Output details

Probabilistic classification of acute myocardial infarction from multiple cardiac markers

This research was in collaboration/funded with Randox Laboratories to investigate the inclusion of two new biochemical markers in a panel for automatic diagnosis of heart disease. Industry driven, it addresses a real problem using clinical data. The simplicity of the new classifier implies robustness and the discriminatory power exhibited by the chosen markers also requires little intervention from a clinician. The improvement found from sequential measures bolsters earlier findings in the literature. The probabilistic framework proposed is important for informing human decision making and is more likely to be taken up by the profession compared to hard yes/no type decisions.