Chinese Journal of Nursing ›› 2021, Vol. 56 ›› Issue (1): 14-20.DOI: 10.3761/j.issn.0254-1769.2021.01.002

• Research Paper • Previous Articles     Next Articles

Construction of a risk prediction model for post-intensive care syndrome-cognitive impairment

WEI Yueqing(),LI Hong(),LI Yun,WU Jingbing,ZHANG Zhihong,ZHENG Yan   

  1. Department of Respiratory and Critical Care Medicine,Shengli Clinical Medical College of Fujian Medical University,Fuzhou,350001,China
  • Received:2020-04-06 Online:2021-01-15 Published:2021-01-15
  • Contact: Hong LI

ICU后认知障碍风险预测模型的构建及验证

魏月清(),李红(),李芸,吴静冰,张智宏,郑艳   

  1. 350001 福州市 福建医科大学省立临床医学院呼吸与危重症医学科(魏月清),护理部(李红),重症医学二科(李芸,吴静冰),血液净化科(张智宏),感染性疾病科(郑艳)
  • 通讯作者: 李红
  • 作者简介:魏月清:女,本科(硕士在读),主管护师,E-mail: 272205811@qq.com

Abstract:

Objective To explore the risk factors of Post-Intense Care Syndrome-Cognitive Impairment(PICS-CI) in critically ill patients,and to build their risk prediction model. Methods A total of 481 ICU patients from 2 Level A tertiary hospitals in Fujian Province were selected,and divided into a cognitive impairment group(n=215) and a non-cognitive impairment group(n=266) according to their cognitive score of 7 days after being transferred out of ICU. The demography,disease,treatment,physiological and laboratory indicators between the 2 groups were compared;the risk factors of PICS-CI were screened out,and Logistic regression was used to establish a risk prediction model. 118 patients from another 4 hospitals were selected to verify the model prediction. Results Age(OR=1.035),delirium(OR=10.488),sepsis(OR=1.925),propofol dose(OR=1.098),sleep disorder(OR=0.932) are the independent risk factors of PICS-CI. These 5 factors are used to construct a prediction model,which was internally verified by the modeling group. The calibration curve of the calibration chart is close to the ideal curve;the area under the ROC curve is 0.838;the risk prediction value 0.521 corresponding to the maximum Youden index is the best value;the prediction critical value is 50 points. The external verification shows that the calibration curve of the calibration chart is near the ideal curve,and the area under the ROC curve drawn is 0.797. Conclusion The PICS-CI prediction model constructed in this study shows good prediction efficiency. ICU patients with scores≥50 should receive close attention and early interventions in early stage.

Key words: Intensive Care Units, Cognition Disorders, Risk Factors, Critical Patients, Prediction Model, Nursing Gare

摘要:

目的 探讨危重症患者发生ICU后认知障碍的危险因素,并构建其风险预测模型。 方法 2019年1月—10月,选取福建省某2所三级甲等综合性医院的危重症患者481例,根据患者转出ICU后7 d认知功能分为认知障碍组(n=215)和非认知障碍组(n=266),将两组间一般资料、疾病因素、治疗因素、实验室检查指标等进行比较,筛选出独立危险因素并利用Logistic回归进行风险预测模型的构建,2019年11月—12月,选取福建省另外4所医院的危重症患者118例对模型进行验证。 结果 最终纳入年龄(OR=1.035)、谵妄(OR=10.488)、脓毒症(OR=1.925)、丙泊酚累积剂量(OR=1.098)、睡眠评分(OR=0.932)5个因素构建出风险预测模型。内部验证:Calibration图示校准曲线与理想曲线接近重合;ROC曲线下面积为0.838;最大Youden指数所对应的风险预测值为0.521,预测临界值为50分。外部验证:Calibration图示校准曲线在理想曲线附近;ROC曲线下面积为0.797。 结论 该研究所构建的ICU后认知障碍风险预测模型一致性和预测效能较好,对于分数≥50分的危重患者,应高度关注其是否有认知障碍,并实施早期干预。

关键词: 重症监护病房, 认知障碍, 危险因素, 危重患者, 预测模型, 护理