Ambient AI reduces documentation time and enhances quality in a simulated inpatient setting

Area of Science:

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

Background:

• Clinical documentation is vital but burdensome for healthcare professionals.
• Ambient AI systems show promise in outpatient settings but are understudied in inpatient environments.
• This study investigates ambient AI's impact on inpatient orthopaedic documentation.

Purpose of the Study:

• To evaluate the effect of an ambient AI scribe on documentation time.
• To assess the quality of clinical notes generated by ambient AI.
• To measure the impact of ambient AI on junior doctors' workload in a simulated inpatient setting.

Main Methods:

• Seven junior doctors used an ambient AI scribe during simulated orthopaedic ward rounds.
• 150 clinical documents (progress notes and discharge summaries) were generated manually and with AI.
• Documentation time, quality (PDQI-9), and workload (NASA-TLX) were measured.

Main Results:

• Ambient AI drastically cut documentation time for progress notes (27s vs. 128s) and discharge summaries (114s vs. 459s).
• AI-generated notes showed improved quality scores (PDQI-9) in thoroughness, currency, usefulness, comprehensibility, and organization.
• Junior doctors reported significantly reduced workload, particularly in frustration (-79%) and effort (-81%).

Conclusions:

• Ambient AI substantially decreases documentation time in simulated inpatient orthopaedic settings.
• AI scribes enhance the quality of clinical documentation, including progress notes and discharge summaries.
• The use of ambient AI alleviates clinician workload, positively impacting junior doctors' experience.