Description

Cardiopulmonary failures are major public health concerns, due to the aging population. Each of these situations is burdened with a poor prognosis in the medium term and a source of prolonged hospitalizations, generating significant health costs. Early detection and prediction of organ failure could reduce health costs and risks for the patient, offering a reaction early and appropriate medical technology. The proposed approach aims to optimize the knowledge of a complex physiological domain and multi-system, while promoting the automatic transfer of knowledge. The approach proposed data-mining and development of algorithms for detecting and / or predicting a strong potential for disruption because it proposes to apply innovative automated analysis procedures to a fragile patient population, and then a transfer to the medical device industry.

From communicating tools of recording of the signals, the investigator envisage in a global way:

1. the constitution of a warehouse of physiological data of grown-up patients in acute situation (intensive care unit);
2. the development by data mining of a system of detection of organs failures or adverse events basing itself on the application of innovative algorithms, allowing the decision-making operational, from the fusion of arisen ill-assorted events;
3. the use of intelligent tools of auto-learning and elaboration of complex multimodal models for purposes of prediction of events;