2022 年第 11 期 第 17 卷

急诊多发伤患者谵妄的影响因素及其风险预测列线图模型的构建

Influencing factors of delirium in emergency patients with multiple trauma and the construction of its risk prediction nomogram model

作者：李春萍曾艳蔡少青杨露

英文作者：Li Chunping Zeng Yan Cai Shaoqing Yang Lu

单位：海南省人民医院急诊科，海口570311

英文单位：Department of Emergency Hainan Provincial People′s Hospital Haikou 570311 China

关键词：多发伤；急诊；谵妄；危险因素；列线图

英文关键词：Multipletrauma;Emergency;Delirium;Riskfactors;Nomogram

• 摘要：

• 目的 探讨急诊多发伤患者谵妄的影响因素，并构建预测谵妄发生风险的列线图模型。方法 选择2016年1月至2021年12月海南省人民医院急诊科重症监护病房（ICU）收治的多发伤患者426例为研究对象行回顾性研究，其中2016年1月至2020年12月的300例作为建模队列，2021年1—12月的126例作为验证队列，采用ICU患者意识模糊评估量表评估患者是否存在谵妄，并将患者分为谵妄组和非谵妄组，收集患者的基本资料和临床资料，采用单因素分析和Logistic回归方法分析多发伤患者谵妄的危险因素，应用R语言软件建立列线图预测模型，采用一致性指数验证模型区分度。结果 426例患者中发生谵妄137例，谵妄发生率为32.2%，建模队列和验证队列谵妄发生率分别为32.7%（98/300）和31.0%（39/126）。Logistic回归分析结果显示，年龄≥60岁、损伤严重程度评分（ISS）≥20分、急性生理学与慢性健康状况评分系统Ⅱ（APACHE Ⅱ）评分≥13分、入住ICU时间≥9 d、Richmond躁动-镇静评分（RASS）≥0分、术后疼痛视觉模拟量表评分≥7分是谵妄的独立危险因素（均P<0.05）。将上述6个危险因素构建列线图模型，一致性指数为0.815，说明模型具有良好的区分度；预测建模队列和验证队列多发伤患者发生谵妄风险的受试者工作特征曲线下面积分别为0.817、0.875，均>0.8，模型准确度高、拟合度较好。结论 年龄≥60岁、高ISS、高APACHE Ⅱ评分、入住ICU时间长、高RASS及疼痛是多发伤患者发生谵妄的独立危险因素，基于多种因素构建的列线图可用于预测多发伤患者发生谵妄的风险。

• Objective To investigate the influencing factors of delirium in emergency patients with multiple trauma, and to construct nomogram model for predicting delirium risk. Methods A total of 426 patients with multiple trauma who were admitted to the intensive care unit(ICU) of the Department of Emergency, Hainan Provincial People′s Hospital from January 2016 to December 2021 were selected retrospectively, of which 300 patients from January 2016 to December 2020 were divided into the modeling cohort, and 126 patients from January to December 2021 were divided into validation cohort. The Confusion Assessment Method for the ICU was used to assess delirium, and patients were divided into delirium group and non-delirium group. The general and clinical data of patients were collected. Univariate analysis and Logistic regression analysis were used to analyze the risk factors of delirium in patients with multiple trauma. The nomograph prediction model was established by R language software, and the consistency index was used to verify the model differentiation. Results Delirium occurred in 137 of the 426 patients, and the incidence of delirium was 32.2%. The incidence of delirium in the modeling cohort and validation cohort was 32.7%（98/300） and 31.0%（39/126）, respectively. Logistic regression analysis showed that age≥60 years old, injury severity score (ISS)≥20, acute physiology and chronic health evaluation system Ⅱ (APACHE Ⅱ) score≥13, ICU stay≥9 d, Richmond agitation-sedation score (RASS)≥0 and postoperative pain visual analogue scale score≥7 were independent risk factors for delirium(all P<0.05). A nomogram model was constructed by the above 6 risk factors, and the consistency index was 0.815, indicating that the model had a good degree of discrimination; the areas under the receiver operating characteristic curve for predicting the delirium risk of patients with multiple trauma in the modeling cohort and the validation cohort were 0.817 and 0.875 respectively, both of which were>0.8, and the accuracy of the model was high and the fitting was good. Conclusion Age≥60 years, high ISS, high APACHE Ⅱ score, long stay in ICU, high RASS and pain are independent risk factors for delirium in patients with multiple trauma.