects on efficiency, documentation workload, quality, and...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Clinical Documentation Improvement

Background:

• Clinicians dedicate over 30% of their workday to electronic health records, impacting patient interaction and contributing to burnout.
• Preanesthetic consultations require extensive documentation, presenting an opportunity for AI-driven support.

Purpose of the Study:

• To evaluate a generative AI application for automating documentation during preanesthetic consultations.
• To assess the AI's impact on consultation efficiency, clinician workload, physician-patient interaction, documentation quality, and user experience.

Main Methods:

• A randomized simulation study involving 30 anesthesiologists at University Hospital Zurich.
• Each anesthesiologist conducted standardized consultations with and without AI assistance (Isaac).
• Outcomes measured included consultation duration, visual attention, workload, and documentation quality.

Main Results:

• AI-assisted documentation reduced consultation duration by 18% and significantly decreased screen fixation, keyboard input, and mouse clicks.
• Clinicians reported a trend toward lower workload and better patient engagement with AI.
• However, manual documentation was rated higher in quality by external raters, and clinicians had concerns about AI formatting.

Conclusions:

• Generative AI-supported documentation enhances efficiency and user experience in simulated preanesthetic consultations.
• Physician review is necessary to ensure AI-generated documentation quality in clinical practice.
• AI tools can reduce typing effort and screen interaction time, potentially streamlining workflows.