Automatic ECG Signal Quality Assessment in Mountain Rescuers Through the Computation of Sample Entropy

In the present study we propose a novel method to automatically assess the quality of ECG signals collected through a wearable device in typical mountain rescuers activities. ECGs signals have been obtained during sessions of programmed field tests at the Bormio Ski Resort (Valtellina, Lombardy, Italy) in the month of March. Here, following the defined protocol, a group of 15 mountain rescuers has carried out daily rescuers’ activities, while wearing wearable textile system by Smartex Srl. The test protocol was designed to simulate the real physiological demands of mountain rescuers during their emergency deployments. Among the activities performed rescuers had to walk up and down hill in snow-covered trails, carrying stretchers onto which simulated victims were located etc… To infer the quality of ECG signals recorded we developed an algorithm for the automatic evaluation of collected signal deterioration. This method is based on the analysis of regularity of ECGs’ P-QRS-T complexes pattern. To estimate the maintenance of typical ECGs pattern shape, Sample Entropy (SampEn) was computed in moving fixed-length windows sliding along the signal, obtained after applying wavelet transform of the row ECG. The SampEn indices series was then thresholded to spot ECG points where P-QRS-T complexes were more or less easy to identify, respect to points where signal quality was completely deteriorated. Moreover, we evaluated signal quality maintenance while performing low and high intensity activities.