In the past decade, technologic advancements have enabled improvements in health care, particularly for evaluating, diagnosing, and monitoring patients, according to a perspective article published in the New England Journal of Medicine.

These improvements in care have been fueled by technology’s ability to reduce clinical errors and allow access to patient data more quickly than humans are capable of. An artificial intelligence domain known as “computer vision,” which has been facilitated by deep learning, is a technologic advancement further supporting the clinician in bedside care. This technology allows clinicians to observe patterns in patients’ attitude, behavior, and symptoms.

One of the main technologic advancements fueling the evolution of artificial intelligence computer vision is the computer vision system in Google’s self-driving cars. This computer vision can immediately initiate changes in steering when the car temporarily initiates a dangerous lane change. It is theorized that the same technology can immediately detect and alert clinicians on deviations in either clinician or patient behavior. These changes, which can often go unnoticed in human surveillance, can then be rectified, resulting in a lower risk for potential adverse effects.

Currently, computer vision is being tested as a promising tool for assisting clinicians in analyzing screened medical images. For instance, computer vision may help to accurately identify cancerous skin lesions to a degree comparable to that of dermatologists. Additionally, computer vision may help guide surgical workflow and help surgeons modify their behavior if the workflow accidently deviates from standard protocol. Computer vision has also been tested for improving hand hygiene compliance in the hospital setting, which if successful, may result in reducing the risk for infection at the bedside.

Related Articles

Computer vision technology represents a low-fatigue, low-cost method for monitoring clinician and patient behavior and improving care at the bedside and beyond. Because the data supporting its use in healthcare are limited, the limitations as well as the benefits of computer vision have yet to be fully elucidated. Prior to wide-scale adoption, rigorous testing in the real world and in clinical trial environments is necessary to ensure accurate delivery and results.

“Although safe hospital care presents unique challenges, if productivity gains seen in other industries are an indication, computer vision may contribute significantly to clinical quality and efficiency while freeing clinicians to focus on nuanced decision-making, engaging with patients, and delivering empathic care,” the investigators concluded.

Reference

Yeung S, Downing NL, Fei-Fei L, Milstein A. Bedside computer vision — moving artificial intelligence from driver assistance to patient safety. N Engl J Med. 2018;378(14):1271-1273.