Abstract 8620: Real-Time Prediction of Cardiac Arrest in Critically Ill Patients Presenting to the Emergency Department Using a Neural Network Score Incorporating Heart Rate Variability Compared with the Modified Early Warning Score

Abstract

Aims A key aim of triage is to identify those with high risk of cardiac arrest as they require intensive monitoring, resuscitation facilities and early intervention. We aim to validate a novel neural network (NN) score incorporating Heart Rate Variability (HRV) for triage of critically ill patients presenting to the emergency department by comparing the area under the curve, sensitivity and specificity with the modified early warning score (MEWS). Methods We conducted a prospective observational study of critically ill patients in an emergency department of a tertiary hospital. At presentation, HRV parameters generated from a 5-minute electrocardiogram recording are incorporated with age and vital signs to generate the NN score for each patient. The patients are then followed up for outcomes of cardiac arrest or death. Results From June 2006 to June 2008, we enrolled 925 patients. The area under the receiver operating characteristic (ROC) curve for NN scores in predicting cardiac arrest within 72 hours is 0.781 (95% CI: 0.724-0.839) as compared to 0.680 (95% CI: 0.597-0.762) for MEWS. As for in-hospital death, the area under the curve for NN score is 0.741 (95% CI: 0.688-0.795) as compared to 0.693 (95% CI: 0.638-0.748) for MEWS. A cutoff NN score of ≥ 60 predicted cardiac arrest with a sensitivity of 84.1% (95% CI: 66.1-91.1), specificity of 72.3% (95% CI: 69.2-75.2) and NPV of 98.8% (95% CI: 97.5-99.4). A cutoff MEWS of ≥ 3 predicted cardiac arrest with a sensitivity of 74.4% (95% CI: 58.5-86.0), specificity of 54.2% (95% CI: 50.8-57.5) and NPV of 97.8% (95% CI: 95.9-98.8). Conclusion We found NN scores to be more accurate than MEWS in predicting cardiac arrest within 72 hours and death. This has potential to develop real-time bedside device for triage based on cardiac arrest prediction.