Use of ECG quality metrics in clinical trials

We present a software package aimed to provide a viable mean to manage and automatically classify large amounts of ECGs on the basis of analytical metrics directly computed from the digital waveforms (such as noise content, heart rate regularity, repolarization complexity) and parametric metrics derived from ECG measurements (such as R and T wave amplitudes, QT/QTc intervals). Single or combined sets of metrics are used to generate ECG quality classes using, for each of the considered metric, predefined thresholds. The number of ECG subsets, as well as the thresholds to be applied, can be determined from the respective built-in distributions derived from an internal database of approximately 300.000 digital ECGs selected from previously conducted clinical trials. The use of this technique can significantly optimize the quality assessment and the management of digital ECG and secondarily improve the quality of recorded ECG in both clinical trials and common clinical practice.