Representation Learning Approaches to Detect False Arrhythmia Alarms from ECG Dynamics
Proceedings of the 3rd Machine Learning for Healthcare Conference, PMLR 85:571-586, 2018.
The high rate of intensive care unit false arrhythmia alarms can lead to disruption of care and slow response time due to desensitization of clinical staff. We study the use of machine learning models to detect false ventricular tachycardia (v-tach) alarms using ECG waveform recordings. We propose using a Superv sed Denoising Autoencoder (SDAE) to detect false alarms using a low-dimensional representation of ECG dynamics learned by minimizing a combined reconstruct on and classification loss. We evaluate our algorithms on the PhysioNet Challenge 2015 dataset, containing over 500 records (over 300 training and 200 testing) with v-tach alarms. Our results indicate that using the SDAE on Fast Fourier Transformed (FFT) ECG at a beat-by-beat level outperforms several competitive baselines on the task of v-tach false alarm classification. We show that it is important to exploit the underlying known physiological structure using beat-by-beat frequency distribution from multiple cardiac cycles of the ECG waveforms to obtain competitive results and improvement over previous entries from the 2015 PhysioNet Challenge.