Establishment and validation of a prediction model for geriatric frailty syndrome in elderly patients with AIS after treatment

Chen Zhangjing; Kong Xianbo; Wang Guopin; Zhou Liqun; Wang Shanshan

doi:10.3969/j.issn.1009-0126.2023.12.025

Abstract：Objective To establish a prediction model for geriatric frailty syndrome(GFS)in elder-ly patients with acute ischemic stroke(AIS)after treatment.Methods Clinical data of 156 elderly AIS patients admitted to our hospital from January 2020 to December 2022 were collected and ret-rospectively analyzed.According to occurrence of GFS or not,they were divided into GFS group(n=57)and control group(n=99).The differences of clinical features were recorded and com-pared between the two groups of elderly AIS patients.Multivariate logistic regression model was used to analyze the risk factors for GFS in the elderly AIS patients.And a prediction model for GFS was constructed.Results Larger proportions of aged ≥80 years,diabetes,massive cerebral infarction and dysphagia were observed in the GFS group than the control group(P＜0.05,P＜0.01).Multivariate logistic regression analysis showed that aged ≥80 years(OR=2.890,95％CI:1.306-6.395,P=0.009),diabetes(OR=4.892,95％CI:2.172-11.018,P=0.000),massive cere-bral infarction(OR=3.363,95％CI:1.418-7.977,P=0.006)and dysphagia(OR=2.772,95％CI:1.123-6.844,P=0.027)were independent risk factors for GFS in the elderly AIS patients after treatment.A nomogram prediction model was constructed.Then the dataset was randomly divided into a training set and a validation set in a ratio of 7∶3.The AUC value was 0.840(95％CI:0.754-0.927)in the training set,and 0.676(95％CI:0.518-0.833)in the validation set.Hos-mer-Lemeshow Goodness-of-Fit test indicated that when the model was subjected to the valida-tion set,a Chi-square value of 14.394 and a P value of 0.072 were obtained.Conclusion Our no-mogram prediction model has good value in predicting the occurrence of GFS in elderly AIS pa-tients after treatment.