Extra-corporeal membrane oxygenation (ECMO) has since been increasingly deployed, to become integral to the therapeutic arsenal of the intensive care unit for patients suffering from cardiac and/or respiratory insufficiency. The inventors endeavored to explore the potential of hemolysis index (HI) monitoring during ECMO in a retrospective study including 115 patients for which EMCO was deployed. In particular, the inventors the inventors used machine-learning algorithms to build a decisional tree including the HI level as a predictive factor for patients under ECMO. Thus the present invention relates to use of algorithms for predicting survival of a patient having extracorporeal membrane oxygenation.