Risk Factors Of Frailty In Elderly Patients With Heart Failure And Construction And Validation Of Risk Prediction Model

Objective:The risk nomogram prediction model of frailty in elderly patients with heart failure was constructed to provide a screening tool for medical workers to identify frailty in patients with heart failure early,in order to improve the long-term prognosis and quality of life of elderly patients with heart failure.Methods:In this study,330 elderly patients with heart failure who were hospitalized in the Department of Cardiology and Geriatrics of a tertiary hospital in Lanzhou from August2021 to November 2022 were collected by convenient sampling method and divided into modeling group and verification group.Questionnaire data were collected one by one by the researcher himself.The survey tools were:general information questionnaire,Tilburg Frailty Indicator(TFI),Charlson comorbidity index(CCI),Activities of daily living Scale(ADL).In this study,the collected data were entered by Epidata 3.1 software,and analyzed by SPSS24.0statistical software.The measurement data were expressed as mean±standard deviation,and the frequency and percentage of count data were expressed as frequency and percentage.With the presence or absence of frailty as the dependent variable,all independent variables were subjected to univariate Logistic regression analysis,and the statistically significant single factors were further included in Logistic regression for multivariate analysis to establish a risk nomogram prediction model.Hosmer-Lemeshow goodness of fit test and receiver operating characteristic curve(ROC)were used to evaluate the predictive efficacy of the model.The application efficiency of the model was verified by sensitivity and specificity.Results:(1)General information of the research object:This study finally included 330 elderly patients with heart failure as the research object.,including 149 males and 181 females.According to the Tilburg Frailty Assessment Scale(TFI),the patients were divided into frailty group(n=179,54.2%)and non-frailty group(n=151,45.8%).(2)Single factor analysis results:Univariate analysis was performed on elderly heart failure patients with and without frailty.There were significant differences in age,marital status,living conditions,fall history,cardiac function classification,hospitalization times,CCI score and ADL score between the two groups(P<0.05).There was no significant difference in gender,education level,sleep status,smoking history,drinking history,physical activity,medication type,hemoglobin,creatinine,N-terminal B-type natriuretic peptide(NT-pro BNP),left ventricular ejection fraction(LVEF)and uric acid between the two groups(P>0.05).(3)Multivariate analysis results:Multivariate Logistic regression analysis showed tha t a total of 4 variables entered the final prediction model.Age(OR=1.122,95%CI:1.038～1.212,P=0.004),history of falls(OR=1.330,95%CI:1.124～1.884,P=0.027),Charlson com orbidity index questionnaire(OR=1.952,95%CI:1.173～3.246,P=0.010)and activity of dai ly living scale(OR=1.196,95%CI:1.110～1.288,P<0.001)were independent risk factors for frailty in elderly patients with heart failure.(4)Parameters of risk nomogram prediction model:According to the results of Logistic regression analysis,age,fall history,Charlson comorbidity index questionnaire(CCI)and activities of daily living scale(ADL)were used as predictors,and whether frailty occurred was used as outcome index.The regression coefficient of each variable obtained by logistic regression analysis was used as its weight in the prediction model to establish a risk nomogram prediction model for frailty in elderly patients with heart failure.(5)Verification of the model:Hosmer-Lemeshow test showed that there was no sign ificant difference between the predicted occurrence of frailty and the actual occurrence(χ~2=2.696,P=0.952).The area under the ROC curve was 0.949(95%CI:0.912～0.987).Wh en the optimal cutoff value of the prediction model was 0.4527,the Youden index reac hed the maximum value of 0.744,and the sensitivity was 82.4%and the specificity wa s 92.0%.It shows that the model has predictive ability and discrimination ability.Conclusion:(1)In this study,the statistical method of Logistic regression was used to construct the model,and a risk nomogram prediction model suitable for frailty in elderly patients with heart failure was established.The final independent variables included age,fall history,Charlson comorbidity index questionnaire(CCI),and activity of daily living scale(ADL).(2)The model has good predictive ability and has certain popularization and application.It can provide a screening tool for clinical medical workers to identify frailty in patients with heart failure early.