An Intelligent Cardiac Health Monitoring and Review System

Cardiac monitoring is an important application of mobilemonitoring systems. Such systems typically require experts to label electrocardiogram (ECG) data; however, large interand intra-expert variation limits the reliability and accuracy of diagnosis. This paper presents a process of integrating the mobile end of the system with a back-end annotation system for reviewing and scoring the quality of an ECG signal. This process serves as a platform for remote cardiac health monitoring and diagnosis. Furthermore, a “crowdsourcing” methodology was used to provide an adjudication of ECG annotations from a set of “experts”, comprising human trainees and automated algorithms. We show that this can provide diagnoses with equivalent accuracy to that of experts, at a substantially reduced cost.