A Study on Predicting Cause of Side Effects in ACS Patients Using Machine Learning

Abstract

Cardiovascular disorders are the 1st ranked cause of death throughout the world and, in particular, ACS (Acute Coronary Syndrome) is the key risk factor of cardiovascular disorders. ACS is a fatal disorder with high risk of manifestation of MI (Myocardial Infarction) and death rate that imparts significant burden on individuals and society due to high cost incurred in treatment. Clopidogrel or Ticagrelor is used as anti-platelet drug. Recently, Ticagrelor known to be high safe and effective has been administered predominantly. However, it has become difficult to administer this drug because of symptoms of dyspnea induced in some patients. Accordingly, statistically significant attributes were sought with Pearson correlation technique to assess the cause of dyspnea in patients and the causes of dyspnea in patients were predicted by applying Two-class Decision Forest algorithm. As a result, it was found that 7 attributes including Uric acid, CPK, HDL cholesterol, TSH, Age, AST and LA Volume iex are related to dyspnea, and in particular, Uric Acid imparted greatest effect on dyspnea. It would be possible to minimize the side effects by screening administration of Ticagrelor to ACS patients by utilizing such results of research.