老年骨折患者术后低钾血症风险预测模型的构建

The construction of a risk prediction model for postoperative hypokalemia in elderly patients with fracture

目的探讨影响老年骨折患者术后低钾血症发生的危险因素,构建预测老年骨折术后低钾血症风险的预测模型.方法回顾性研究两家三级甲等综合医院的骨科病房符合标准的患者病历资料260例作为建模数据集,102例作为验证数据集.采用Logistic多元回归分析低钾血症的独立危险因素并构建风险预测模型,应用受试者操作特征(ROC)曲线分析该模型的预测效能,并应用STATA软件建立老年骨折患者术后低钾血症列线图.结果多因素Logistic回归分析显示年龄、既往史疾病种类、利尿剂、关节置换术、手术持续时间长是老年骨折患者术后低钾血症的独立危险因素(P<0.05).构建老年骨折患者术后低钾血症风险预测模型及列线图,ROC曲线下面积为0.859,约登指数为0.593,敏感度为0.800,特异度为0.793,准确率为79.4%,提示该模型具有较好的精准度和区分度.结论本资料建立的老年骨折患者术后低钾血症风险预测模型具有良好的区分度和精准度,列线图为临床识别低钾血症高风险患者提供可视化工具,对制定相应的预防和干预措施有临床指导意义.

Objective To explore the risk factors affecting postoperative hypokalemia in elderly patients with fracture,and to construct a prediction model for predicting the risk of postoperative hypokalemia in elderly patients with fracture.Methods The convenience sampling method was used to retrospectively study the medical records of 260 patients who met the criteria in the orthopedic wards of two tertiary general hospitals in a city as the modeling data set and 102 as the validation data set.Logistic multiple regression analysis was used to analyze the independent risk factors of hypokalemia and a risk prediction model was constructed.The receiver operating characteristic(ROC)curve was used to analyze the prediction efficiency of the model,and STATA software was used to establish the line chart of postoperative hypokalemia in elderly patients with fracture.Results Multivariate Logistic regression analysis showed that age,type of medical history,diuretics,joint replacement and duration of operation were independent risk factors for postoperative hypokalemia in elderly fracture patients(P<0.05).A risk prediction model and nomogram of postoperative hypokalemia in elderly fracture patients were constructed.The area under the ROC curve was 0.859,the Youden index was 0.593,the sensitivity was 0.800,the specificity was 0.793,and the accuracy was 79.4%,which suggested that the model had good prediction performance.Conclusion This study establishes a risk prediction model of postoperative hypokalemia elderly patients with fracture,which has good discrimination and accuracy.The nomogram provides a visual tool for clinical identification of patients with high risk of hypokalemia,and has clinical guiding significance for formulating corresponding prevention and intervention measures.