ICU患者医用粘胶相关性皮肤损伤发生风险预测模型的建立与验证  

Establishment of risk prediction model for medical adhesive-related skin injuries in ICU patients and its validation

作　　者：谢婷 张桂屏[1] 张娟[1] 李义稳 邹旭 XIE Ting;ZHANG Guiping;ZHANG Juan;LI Yiwen;ZOU Xu(EICU,Guizhou Provincial People's Hospital,Guiyang 550002,Guizhou,China;Guizhou Nursing Vocational College,South Guizhou 551304,Guizhou,China;Guizhou Education University,Guiyang 550018,Guizhou,China)

机构地区：[1]贵州省人民医院EICU,贵州贵阳550002 [2]贵州护理职业技术学院,贵州黔南551304 [3]贵州师范学院,贵州贵阳550018

出　　处：《护士进修杂志》2024年第20期2173-2178,共6页Journal of Nurses Training

基　　金：贵州省卫生健康委科学技术基金项目(编号:gzwkj2022-243)。

摘　　要：目的构建ICU患者医用粘胶相关性皮肤损伤(medical adhesive-related skin injuries,MARSI)发生风险列线图预测模型,并对其进行验证。方法选取2022年3-12月入住贵州省某三甲医院治疗的261例患者进行数据监测,通过logistic回归分析法筛选相关因素建立MARSI发生风险列线图预测模型,并采用受试者工作特征ROC曲线、校准曲线及决策曲线对预测模型进行拟合效应检验。结果回归分析结果显示:年龄、机械通气、APACHEⅡ评分、过敏史、发热、低蛋白血症、皮肤水肿及使用镇静剂是MARSI发生的影响因素。以此8个变量建立ICU患者MARSI发生风险列线图预测模型,模型的ROC曲线下面积为0.85(95%CI=0.824~0.886,C-index指数为0.85),模型拟合度检测(Hosmer-Lemeshow)χ2=4.621,P=0.797,预测值与实际值基本一致,此模型具有较好的预测效果。结论本研究构建的风险列线图预测模型具有较好的预测效果,可为临床护理人员预测ICU患者MARSI的发生风险,以尽早制订个性化预防策略提供参考。Objective To establish a prediction model of the risk of medical adhesive-related skin injury(MARSI)in ICU patients,and verify it.Methods A total of 261 patients admitted to a top-three hospital in Guizhou province from March to December 2022 were selected for data monitoring,and logistic regression analysis was used to screen relevant factors to establish a nomogram prediction model for MARSI,and receiver operating characteristic(ROC)curve,calibration curve,and decision curve were used to test the fitting effect of the prediction model.Results The results of logistic regression analysis revealed that age,mechanical ventilation,APACHEⅡscore,history of allergy,fever,hypoproteinemia,skin edema,and use of sedatives were the factors influencing the development of MARSI.Using these 8 influencing factors identified,a well-fitting nomogram model for MARSI risk prediction was established,with the area under the ROC curve of 0.85(95%CI was 0.824-0.886)and the C-index of 0.85,and there was no significant difference between the model-predicted data and actual data by model fit degree detection(Hosmer-Lemeshow test;χ2=4.621,P=0.797)and the model had a good prediction effect.Conclusion The nomogram prediction model established in this study is of good prediction effect,which can provide a reference for clinical nurses to predict the risk of MARSI in ICU patients and formulate personalized prevention strategies as soon as possible.

关 键 词：重症监护 医用粘胶相关性皮肤损伤 列线图 预测模型 

分 类 号：R473[医药卫生—护理学] R473.75[医药卫生—临床医学]