The QRS detection using the Modified Pan-Tompkins algorithm

An accurate detection of QRS complexes during long-term ECG monitoring is a challenging task due to cardiac diseases such as cardia arrythmias, coronary artery blockage, etc., that change the signal pattern and displace the fiducial points in the ECG. We proposed a new QRS detection algorithm (MPT), being a modification of the standard Pan-Tompkins algorithm (SPT) consisted in the normalization of ECG signals in short windows (1200 ms) and the averaging of the aligned QRS patterns in relation to R maxima identified in these segments using SPT. The analysis was performed using 8-hour ECG signals from a wearable ECG device manufactured by Movesense recorded in healthy volunteers at the Universita Campus Bio-Medico di Rome. The performance of the MPT algorithm was high. The sensitivity, positive predictive value and accuracy were 99.98, 99.98%, and 99.96%, respectively. These values are comparable with values obtained using other two algorithms - the standard Pan-Tompkins (SPT), and an algorithm based on the wavelet transform (WT). Moreover, in contrast to other methods, the MPT allowed for determination of Q and S points in the identified QRS patterns.