Feasibility of Tailoring Artificial Intelligence–Assisted Ambient Scribes for Intensive Care Unit Rounds: Algorithm Development and Validation - Summary - MDSpire
Advertisement
Feasibility of Tailoring Artificial Intelligence–Assisted Ambient Scribes for Intensive Care Unit Rounds: Algorithm Development and Validation
To assess the feasibility of AI-assisted ambient scribes for capturing interprofessional rounds and synthesizing a document in the ICU.
Approach:
Phase 1: Focused on customizing the prompt for the LLM component of AI-assisted scribes for ICU rounds.
Phase 2: Evaluated the quality of the generated notes using the customized prompt after selecting the best-performing model.
Key Findings:
AI-assisted scribes have the potential to reduce documentation burden in the ICU.
The study utilized software-generated and manually reviewed transcripts from simulated ICU rounds.
Iterative prompt engineering improved the accuracy and quality of generated ICU progress notes.
Interpretation:
The study demonstrates the feasibility of using AI scribes in the ICU setting, highlighting the need for tailored approaches to address unique challenges.
Limitations:
The study was conducted in a simulated environment, which may not fully replicate real-world ICU conditions.
Only a limited number of models were evaluated.
Conclusion:
The pilot study indicates that customized AI-enhanced ambient scribes could support documentation in ICU rounds.
