핵심 개념
Real-time ICU acuity prediction model APRICOT-Mamba developed and validated for improved patient outcomes.
초록
APRICOT-Mamba predicts patient acuity, transitions, and life-sustaining therapies in real-time in ICU patients.
Utilizes state space-based neural network with data from prior four hours and patient information at admission.
External validation on 75,668 patients, temporal validation on 12,927 patients, and prospective validation on 215 patients.
Comparable performance to state-of-the-art models for mortality, acuity, transitions, and therapy predictions.
Provides real-time monitoring and timely interventions for critically ill patients.
통계
APRICOT-M의 외부 AUROC 0.94-0.95, 시간적 AUROC 0.97-0.98, 전망적 AUROC 0.96-1.00
APRICOT-M은 안정 상태, 불안정 상태, 사망, 퇴원을 예측하기 위해 사용됨
APRICOT-M은 안정-불안정 전환, 생명 유지 요법에 대한 예측도 제공
인용구
"Early detection of deteriorating conditions can result in providing more timely interventions and improved survival rates."
"APRICOT-M allows for real-time acuity monitoring in critically ill patients and could help clinicians make timely interventions."