AbstractsEngineering

In the last decade, heart disease has been the leading cause of death all over the world. However, it is among the most preventable and controllable diseases. The World Health Organization reported that early detection of heart disease reduces progression to severe and costly illness and complications. Early detection of heart disease patients helps in recovering the patients’ health and decreasing the mortality rate from heart disease. Although heart disease can be detected by several tests, such as electrocardiogram, stress tests, and cardiac angiogram, these tests are expensive and cannot be used as community-level screening tests. The Framingham Heart Disease Risk Evaluation Tool and the Australian Absolute Cardiovascular Risk Calculator are two common heart disease risk evaluation screening tests. However, both tests need prior blood sample investigations, an invasive and relatively costly process, which reduces their usability in other than medical settings. Motivated by the increasing mortality rates of heart disease patients, researchers have been applying different data mining techniques in the diagnosis of heart disease. Research finds that the same data mining technique shows different results across different heart disease datasets indicating that there can be significant attributes for heart disease diagnosis. Furthermore, researchers suggest that hybrid data mining techniques show better performance in the diagnosis of heart disease patients. This research seeks to help healthcare professionals in the early detection and risk evaluation of heart disease patients using data mining analysis. To achieve this objective, the significant attributes in the diagnosis of heart disease patients are identified using a benchmark dataset and a new larger dataset, the reliability of non-invasive attributes in the diagnosis of heart disease is investigated, the enhancement of applying hybrid data mining model to the non-invasive attributes is tested, and a heart disease expert system risk evaluation tool (HD - ESRET) using hybrid data mining model on non-invasive data attributes is constructed. Although this research builds a low-cost heart disease expert system risk evaluation tool using a novel non-invasive data attributes combination, its usability testing among healthcare providers still needs further investigation.