The predictive alert system for patients with cardiovascular conditions encompasses the development of an analytical and predictive medicine module designed to support clinical diagnosis, as well as improvements to how data is presented in existing healthcare tools.

Among its goals are automating the data upload process so it is seamless for the physician and does not require time spent entering data, and running automatic analysis and reclassification of alerts and events emitted by Holter monitors, ICDs, and implantable pacemakers in under 1 minute, without the patient needing to visit the medical center, drastically reducing the risk of transmitting diseases such as COVID-19.

It also allows executions to be customized, so that healthcare staff, without specific technical knowledge, can segment and tailor the automatic processing to a specific center, staff member, and/or patient. Additionally, within its functional scope, the system will focus on reclassifying alerts emitted by devices from the leading manufacturers. However, it can be used in other specialties after proper training based on a dataset from that specialty.

Among the technical highlights is the creation of a diagnostic support engine based on machine learning techniques: this engine will work as a black box that, given a set of data or knowledge (corpus), can extrapolate a series of conclusions that help predict or support real-time diagnosis. Finally, a graphical flow control interface will be created: this tool allows the data ingestion, learning, and labeling algorithms to be managed graphically. Thanks to this tool, an experienced non-technical user is able to make small modifications to the system.