成人颅内肿瘤手术患者非计划重返ICU风险预测模型的构建及验证  

Construction and validation of risk prediction model for unplanned ICU readmission in patients undergoing intracranial tumor surgery

作　　者：刘云云[1] 卢舒颖 欧梦仙 张娜芹 纪媛媛 崔丽 王军[1] LIU Yunyun;LU Shuying;OU Mengxian;ZHANG Naqin;JI Yuanyuan;CUI Li;WANG Jun(Department of Neurosurgery,Xuanwu Hospital Capital Medical University,Beijing,100053,China)

机构地区：[1]首都医科大学宣武医院神经外科,100053

出　　处：《中国护理管理》2025年第2期283-289,共7页Chinese Nursing Management

基　　金：首都卫生发展科研专项(首发2022-4-20114);首都医科大学宣武医院2024年度护理专项课题(HLQN2024008)。

摘　　要：目的 :分析神经外科成人颅内肿瘤患者术后非计划重返ICU的现状和影响因素,构建重返ICU风险预测模型并验证,以期为临床干预提供依据。方法 :采用便利抽样法,选取2019年12月至2022年9月北京市某三级甲等医院神经外科术后由ICU转出的成年颅内肿瘤术后患者2 383例。采用倾向性评分匹配法按1∶4成功匹配420例,前300例为训练集,后120例为验证集。采用二分类Logistic回归分析成人颅内肿瘤手术患者非计划重返ICU的影响因素,构建并验证预测模型。结果 :共纳入2 383例患者资料,非计划重返ICU 84例,成人颅内肿瘤手术患者非计划重返ICU发生率为3.52%。倾向性匹配评分后共420例纳入分析,再入其中84例纳入重返ICU组,336例纳入非重返ICU组。回归分析结果显示:气管插管、机械通气时长、转出ICU时存在中/重度意识障碍是其独立危险因素(P<0.05)。训练集所建立模型的ROC曲线下面积(AUC)为0.834,灵敏度为81.2%,特异度为77.0%,最佳临界值为0.206。外部验证集建立模型的AUC值为0.854,灵敏度为81.4%,特异度为82.6%。结论 :本研究建立的成人颅内肿瘤术后患者非计划重返ICU风险预测模型科学合理,预测效果良好,适用于临床筛查,可为医护人员早期识别高危人群并采取有针对性的干预措施提供参考。Objective:To explore the status quo and influencing factors of unplanned ICU readmission in patients undergoing intracranial tumor surgery so as to establish a prediction model and validate it,in order to provide a basis for clinical intervention.Methods:A total of 2383 patients undergoing intracranial tumor surgery admitted to the neurosurgery ICU of a tertiary grade A hospital in Beijing from December 2019 to September 2022 were enrolled by convenience sampling method.Propensity Score(PSM)was used to match 420 cases at 1:4,with the first 300 cases as the training set and the last 120 cases as the validation set.Univariate and multivariate logistic regression models were used to evaluate the independent influencing factors of unplanned readmission to ICU in patients undergoing intracranial tumor surgery,and the prediction model was constructed and validated.Results:The incidence of unplanned readmission to ICU in patients with intracranial tumor surgery was 3.52%(84/2383).A total of 420 cases were included in the analysis after PSM,of which 84 cases were included in the ICU readmission group,and 336 cases were included in the non-readmission group.Regression analysis showed that tracheal intubation,duration of mechanical ventilation,and the GCS score before ICU discharge were the influencing factors(P<0.05).The Area Under the ROC Curve(AUC)of the model established by the training set was 0.834 with the sensitivity being 81.2%,the specificity being 77.0%,and the best critical value being 0.206.The AUC value of the model established in the external validation set was 0.854 with a sensitivity of 81.4%and a specificity of 82.6%.Conclusion:The risk prediction model established in this study for patients with craniocerebral tumor who are not scheduled to return to ICU after surgery is scientific and reasonable,with good prediction effect,which is suitable for clinical screening,and can provide reference for medical staff to identify high-risk groups early to take targeted intervention measures.

关 键 词：成人 颅内肿瘤 非计划重返ICU 影响因素 预测模型 护理 

分 类 号：R47[医药卫生—护理学] R197[医药卫生—临床医学]