目的 基于Logistic回归和人工神经网络构建老年糖尿病足（diabetic foot,DF）患者衰弱风险预测模型，并比较两种模型预测效能，为早期识别并预防老年DF患者衰弱的发生提供依据。方法 2023年5-10月，采用便利抽样法选取天津市某两所三级甲等医院内491例老年DF患者为研究对象。通过问卷调查及病历记录收集资料，绘制列线图模型及人工神经网络模型；受试者工作特征曲线和曲线下面积评估模型预测能力，敏感度和特异度评估模型预测价值。结果 建模组列线图和人工神经网络模型的曲线下面积（area under curve,AUC）分别为0.973、0.742，敏感度分别为92.90%、95.50%，特异度分别为91.10%、50.50%。结论 构建的老年DF患者衰弱风险预测的列线图模型预测性能较好，对有效识别高衰弱风险的老年DF患者有临床价值。

Objective To construct a frailty risk prediction model in elderly patients with diabetic foot (DF) based on logistic regression and artificial neural network,and to compare the prediction efficacy of the two models,in order to provide a basis for the early identification and prevention of the occurrence of frailty in elderly DF patients.Methods Convenience sampling method was used to select 491 elderly DF patients in two tertiary A hospitals in Tianjin from May to October 2023 as research objects.Data were collected through questionnaires and medical records,and the nomogram model and artificial neural network model were drawn.The receiver operating characteristic curve and area under curve (AUC) were used to assess the predictive ability of the models ,and the sensitivity and specificity were used to assess their predictive value.Results The AUC of the modeling group’s nomogram and artificial neural network model were 0.973 and 0.742,with a sensitivity of 92.90% and 95.50% and a specificity of 91.10% and 50.50%,respectively.Conclusions The nomogram model constructed for frailty risk predicting in elderly DF patients has better predictive performance and clinical value for effectively identifying elderly DF patients with high frailty risk.

R473.58

天津市研究生科研创新项目（2022SKY239）；天津中医药大学研究生科研创新项目（YJSKC-20221033）