Clinical Safety & Quality
Validation of Computer Vision in the Operating Room for Ambient Documentation
Chloe K. Nobuhara, MD; Yann Sakref, PhD; Joseph D. Forrester, MD, MSc; Ashley Peterson, MD, MS; Serena Yeung-Levy, PhD; Michael A. Pfeffer, MD; Samuel H. Wald, MD, MBA; Mary T. Hawn, MD, MPH
Published by
Overview
The use of ambient monitoring and artificial intelligence (AI) is rapidly transitioning from theory into clinical practice. One promised use case is automating documentation with recording devices, such as AI scribes in ambulatory care settings.1 Ambient AI documentation in the operating room (OR), however, is more complex as many of the events worth documenting rely on visual cues, such as patient entry and exit. Recent technological advances both in audiovisual recording devices and in computer vision have allowed for progress in intraoperative automated documentation and timestamping of patient events.2,3 However, real-world implementation and validation studies remain scarce.4 Here, we present our single-center experience validating ambient timestamp documentation in the OR for a variety of procedures at scale and discuss the framework of governance enabling its implementation.
Elevate your entire health system.
Deploy the definitive architecture for procedural ambient AI.
