Abstract
<title>Abstract</title> <p>Premature Ventricular Contractions (PVCs) are a type of arrhythmia which are known simply as abnormal extra heartbeats. These are abnormal ventricular beats that are caused mainly by changes in the electrocardiographic (ECG) morphology along with the timing of the surrounding heart rhythm. Even though the morphology of the waveform provides excellent information about the automatic detection of PVCs through software, its interpatient variation limits the performance for previously unintroduced patient data which the model was never trained with. The main purpose of this study was to investigate if beat to beat rhythm context can recover missed PVCs by EKG morphology and to identify the temporal properties of the recovered beats. Ambulatory ECG recordings from the MIT-BIH Arrhythmia database that were annotated by cardiologists were analyzed in patients that experienced frequent PVCs. A classification model using waveform technology was developed to differentiate between PVC and non PVC heartbeats. The heartbeat rhythm was then displayed using wave technology and RR intervals under morphology and time constraints. The final group consisted 24 patients and 59,021 heartbeats were found with 6,880 PVCs identified. The morphology model resulted in a PVC false rate of 17.54%, and reduced to 15.67% when the RR timing context was incorporated. Of those 1207 undetected PVCs, 301 were recovered successfully and all of them had a longer RR interval and followed a local rhythm.This RR interval definitely brings some valuable information to the classification and can also be used to help retrieve missing PVCs.</p>