Decision Tree Model Of Influencing Factors Of Carotid Atherosclerosis

Objective:The purpose of this study was to investigate the effects of various factors especially behavioral habits,behavioral approach /inhibition system sensitivity(BAS/BIS),common diseases on carotid atherosclerosis(CA),establish a decision tree,and analyze the characteristics of its multi-factor action on the occurrence of CA.Methods:From July 2019 to February 2021,573 participants(329 CA and 244 non-CA)were recruited from the Physical Examination Center of the Affiliated Hospital of Qingdao University.Demographic factors,such as age,gender and body mass index(BMI),etc.;behavioral factors,such as smoking,alcohol consumption,tea drinking,high-fat diet,physical exercise,etc.;concomitant disease factors,such as hypertension,diabetes,hyperlipidemia,etc.;the scores of the Reward and Punishment Sensitivity Scale were collected by questionnaire and the intra-media thickness of carotid artery were analyzed by ultrasound.The data were analyzed using SPSS 25.0 software,measurement data were expressed as mean ± standard deviation using independent sample t-test;enumeration data were analyzed by chi-square test.Multivariate Logistic regression was used to analyze the relationship between the above factors and carotid atherosclerosis,and risk factors and protective factors were found.P < 0.05 was considered statistically significant.IBM SPSS Modeler 18.0 decision tree C5.0 algorithm was used to establish the decision tree model of carotid atherosclerosis.The synergistic influence pattern of behavior habit,reward and punishment sensitivity and common diseases on carotid atherosclerosis was observed.Med Calc 15.2.2 software was used to draw the receiver operating characteristic curve(ROC)of Logistic regression model and decision tree model in the occurrence of carotid atherosclerosis.By comparing the AUC of the two models,the accuracy of the two models to predict the occurrence of carotid atherosclerosis was testified.Results:1.Multivariate Logistic regression analysis showed that aging,high-fat diet,hyperlipidemia and hypertension were the risk factors for carotid atherosclerosis,while physical exercise was the protective factor.2.The decision tree model showed that hypertension was a root node,age,hyperlipidemia,behavioral inhibition system,physical exercise,diabetes,alcohol consumption,tea drinking,behavioral activation system,high fat diet,gender and BMI were child nodes.Respectively,for the occurrence of carotid atherosclerosis the above factors have synergistic effect in different ways,for example,people aged < 70 years,have high behavioral inhibition sensitivity,physical exercise habits,tea drinking habits and high-fat diet,the probability of carotid atherosclerosis is87.50%;The probability of carotid atherosclerosis was 61.91% for those aged less than 70 years,with high scores of behavioral inhibition system,no physical exercise habit,no tea drinking habit,and no high-fat diet.However,smoking was not included in the decision tree model of this study.3.The area under the ROC curve comparison showed that the two models had statistically difference in predicting carotid atherosclerosis(P < 0.05).The area under the curve of the decision tree model was greater than that of the Logistic regression model,which confirmed that the prediction accuracy of the decision tree model was better.Conclusion:1.Decision tree model shows that high blood pressure,age,hyperlipidemia,diabetes,alcohol consumption,high fat diet,male,high body mass index can increase the risk of the occurrence of carotid atherosclerosis,and physical exercise,drinking tea,female,high sensitivity,low reward sensitivity can reduce the occurrence of carotid atherosclerosis risk,shows a variety of synergistic or superimposed modes among the factors.2.Each node factor in the decision tree model can be used as an indicator to identify high-risk groups of diseases,and the intervention of node factors is conducive to the prevention and control of carotid atherosclerosis in corresponding groups of people.3.In the aspect of predicting the occurrence of carotid atherosclerosis,the prediction accuracy of decision tree model is higher.Decision tree model can simply and intuitively predict the occurrence risk of carotid atherosclerosis and the correlation among influencing factors.