A new major version 2.0 of the Cardiolund ECG Parser is released today. The focus of this release is on improving the specificity of the review groups used in atrial fibrillation (AF) screening of large databases with home based thumb-ECG recordings.
The major challenges in screening for AF in thumb-ECG recordings are 1) discerning between disturbances (noise) looking like irregular rhythm and actual rhythm irregularity, and 2) classifying arrhythmias, to leave only the unexplainable rhythm irregularities (namely AF) for manual clinical review.
With this new ECG Parser version we have managed to significantly reduce the size of the review groups by 15% to 20%, depending on the application, based on data from the STROKESTOP study. This implies a major reduction in time spent on manually interpreting ECGs in AF screening applications.
The release also includes a number of other improvements, such as new markers indicating baseline disturbances and position of AF on the beat level, significantly improved SVES and VES event and run detections, and improved stim detection and cancellation.
For 12-lead systems, the electrical axis of the heart is now calculated, and the AFR Tracker is available as an add-on.
Further, there is now has a new default output format that is well suited for longer recordings, such as Holter.
This release is a milestone in improved performance, which is now also documented according to the ANSI/AAMI EC57 requirements.