Evaluating heart failure predictors using quantitative approaches

Abstract

Heart failure (HF) is a type of disease which means that the human heart did not perform its function like it supposed to be. There are several factors that are known to affect the HF but which factors that really affect are unknown. This study analysed the factors that really contribute to the failure of a human heart. The variables that had been analysed in this study were age, the decrease of red blood cells (RBCs), the patient has hypertension or not, level of the creatinine phosphokinase (CPK) enzyme in the blood, the patient has diabetes or not, percentage of blood leaving the heart at each contraction, platelets in the blood, gender, the level of serum creatinine and serum sodium in the blood, the patient smokes or not and the follow - up period. The method used in this research were Decision Tree by using Chi – squared automatic interaction detection (CHAID) growing method to determine the significant contributing factors towards HF and Binary Logistic Regression to develop a model using the significant contributing factors towards HF. From the analysis of decision tree, we can see the results for the independent variables included were time to follow – up period, ejection fraction and diabetes. As a conclusion, the results from this study can be an indicator to whom needed, to know whether a person with HF can survive or not. It is recommend that future researcher can use the method mentioned to analyse new HF predictors.