The Past, Present, and Future of Automated Electrocardiogram Diagnostics

Abstract

　Electrocardiography was invented more than 100 years ago by Einthoven et al. It is still in use today as an important medical examination. Attempts to use computers to automatically diagnose electrocardiograms began in the 1960s. In the 1970s, an automated diagnostic system consisting of an electrocardiograph, a minicomputer, and a teletypewriter was developed, and in the 1980s, automated diagnostic and printing functions were built into the body of a single electrocardiograph. Although the accuracy of automatic diagnosis has been high for normal waveforms since its early development, there are still many cases of difficulty in diagnosing complex or small waveforms. In order to study the problems related to automatic ECG diagnosis, a "The task force on automated diagnosis of the electrocardiogram in Japan." consisting of researchers, company engineers, and clinicians was formed, and discussions were held, and an expert consensus statement was issued on the unification of the names of findings and diagnoses in automatic diagnosis. Furthermore, the study of inappropriate cases showed that the accuracy of automated diagnosis is not only over- or under-diagnosis of myocardial ischemia, but also misdiagnosis of atrial fibrillation and missed pacing spikes, although these cases are less frequent. In considering the future of automated diagnosis, it is important to note that the widespread use of smartwatches, including the Apple Watch, has created a situation in which individuals record and own ECG waveforms using ECG functions. To improve the accuracy of the current automatic diagnosis, it is suggested that in addition to improving the conventional algorithm, a program using AI diagnostic theory could also be useful.