A scoping review of frailty risk prediction models in Chinese patients with diabetes

HU Shiya; WU Peiyun; HUANG Renling; LIN Fenglan

doi:10.3969/j.issn.2096-9058.2025.06.002

Abstract：Objective To conduct a scoping review of the risk prediction models for frailty in Chinese patients with diabetes.Methods PubMed,Embase,Web of Science,EBSCO,The Cochrane Library,CNKI,Wanfang,VIP,and CBM were systematically searched to collect Chinese and English literature on the risk prediction models for frailty in Chi-nese patients with diabetes.The retrieval period was from the establishment of the database to February 22,2025.Two re-searchers independently screened and extracted information from the literature,and the quality of the included literature was evaluated using the prediction model bias risk assessment tool.Results A total of 17 studies and 26 diabetes frailty risk prediction models were included,and the incidence of frailty in Chinese diabetic patients was 10.9%-51.2%.The modeling methods included logistic regression and machine learning algorithms.All models reported the AUC(0.768-0.975),and 15 studies calibrated the models.The risk of bias of the included literature was high,mainly in the research subjects and analysis fields,and the overall applicability was good.Age,depression and glycated hemoglobin level were the important predictors of frailty in Chinese patients with diabetes.Conclusions The included models have good predic-tive performance,but the risk of bias is high.Future research should focus on developing or validating risk prediction mod-els with low risk of bias and high applicability.