Crit Care. 2023;27(1):346.

Prospective, multicenter validation of the deep learning-based cardiac arrest risk management system for predicting in-hospital cardiac arrest or unplanned intensive care unit transfer in patients admitted to general wards.

, , , , , , , , ,

제1저자 조경재, 김정수 · 교신저자 이연주

PMID 37670324 PMC10481524 DOI 10.1186/s13054-023-04609-0 IF 11.4 호흡기내과

소속 (8)

VUNO, Seoul, Republic of Korea.
Division of Critical Care Medicine, Department of Hospital Medicine, Inha College of Medicine, Incheon, Republic of Korea.
Department of Intensive Care Medicine, Dong-A University Hospital, College of Medicine, Busan, Republic of Korea.
Department of Critical Care Medicine, Seoul National University Hospital, Seoul, Republic of Korea.
Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Republic of Korea.
Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, Republic of Korea.
Department of Emergency Medicine, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, Republic of Korea.
Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, Republic of Korea. yjlee1117@snubh.org.

Abstract

BACKGROUND Retrospective studies have demonstrated that the deep learning-based cardiac arrest risk management system (DeepCARS™) is superior to the conventional methods in predicting in-hospital cardiac arrest (IHCA). This prospective study aimed to investigate the predictive accuracy of the DeepCARS™ for IHCA or unplanned intensive care unit transfer (UIT) among general ward patients, compared with that of conventional methods in real-world practice.

METHODS This prospective, multicenter cohort study was conducted at four teaching hospitals in South Korea. All adult patients admitted to general wards during the 3-month study period were included. The primary outcome was predictive accuracy for the occurrence of IHCA or UIT within 24 h of the alarm being triggered. Area under the receiver operating characteristic curve (AUROC) values were used to compare the DeepCARS™ with the modified early warning score (MEWS), national early warning Score (NEWS), and single-parameter track-and-trigger systems.

RESULTS Among 55,083 patients, the incidence rates of IHCA and UIT were 0.90 and 6.44 per 1,000 admissions, respectively. In terms of the composite outcome, the AUROC for the DeepCARS™ was superior to those for the MEWS and NEWS (0.869 vs. 0.756/0.767). At the same sensitivity level of the cutoff values, the mean alarm counts per day per 1,000 beds were significantly reduced for the DeepCARS™, and the rate of appropriate alarms was higher when using the DeepCARS™ than when using conventional systems.

CONCLUSION The DeepCARS™ predicts IHCA and UIT more accurately and efficiently than conventional methods. Thus, the DeepCARS™ may be an effective screening tool for detecting clinical deterioration in real-world clinical practice. Trial registration This study was registered at ClinicalTrials.gov ( NCT04951973 ) on June 30, 2021.

Keywords

Artificial intelligenceCardiac arrestDeep learningEarly warning scoreIntensive care unitRapid response system

Figures

**Fig. 1** The performance of each model predicting in-hospital cardiac arrest or unplanned intensive care unit transfer. a The ROC curve for predicting deterioration (IHCA or UIT). b The ROC curve for predicting IHCA. c The ROC curve for predicting UIT. DCARS: deep learning-based cardiac arrest risk score; MEWS: Modified Early Warning Score; NEWS: National Early Warning Score; SPTTS: single-parameter track-and-trigger system; AUROC: area under the receiver operating characteristic curve; ROC: receiver operating characteristic curve; CI: confidence interval; IHCA: in-hospital cardiac arrest; UIT: unplanned intensive care unit transfer

**Fig. 2** Comparison of mean alarm counts/day/1,000 beds at the same sensitivity level for each prediction model. MACPD: mean alarm counts per day per 1,000 beds; DCARS, deep learning-based cardiac arrest risk score; MEWS: modified early warning score; NEWS: national early warning score; SPTTS: single-parameter track-and-trigger system

**Fig. 3** The comparison of appropriate alarm rates at the same sensitivity level for each prediction model. DCARS: deep learning-based cardiac arrest risk score; MEWS: Modified Early Warning Score; NEWS: National Early Warning Score; SPTTS: single-parameter track-and-trigger system; UIT: unplanned intensive care unit transfer; DNR: do not resuscitate; Screen.: Screening; Interv.: Intervention

**Fig. 4** The cumulative percentage of patients with IHCA and UIT for each prediction model. DCARS: deep learning-based cardiac arrest risk score; MEWS: modified early warning score; NEWS: national early warning score; UIT: unplanned intensive care unit transfer; IHCA: in-hospital cardiac arrest

**Fig. 5** The performance of each model predicting IHCA and UIT in different cohorts. AUROC: area under the receiver operating characteristic curve; DCARS: deep learning-based cardiac arrest risk score; MEWS: Modified Early Warning Score; NEWS: National Early Warning Score; UIT: unplanned intensive care unit transfer; IHCA: in-hospital cardiac arrest

그림 출처: PubMed Central 공개 논문 (PMC10481524)

Conflict of interest statement

All authors have disclosed that they have no potential conflicts interest with any companies or organizations.

그 밖의 정보