Output details

A support vector machine for predicting defibrillation outcomes from waveform metrics

This work is funded through charitable trusts (McGrath-Trust/Ulster Garden Villages) which helped establish our Centre for Advanced Cardiovascular Research and is in collaboration with a number of hospitals, INSA-Lyon and Heartsine Technologies Ltd, who sponsored the Medical Research Fellowship for this work. The paper reports patient trials on a robust algorithmic method for ventricular fibrillation management based on machine learning techniques to optimise defibrillation. This is an alternative to conventional threshold-based classification and enables an individualised patient approach to timing of the defibrillation phase during cardiopulmonary resuscitation (CPR). This advancement is linked to Heartsine’s AED devices with CPR Advisor(http://www.heartsine.com/en/products/).