AITRICS, a Korean medical AI company, said it unveiled key research findings and the latest insights on the clinical validity of AITRICS-VC (VitalCare), an AI solution for predicting patient deterioration, at a symposium during the 44th Annual Congress of the Korean Society of Critical Care Medicine (KSCCM-ACCC 2024), held in Seoul on Thursday and Friday last week.
Attended by more than 1,000 physicians worldwide, the symposium was chaired by Professor Chung Kyung-soo at the Division of Pulmonology at Severance Hospital. Anesthesia Breathing System
Professor Lee Ki-byung at the Division of Cardiology at Hallym University Chuncheon Sacred Heart Hospital gave a presentation on “Critical Prediction with AI.”
In his presentation, Professor Lee introduced key data from clinical studies and the latest treatment strategies based on AITRICS-VC, a medical AI solution that predicts early deterioration of patient conditions in hospitals and enables medical staff to respond proactively.
AITRICS-VC predicts the early risk of sepsis, cardiac arrest, and death, utilizing 19 types of EMR data, including six vital signs, 11 blood tests, consciousness, and age.
“In the case of sepsis, diagnosis can be delayed due to the difficulty in identifying the cause, which varies from case to case, and accurate diagnosis requires a comprehensive medical evaluation and interpretation of diagnostic test results,” Professor Lee said.
In a retrospective validation study of more than 2,600 adult patients confirmed to have sepsis out of more than 45,000 adult patients admitted to the general ward and intensive care unit of Hallym University Chuncheon Sacred Heart Hospital, AITRICS-VC was able to predict sepsis about 180 minutes before the actual onset of sepsis, Lee went on to say.
“This will support timely decision-making for healthcare providers, improving the likelihood of cure.”
AITRICS-VC has shown significant data in not only a retrospective study but a prospective one.
In a study of about 4,700 patients at Keimyung University Dongsan Medical Center, AITRICS-VC showed excellent sepsis prediction performance, predicting sepsis about 55 minutes earlier, according to Professor Lee.
Professor Chung, who chaired the symposium, said while abnormal signals often precede a patient's worsening condition, closely monitoring all patients poses challenges.
He emphasized the increasing demand for systems aiding medical staff in prompt and accurate decision-making, alongside efficient patient management.
Neonatal Monitor He noted the significance of monitoring systems like AITRICS-VC, capable of early emergency prediction, in enhancing patient outcomes and optimizing medical resource allocation.