Vuno said Tuesday it has unveiled the results of clinical studies for VUNO Med-Deep Triage and Acuity Scale (DTAS), artificial intelligence-based emergency patient classification software developed jointly with Mediplex Sejong Hospital.
Hospital emergency rooms operate emergency patient classification tools named Triage and Acuity Scale (TAS), to manage the demand for emergency medical care efficiently, such as setting the order of medical care for patients who have visited.
Until now, hospitals have used the Korea Emergency Medicine Association’s Korean Triage and Acuity Scale (KTAS). However, the company’s research on 10 million patients showed that its AI-based DTAS showed a higher accuracy of emergency patient classification than KTAS.
Vuno’s DTAS is an AI platform based on the study of about 80 percent of the data of 11,659,559 patients collected by the national emergency medical information network for three years. It predicts three models – deaths, ICU and general ward admission.
The company conducted an Area Under the Receiver Operating Characteristic (AUROC) evaluation for the remaining 20 percent and found that its platform can predict death 93.5 percent of the death (compared to 78.5 percent for KTAS), 89.4 percent of ICU admission (79.7 percent for KTAS), and 80.4 percent of general admission (68.1 percent for KTAS), respectively.
DTAS also showed a higher accuracy with less information than KTAS.
“The emergency room of the hospital is a place where doctors have to make an accurate judgment in a short time about which patient is more urgent,” said Professor Lee Young-nam, a researcher at Vuno and co-author of the research. “This study is the first case where artificial intelligence was applied in the field of emergency, and the possibility of future performance improvement is very high.”
The research results were published in PLOS ONE, a U.S. online academic journal of SCI class.
<© Korea Biomedical Review, All rights reserved.>