Researchers in Florida are investigating whether artificial intelligence can help shorten the length of ICU stays and reduce hospital-based delirium.
Scientists from the University of Florida are developing a system using remote monitoring and algorithms to predict the outcomes of patients with severe illness or injury, WAFB reported May 30. Patients' pain levels would be observed through such visual cues as facial expressions and muscle twitches, while sensors would record body movements and posture.
The AI "will help the physicians and nurses monitor the patients, and also predict the trajectory of the patients in the ICU," Parisa Rashidi, PhD, a biomedical engineer at the University of Florida, told the TV station.
Algorithms would then analyze the data to "help you predict who is going to get sicker in the next three, four, five hours," said Azra Bihorac, MD, a researcher with the University of Florida College of Medicine.
Researchers are also using AI to identify the potential causes of ICU delirium, such as light levels, noise and odor. The studies started last year and go through 2026. The techniques could be tested in a clinical setting in about three years.