Construction Of A Predictive Model For Frailty Risk In Maintenance Hemodialysis Patients

Research Purpose:1.To understand the current situation of frailty in patients with maintenance hemodialysis and analyze its related influencing factors,establish a frailty risk prediction model and draw a line graph for patients with maintenance hemodialysis,so as to provide medical staff with a screening tool for frailty in patients with maintenance hemodialysis.2.To perform external validation of the model for predicting the risk of frailty in maintenance hemodialysis patients and evaluate its extrapolation and prediction ability.Research Methods:1.Three hundred and twenty maintenance hemodialysis patients in a tertiary general hospital in Zhejiang Province from March 2021 to August 2021 were selected as the research objects of the modeling group.The frailty was evaluated with the frailty phenotype scale by the questionnaire survey and combined with case data for modeling.Single factor variables with statistically significance（P < 0.05）were subjected to the Logistic stepwise regression analysis to identify independent influencing factors.The regression equation was constructed based on the partial regression coefficients and intercepts of the influencing factors,to complete the construction of the frailty risk prediction model for maintenance hemodialysis patients and the R software was used to construct the alignment diagram.2.One hundred maintenance hemodialysis patients in a tertiary hospital in Zhejiang Province from September 2021 to December 2021 were selected as the research objects of the validation group.The frailty was evaluated with the frailty phenotype scale by the questionnaire survey and combined with case data for the external verification of the model.The areas under the ROC curve,the Hosmer-Lemeshow goodness-of-fit test and the calibration diagram were used to evaluate the actual prediction ability of the model.Research Results:1.The incidence of frailty in 320 maintenance hemodialysis patients in the modeling cohort was 74.06%（frailty 31.87% and pre-frailty 42.19%）.Logistic regression analysis showed that age,cerebrovascular disease and heart disease were the risk factors for frailty in maintenance hemodialysis patients（P < 0.05）.Physical activity score and perceived social support score were protective factors（P < 0.05）.The risk prediction model was constructed based on the five influencing factors.The results of the model showed that the area under the ROC curve was 0.893（95%CI 0.856～0.931）,indicating good differentiation.The Hosmer-Lemeshow test value was10.638（P=0.155 > 0.05）,indicating good calibration.The Joden index of the model and the critical value were 0.676 and 0.778,respectively,and the corresponding sensitivity and specificity were77.2% and 90.4%.2.In the validation cohort,the incidence of frailty in the one hundred patients with maintenance hemodialysis was 75.00%（frailty 33.00% and pre-frailty 42.00%）.The area under the ROC curve was 0.865（95 % CI 0.791～0.940）,indicating a good discrimination of the model.The Hosmer-Lemeshow test P = 0.502,indicating that the model had good calibration.The calibration graph showed that the predicted probability of frailty in maintenance hemodialysis patients was consistent with the actual probability,indicating that the model had a certain extrapolation ability.Research Conclusion:The frailty risk prediction model of maintenance hemodialysis patients: Logit（P）= 8.936 +0.564 * age（40-60 years old）+ 1.672 * age（>60 years old）+ 1.047 * cerebrovascular disease +1.481 * heart disease-0.842 * international physical activity score-0.167 * perceived social support score.The model shows a good discrimination and calibration by external validation and has certain extrapolation ability.