Evaluating the risk of hypertension using an artificial neural network method in rural residents over the age of 35 years in a Chinese area

Hypertens Res. 2010 Jul;33(7):722-6. doi: 10.1038/hr.2010.73. Epub 2010 May 27.

Abstract

Hypertension (HTN) has been proven to be associated with an increased risk of cardiovascular diseases. The purpose of the study was to examine risk factors for HTN and to develop a prediction model to estimate HTN risk for rural residents over the age of 35 years. This study was based on a cross-sectional survey of 3054 rural community residents (N=3054). Participants were divided into two groups: a training set (N1=2438) and a validation set (N2=616). The differences between the training set and validation set were not statistically significant. The predictors of HTN risk were identified from the training set using logistic regression analysis. Some risk factors were significantly associated with HTN, such as a high educational level (EL) (odds ratio (OR)=0.744), a predominantly sedentary job (OR=1.090), a positive family history of HTN (OR=1.614), being overweight (OR=1.525), dysarteriotony (OR=1.101), alcohol intake (OR=0.760), a salty diet (OR=1.146), more vegetable and fruit intake (OR=0.882), meat consumption (OR=0.787) and regular physical exercise (OR=0.866). We established the predictive models using logistic regression model (LRM) and artificial neural network (ANN). The accuracy of the models was compared by receiver operating characteristic (ROC) when the models were applied to the validation set. The ANN model (area under the curve (AUC)=0.900+/-0.014) proved better than the LRM (AUC=0.732+/-0.026) in terms of evaluating the HTN risk because it had a larger area under the ROC curve.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • China / epidemiology
  • Cross-Sectional Studies
  • Educational Status
  • Exercise
  • Female
  • Humans
  • Hypertension / epidemiology*
  • Male
  • Middle Aged
  • Neural Networks, Computer*
  • Overweight / epidemiology
  • Prognosis
  • Risk Factors
  • Rural Health / statistics & numerical data*
  • Sedentary Behavior