Analysis Of Influencing Factors Of Mild Cognitive Impairment And Study On Decision Tree Model

Objective: To explore the factors affecting the occurrence and progression of mild cognitive impairment(MCI),establish a decision tree model,and analyze its practical significance to the occurrence and development of MCI.Methods: From October 2015 to February 2020,there were 330 patients in MCI group and 295 patients in Normal cognitive(NC)group who met the diagnosis criteria from the inpatient department and physical examination center of the affiliated Hospital of Qingdao University.Cognitive function was evaluated by Mini Mental State Examination scale(MMSE)and Clinical Dementia Rating scale(CDR),evaluation of activities of daily living by Activity of Daily Living scale(ADL),and demographic,lifestyle and previous chronic diseases data were collected.Chi-square test and Logistic regression were used.The decision tree model of the influencing factors of MCI was established,and the curve of the receiver operating characteristic(ROC)curve was drawn to compare the prediction ability of the Logistic regression model and the decision tree model.The patients in MCI group were followed up,with the date of filling in the data of MCI group as the starting point and the deterioration of cognitive function as the cut-off point.Kaplan-Meier survival analysis and COX regression were used to analyze the influencing factors of MCI progress,the decision tree model of influencing factors of MCI progress was established,and the ROC curve was drawn to evaluate the prediction ability of the decision tree model.Results: Logistic regression analysis showed that education,living alone,smoking,high-fat diet,hyperlipidemia,age,tea drinking,siesta,social interaction and hobbies were the independent influencing factors of MCI.Social interaction,high-fat diet,hobbies,living conditions,tea drinking and smoking enter all nodes of the decision tree model,in which social communication is the root node variable,which is the most important factor to predict the occurrence of MCI.The ROC curve shows that the AUC of the decision tree model for predicting the occurrence of MCI is 0.811(95% CI: 0.778-0.841).The AUC of the logistic regression model for predicting the occurrence of MCI is 0.809(95% CI: 0.775-0.839).The comparison between the two shows that there is no significant difference in the ability to predict the results.MCI follow-up results: 140 MCI patients progressed during the follow-up,and Kaplan-Meier survival analysis showed that the survival probability was 33.2% and the annual conversion rate of MCI progression was 20%.COX regression showed that diabetes was a risk factor for MCI progression,while physical exercise,social interaction and hobbies were protective factors for MCI progression.Hobbies,social interaction,physical exercise,diabetes and hypertension enter the nodes of the decision tree model,and hobbies are the root node variables,which are the most important factors to predict the progress of MCI.The decision tree predicts that the AUC for predicting MCI progress is 0.728(95% CI: 0.6747-0.775).Conclusion: 1.Low education,living alone,smoking,high-fat diet and hyperlipidemia are the risk factors for the occurrence of MCI.Low age,tea drinking,siesta,social communication and hobbies are the protective factors for the occurrence of MCI.2.Hypertension and diabetes are the risk factors for the progression of MCI,while hobbies,social interaction and physical exercise are the protective factors for the progression of MCI.3.In the process of the occurrence and progress of MCI,the influencing factors and the effects of factors are different in different stages,so the intervention of influencing factors should be more targeted in order to prevent the occurrence of MCI and delay the progress of MCI more effectively.4.Introducing the decision tree model into the study of influencing factors of MCI can intuitively and succinctly explore the influencing factors of the occurrence and development of MCI,the interaction between factors and the prediction of the occurrence and development of MCI,which reflects the advantage and feasibility of decision tree model in clinical medical research.