Machine Learning Techniques for the Detection of Shockable Rhythms in Automated External Defibrillators

Abstract

Early recognition of ventricular fibrillation (VF) and electrical therapy are key for the survivalof out-of-hospital cardiac arrest (OHCA) patients treated with automated external defibrilla-tors (AED). AED algorithms for VF-detection are customarily assessed using Holter record-ings from public electrocardiogram (ECG) databases, which may be different from the ECGseen during OHCA events. This study evaluates VF-detection using data from both OHCApatients and public Holter recordings. ECG-segments of 4-s and 8-s duration were ana-lyzed. For each segment 30 features were computed and fed to state of the art machinelearning (ML) algorithms. ML-algorithms with built-in feature selection capabilities wereused to determine the optimal feature subsets for both databases. Patient-wise bootstraptechniques were used to evaluate algorithm performance in terms of sensitivity (Se), speci-ficity (Sp) and balanced error rate (BER). Performance was significantly better for publicdata with a mean Se of 96.6%, Sp of 98.8% and BER 2.2% compared to a mean Se of94.7%, Sp of 96.5% and BER 4.4% for OHCA data. OHCA data required two times morefeatures than the data from public databases for an accurate detection (6 vs 3). No signifi-cant differences in performance were found for different segment lengths, the BER differ-ences were below 0.5-points in all cases. Our results show that VF-detection is morechallenging for OHCA data than for data from public databases, and that accurate VF-detection is possible with segments as short as 4-s.