Status Of Cognitive Frailty In Elderly Patients With Chronic Kidney Disease And Construction Of Risk Prediction Model

Objectives:In this study,a cross-sectional survey was conducted to know the status of cognitive frailty in elderly patients with chronic kidney diseases(CKD),and to explore the risk factors affecting cognitive frailty.CART decision tree classification algorithm based on the R software was used to construct a model of risk prediction in elderly patients with CKD about cognitive frailty.It can facilitate clinical medical staff quickly intuitively judge in elderly patients with CKD risk of cognitive frailty,provide a new reference tool for the early prediction of cognitive frailty,and provide a decision basis for subsequent clinical nursing work.Methods:Convenience sampling method was used in this study to select 450 elderly patients with CKD from the Department of Nephrology of the First Affiliated Hospital of Shantou University Medical College and shantou Central Hospital,two grade III,Class A hospitals from October 2020 to July 2021 as the research objects.The status of cognitive frailty,general data,scale data and laboratory indicators were collected.Variables were screened by univariate and Logistic regression analysis,and the final risk variables of cognitive frailty were included in the model.The data set was divided into training set and verification set in a ratio of 7:3,and RStudio4.1.2software was used to construct a decision tree model of cognitive frailty in elderly patients with CKD based on CART classification tree algorithm.Confusion matrix,receiver operating characteristic curves(ROC)and area under the ROC curve were used to evaluate the prediction efficiency of the model.Results:(1)The incidence of cognitive frailty in elderly patients with CKD was 28.2%.(2)Univariate analysis showed that 26 variables with statistical significant(p<0.05)include age,marital status,way of living,education level,drinking,auxiliary tools,exercise,heart failure,diabetes,high blood uric acid,secondary hyperparathyroidism,kinds of drug,chronic pain,Barthel index,GDS-15 score,SSRS score,C-reactive protein,ferritin and albumin.Binary Logistic regression analysis showed that variables with statistical significant(p<0.05)include educational level,exercise,chronic pain,Barthel index,GDS-15 score and SSRS score.(3)The decision tree model consists of 6 intermediate nodes,8 leaf nodes and14 decision paths,with a depth of 4 layers.The basis of node segmentation is indicated in each branch.Among the 8 variables included in the model,Barthel index,GDS-15 score,SSRS score,albumin and education level were finally included in the decision tree model,while the other 3 variables including exercise,chronic pain and secondary hyperparathyroidism were not included in the final decision tree model.(4)The prediction accuracy of the decision tree model constructed in this study was 81.7%,the area under ROC curve was 0.791,the sensitivity was 0.719,the specificity was 0.828,the positive predictive value was 61.1%,and the negative predictive value was 89.5%.Conclusions:The prediction efficiency evaluation index of the decision tree model constructed in this study is medium,indicating that the model has good predictive ability.Barthel index,GDS-15 score,SSRS score,educational level and albumin were the decision factors for screening the decision tree model of cognitive frailty in elderly patients with CKD.Among them,Barthel index was the most important decision factor affecting cognitive frailty.Decision factors and decision paths were selected based on decision tree model,the combination of each branch node value segmentation,coupled with the visual characteristics of the decision tree model.It can facilitate clinical medical staff quickly intuitively judge in elderly patients with CKD risk of cognitive frailty,provide a new reference tool for cognitive frailty,and provide a basis for the formulation and implementation of early personalized interventions.