Wearable hemodynamic monitoring device and system

A backend analysis system configured to receive data from a wearable patch for predicting heart failure events (HFE) comprising:

• a waveform predictor module, wherein the waveform predictor module is configured to take physiological waveforms from the wearable patch worn by a patient over past several days (cycles) as input and uses it to predict the waveforms the patient is likely to develop over the next few days (cycles);
• a computational module, wherein the computational module is configured to calculate relevant parameters from characteristic points of all of the predicted waveforms;
• a set of comparator engines, wherein one of the comparator engines is configured for a stochastic comparison of the parameters obtained at different time cycles from the predicted waveforms;
• and the other engine is configured for a stochastic comparison of the parameters obtained at different time cycles from predicted ECG/EKG and thoracic bioimpedance waveforms;
• a set of recurrent neural network (RNN) based learning modules, wherein the learning modules are configured to estimate a baseline model where baseline parameters are generated, based on a stochastic interdependent relationship between the predicted waveforms and the one between ECG/EKG and thoracic bioimpedance waveforms;
• a baseline module, which is configured to compare the parameters obtained from the predicted waveforms against their respective baseline characteristics to determine the changes in cardiovascular or circulatory conditions;
• an anomaly detector module, wherein the anomaly detector module is configured to detect abnormalities in the parameters obtained from the predicted waveforms;
• and
• an alert module for prompting an alert to the patient if the abnormalities detected by the anomaly detector module exceed a threshold value.