High-Risk People Identification and Key Risk Factor Comparison of Cardiovascular Diseases in Northern China Rural and Urban Areas

Abstract

Data mining techniques were utilized to improve the regional cardiovascular diseases (CVDs) prevention and in addition to find why rural residents are in worse statues of CVDs. Decision tree (DT) was chosen to build a high-risk CVDs people identification model, and the steps of identifying are reduced, with 99.4% precision in the classification. In comparison with the top 18 key features assessed by Random Forest, rural people were at higher risk due to the factors such as higher total cholesterol, blood glucose, blood pressure, body mass index(BMI) and waistline. Urban people had higher hypertensive historical prevalence but lower hypertensive comorbidity prevalence than rural people because of better control of high blood pressure. Smoking, drinking and medicine taking affected current blood pressure and lipid. More tobacco use of rural female and more alcohol intake of rural male can take part of responsibilities for worse CVDs prevalence and control in rural area. In the intervention and prevention of CVDs in a region, high-risk people should be more focused on. And rural and urban residents should be separately implemented with intervenable methods because of different disease history statues, control effect and current statue. Gender also should be considered when lifestyles are significantly different.

Keywords

regional cardiovascular diseases prevention / data mining / high-risk group identification / key risk factors identification / intervenable risk factor analysis