Construction and verification of risk warning model for secondary vascular cognitive impairment in stroke patients
Objective To establish a risk early warning model for secondary vascular cognitive impairment(VCI)in stroke patients,and to verify the predictive effect of this model.Methods A total of 374 stroke patients from General Medicine Department of Gulou Hospital Affiliated to Nanjing University School of Medicine were selected as the study objects,and 262 of them were model constructed by Logistic regression analysis.The predictive ability and fit of the model were evaluated by area under ROC curve and Hosmer-Lemeshow test.The model was verified by 112 stroke patients.Results The incidence of secondary VCI in stroke patients was 48.7％.The influencing factors of secondary VCI in stroke patients included age,hypertension,diabetes,cerebrovascular stenosis and stroke frequency.P=0.683 in Hosmer-Lemeshow test,the area under ROC curve was 0.860,and the maximum of Yoden index was 0.647.The optimal critical value was 0.460,the sensitivity was 0.800,and the specificity was 0.760.Conclusion The risk warning model of secondary VCI in stroke patients established in this study has good predictive effect,and can be widely used in the screening of secondary VCI in stroke patients.
StrokeVascular cognitive function impairmentEarly warning model
NETL
NSTL
万方数据
