The development of automated ECG analysis purposes and its validation and implementation into clinical practice

Each year, an estimated 300 million electrocardiograms (ECGs) are obtained worldwide, demonstrating the ECG to be a fundamental tool in everyday clinical practice. Correct interpretation is crucial for accurate diagnosis of a variety of cardiac abnormalities and the life-threatening nature of cardiac disorders also requires timely assessment. Although acquisition of the ECG is well reproducible and standardized, human interpretation of the ECG varies greatly dependent on the level of experience and expertise. ECGs are often with discrepancy in interpretation between physicians and expert knowledge to correctly interpret ECGs is often lacking, especially in non-cardiology departments and pre-hospital care. Computer-generated interpretation of the ECG (CIE) has demonstrated to be able to support the physician in clinical diagnosis and decision making, with a significant rise in new automated detection approaches seen during the last decade. In this thesis, two automated ECG analysis purposes as being developed in the University Medical Center Utrecht (UMCU) will be discussed, with a focus on their validation and implementation into clinical practice.