A stand-alone ECG abnormality detector

Monitoring the ECG (Electrocardiogram) of patients with cardiac disorder is of paramount importance since they may undergo cardiac arrhythmias even without noticing it. Majority of existing devices for this purpose are only capable of recording ECG of patients which are analyzed later on by cardiologists. This paper presents a new system developed to continuously monitor the ECG of patients and analyze them in real time to identify any abnormalities. The new system consists of two parts: a portable device worn by the patient and a server side component. The arrhythmia detection algorithm used in the portable device shows 99.78% sensitivity, 99.65% positive predictivity for QRS complex detection and 93.48% sensitivity, 96.45% positive predictivity for ventricular beat classification when evaluated on MIT-BIH arrhythmia database. When an arrhythmia is detected the device automatically uploads the ECG data in the neighborhood of the abnormality to the remote server. This system provides a detailed analysis of the event at the web interface and also provides the cardiologist the opportunity to scrutinize the ECG data in order to make necessary diagnosis even without meeting the patient.