【摘要】 目的：急诊留观的老年患者存在衰弱的不良预后风险。为提升衰弱早期识别与干预效率，本研究旨在构建与验证急诊留观老年患者衰弱的风险预测模型。方法：选择淮安市第二人民医院2023年1月至2024年6月急诊留观的老年患者336例，随机分为训练集(168例)和验证集(168例)。用Logistic回归分析筛选急诊留观老年患者衰弱的独立危险因素，构建列线图并评估其区分度、一致性。结果：训练集168例急诊留观老年患者中，101例(60.12%)存在衰弱，验证集168例急诊留观老年患者中96例(57.14%)存在衰弱。Logistic回归分析结果显示：性别“女”(OR=4.431,P<0.001)、共病(OR=3.369,P=0.003)、营养风险筛查结果“有营养风险”(OR=2.811,P=0.006)、贫血(OR=2.258,P=0.031)、患有肌少症(OR=3.690,P=0.004)是急诊留观老年患者衰弱的独立危险因素。在训练集中，受试者操作特征(receiver operating characteristic,ROC)曲线的曲线下面积为0.791(95%CI 0.716～0.866)；校准曲线显示列线图预测急诊留观老年患者衰弱风险的结果和实际发生的情况相吻合，且Hosmer-Lemeshow拟合优度检验显示χ~2=7.575,P=0.372。在验证集中，ROC曲线下面积为0.786(95%CI 0.714～0.857)；校准曲线显示列线图预测急诊留观老年患者衰弱风险的结果和实际发生的情况相吻合，且Hosmer-Lemeshow拟合优度检验显示χ~2=7.755,P=0.458。训练集阈值概率为0.15～0.81，验证集阈值概率为0.13～0.85，该列线图对急诊留观老年患者的衰弱风险预测更有利。结论：基于女性、共病、营养风险筛查结果“有营养风险”、贫血、患有肌少症构建的预测急诊留观老年患者衰弱风险的列线图可辅助急诊医护人员快速识别高风险患者，从而优化和干预资源分配。更多还原

【Abstract】 Objective: Elderly patients under emergency observation are at risk of frailty, which may lead to adverse outcomes. To enhance early identification and intervention efficiency, this study aims to develop and validate a risk prediction model for frailty in elderly patients under emergency observation.Methods: A total of 336 elderly patients who underwent emergency observation at Huai’an Second People’s Hospital from January 2023 to June 2024 were selected and randomly divided into a training set(n=168) and a validation set(n=168). Logistic regression analysis was used to identify independent risk factors for frailty, and a nomogram was constructed. Its discrimination and calibration were evaluated.Results: In the training set, 101(60.12%) patients were frail, while 96(57.14%) patients were frail in the validation set. Logistic regression analysis revealed that female sex(OR=4.431, P<0.001), comorbidities(OR=3.369, P=0.003), nutritional risk(OR=2.811, P=0.006), anemia(OR=2.258, P=0.031), and sarcopenia(OR=3.690, P=0.004) were independent risk factors for frailty. In the training set, the area under the receiver operating characteristic(ROC) curve(AUC) was 0.791(95% CI 0.716 to 0.866). The calibration curve showed good agreement between predicted and observed frailty, and the HosmerLemeshow test yielded χ~2=7.575, P=0.372. In the validation set, the AUC was 0.786(95% CI 0.714 to 0.857). The calibration curve also showed good agreement, with HosmerLemeshow χ~2=7.755, P=0.458. The decision curve analysis demonstrated that the nomogram provided greater net benefit in predicting frailty within threshold probabilities of 0.15-0.81(training set) and 0.13-0.85(validation set).Conclusion: The nomogram based on female sex, comorbidities, nutritional risk, anemia, and sarcopenia can assist emergency healthcare providers in rapidly identifying elderly patients at high risk of frailty, thereby optimizing resource allocation and facilitating timely interventions.更多还原