Cardiovascular Attributable Risk and Risk Factors Evaluations as a Matter of Statistics and Data Mining Confluences

Abstract

Cardiovascular diseases represent a severe threat for humanity, being the first cause of death and hospitalization in both genders. An impressive number of studies have been developed in order to identify a set of factors causing this kind of illness, but only few of them were able to pay significant resources in analyzing large population samples (tens of thousands) and for longer periods of time (decades). This paper’s objective is to continue the previous researches of the eProCord project and to validate with concrete data the theoretical model developed for the attributable risk (AR). It will consider the same risk factors for myocardial infarction identified by INTERHEART study and the same work hypothesis. We will also evaluate if a certain value of the AR is also confirmed by the invoked disease of the patient. Using statistical and data mining tools we will investigate the prediction potential of the chosen factors and the opportunity to extend them in order to capture any cardiovascular disease. The empirical tests rely for now on a sample of 236 patients.

Keywords

• Cardiovascular Disease
• Myocardial Infarction
• Attributable Risk
• Roc
• Data Mining
• Classification