社区老年高血压患者衰弱的动态列线图风险预测模型构建

Web-based dynamic nomogram for predicting frailty in community elderly patients with hypertension

目的 构建动态列线图预测模型,分析社区老年高血压患者衰弱的影响因素,为制定针对性的干预措施提供参考。方法 从中国健康与养老追踪调查随访数据库中提取高血压患者信息,以7∶3比例随机分为训练集(n=1 160)与验证集(n=494)。采用Lasso法筛选最佳预测变量,使用logistic回归模型分析高血压患者衰弱影响因素,并构建动态列线图。使用ROC曲线的曲线下面积、Hosmer-Lemeshow检验、校准曲线和决策曲线分析评估列线图的预测性能。结果 共筛选出1 654例老年高血压患者,其中560例(33.86%)并发衰弱。受教育程度、握力、BMI、抑郁、认知障碍、自评健康、代谢性疾病、心脑血管疾病、呼吸系统疾病、胃肠道疾病10个变量纳入预测模型。预测模型在训练集和验证集的ROC曲线下面积分别为0.883(95%CI为0.863~0.903)和0.887(95%CI为0.857~0.916);Hosmer-Lemeshow检验值分别为P=0.825和P=0.410;校准曲线显示预测值和实际值之间存在显著一致性。决策曲线分析显示该模型具有良好的净效益和预测准确性。结论 动态列线图具有良好预测性能,可为社区医护人员评估高血压患者衰弱风险提供便捷有效的工具。

Objective To construct a dynamic nomogram prediction model,to analyze the influencing factors of frailty in elderly patients with hypertension in the community,and to provide a reference for developing targeted intervention.Methods Information on patients with hypertension was extracted from the China Health and Retirement Longitudinal Survey and randomly divided into a training set(n=1 160) and a validation set(n=494) in a ratio of 7∶3.The best predictors were screened by using LASSO me-thod,and the logistic regression model was used to analyze the influencing factors of frailty in hypertensive patients and develop a nomogram prediction model.Area under the ROC curve(AUC),Hosmer-Lemeshow test,calibration curve and decision curve analysis(DCA) were used to evaluate the predictive performance of the nomogram prediction model.Results A total of 1 654 elderly hypertensive patients were screened,of whom 560(33.86%) were complicated by frailty.Such 10 variables as education level,grip strength,BMI,depression,cognitive impairment,self-rated health,metabolic diseases,cardiovascular/cerebrovascular diseases,respiratory diseases,and gastrointestinal diseases were included in the prediction model.The AUC of the ROC curve of the prediction model were 0.883(95%CI:0.863-0.903)in the training set and 0.887(95%CI:0.857-0.916) in the validation set respectively.The P value of the Hosmer-Lemeshow test was 0.825 in the training set and 0.410 in the validation set respectively.The calibration curves showed a favorable consistency between predicted and actual values.DCA showed that the model had good net benefits and predictive accuracy.Conclusion This dynamic nomogram has good predictive performance,which can serve as a convenient and effective tool for community medical staff to assess the risk of frailty in patients with hypertension.