Investigation of risk factors for the diagnosis of heart disease

Abstract

Introduction: Limited studies have investigated the association between diagnosis of heart disease and its relating risk factors and haven’t met with robust results. We hypothesized that all or part of the risk factors are correlated with the heart disease. Method: This dataset published in 1998 was obtained from UCI Machine Learning Repository and was collected at University of Irvine. The dataset has 303 instances with 76 attributes, but all published experiments refer to using a subset of 14 of them. Our study focuses on 11 parameters specifically, including body health conditions, historic medical records, and habits. Logistic regression analyses were conducted to assess the relative risks of heart disease. Results: Both chest pain type (p < 0.01) and ST depression (p < 0.05) are positively correlated with the incidence of heart disease. Maximum heart rate, on the other hand, are negatively correlated with the diagnosis of heart disease. Conclusion: Our study suggested that chest pain type, ST depression and maximum heart rate are saliant contributors to indicate the occurrence of heart disease. The findings from our study have implications for the heart disease and call for future studies to explore the underlying prevention strategies of this findings.