老年慢性病住院患者营养不良的风险预测模型构建

Construction of a risk prediction model of malnutrition in elderly hospitalized patients with chronic diseases

迟晨汝 ¹陈梦奇 ²杨丹 ²刘欢 ²周志庆²

扫码查看

作者信息

1. 安徽医科大学第一附属医院
2. 皖南医学院第一附属医院弋矶山医院,安徽芜湖 241001
折叠

摘要

目的:构建老年慢性病住院患者营养不良的风险预测模型,为临床医护人员早期识别及干预患者的营养不良提供科学依据.方法:2023年3月—2023年7月采用便利抽样法选取某三甲医院老年慢性病住院患者为研究对象进行问卷调查,通过单因素及多因素分析筛选预测因素,并构建老年慢性病住院患者营养不良的风险预测模型,分别采用校准曲线、受试者操作特征(Receiver operating characteristic,ROC)曲线和临床决策曲线评价列线图模型的准确性、区分度和有效性.结果:老年慢性病住院患者营养不良检出率为36.1%,单因素分析发现性别、年龄、自理程度、睡眠时长(夜间+午休)、衰弱情况、抑郁水平、血红蛋白水平为老年慢性病住院患者营养不良状况的影响因素(P＜0.001),多因素Logistic回归分析结果显示衰弱、抑郁、女性是老年慢性病住院患者营养状况的危险因素,血红蛋白水平是老年慢性病住院患者营养状况的保护因素(P＜0.05).基于以上回归分析结果构建的预测模型校准曲线拟合度良好;ROC曲线下面积为0.772,95%CI为0.719～0.825,灵敏度为73.1%,特异度为72.0%,约登指数为0.451;决策曲线示模型具有良好的临床实用价值.结论:本研究构建的老年慢性病住院患者营养不良的风险预测模型预测效能良好,可作为临床医护人员早期识别营养不良高风险人群的评估工具.

Abstract

Objective:To construct the risk prediction model of malnutrition in elderly hospitalized patients with chronic diseases to provide scientific basis for clinical staff to identify the occurrence of malnutrition and intervene in the early stage.Methods:From March 2023 to July 2023,the convenience sampling method was used to select elderly hospitalized patients with chronic diseases in a Grade A tertiary hospital as the research objects for questionnaire survey.The predictive factors were screened through univariate and multi-factor analysis,and the risk prediction model of malnutrition in elderly hospitalized patients with chronic diseases was built.Receiver operating characteristic(ROC)curve,calibration curve and clinical decision curve were used to evaluate the differentiation,accuracy and validity of the nomogram model.Results:The detection rate of malnutrition in elderly inpatients with chronic diseases was 36.1%.Multivariate logistic regression analysis showed that gender,age,degree of self-care,length of sleep(night and noon time rest),frailty status,depression level and hemoglobin level were independent influencing factors of malnutrition in elderly inpatients with chronic diseases(P＜0.001);multi-factor Logistic analysis showed that frailty status,depression level and female were the risk factors,hemoglobin level was protective factor(P＜0.05).The area under the ROC curve of the prediction model constructed based on the results of the above regression analysis was 0.772,the 95%CI was 0.719-0.825,the sensitivity was 73.1%,the specificity was72.0%,and the Youden index was 0.451.Decision curve model had good clinical value.Conclusion:The established risk prediction model of malnutrition in elderly hospitalized patients with chronic diseases has good prediction efficiency.It can be used as an assessment tool for clinical medical staff to identify high-risk groups of malnutrition in early stage.

关键词

老年人/慢性病/住院患者/营养不良/风险预测模型

Key words

The aged/Chronic diseases/Hospitalized patients/Malnutrition/Risk prediction model

引用本文复制引用

基金项目

安徽省新时代育人质量工程项目(2022jyjxggyj338)

皖南医学院弋矶山医院管理与服务创新项目(CX2022013)

出版年

2024

赣南医学院学报

赣南医学院

赣南医学院学报

影响因子：0.622

ISSN：1001-5779

段落导航