HF-Predict is the first clinically accurate wearable device and machine learning software for predicting Heart Failure (HF) of a patient. This wearable system acquires relevant vital signs and parameters that measure the heart muscle functionality and cardiovascular system continuously. The machine learning methods combine important features from signals with the patient’s health record to perform accurate prediction of heart failure, days before it happens.
HF-Predict is a project of SmartCardia, a Swiss medical device company with focus on clinical quality data and validation.
Kaleidos designed from scratch a expert system to do preliminary evaluations based on more than 40 relevant parameters extracted from ECG signals. We enabled the use of two wearables to obtain more accurate data. Generated models predicted heart problems with more than a 90% accuracy. We also developed a secure portal for doctors and patients and validated through interviews and usability tests. Doctors can introduce data, manage clinical sessions and exchange information with patients, among other functionalities. Patients can access their real time data and medical reports through the platform, follow part of this process. They also can request doctor consultations. We automated the deployment of Machine Learning Technologies in secure and elastic cloud environments.