Computational efficiency and accuracy for QRS detection algorithms on clinical long term multilead monitoring

A number of relevant clinical measurements are derived from QRS detection. As a consequence, the fast and accurate calculation becomes a key factor to meet this target, which is especially relevant for the huge amount of beats recorded in the increasingly used long term monitoring. In this paper, we propose several algorithms that present high efficiency and accuracy for the QRS detetion in long term, and we benchmark them with some of the most relevant published QRS detection algorithms. All the implemented algorithms were applied to a specifically created gold-standard database. This gold-standard was labelled by expert clinicians, who evaluated manually every beat within 120 records of 48h multilead Holter from Hospital Virgen de la Arrixaca of Murcia (Spain). One of our new methods outperformed the others in terms of accuracy and computational efficiency, by using a multilead processing combining OR-function with a new Pan-Tompkins detector. It presented 99.2% sensitivity, 95.6% specificity, 97.1% accuracy, and 77-s processing time for our database. The QRS detection methods used in short term ECG records or traditional 24h Holter can be limited in long term, whereas the proposed multilead processing can provide with better performance in these monitoring scenarios.