基于潜在药物相互作用的心血管疾病患者住院时间延长风险预测模型的构建  

Construction of Prediction Model for Prolonged Length of Stay in Patients with Cardiovascular Disease Based on Potential Drug Interactions

作　　者：席田兰 幸晓琼 杨佳丹[1] 王红梅[1] XI Tianlan;XING Xiaoqiong;YANG Jiadan;WANG Hongmei(Dept.of Pharmacy,the First Affiliated Hospital of Chongqing Medical University,Chongqing 400000,China;Dept.of Pharmacy,Xichang People’s Hospital,Sichuan Xichang 615000,China)

机构地区：[1]重庆医科大学附属第一医院药学部,重庆400000 [2]西昌市人民医院药学部,四川西昌615000

出　　处：《中国医院用药评价与分析》2023年第4期400-403,共4页Evaluation and Analysis of Drug-use in Hospitals of China

基　　金：重庆市科卫联合医学项目资助(No.2020MSXM120)。

摘　　要：目的:构建基于潜在药物相互作用(pDDIs)等因素的心血管疾病(CVD)患者住院时间延长风险预测模型。方法:收集2019年10月至2021年10月重庆医科大学附属第一医院CVD住院患者人口学资料、相关检查以及pDDIs等信息,采用单因素与多因素Logistic回归分析CVD患者住院时间延长的危险因素,构建列线图预测模型,并采用受试者工作特征(ROC)曲线及校准曲线对模型进行评价。结果:共纳入599例CVD住院患者,其中566例患者(占94.49%)存在至少1种pDDIs。多因素Logistic回归分析显示,禁忌pDDIs、体重指数和用药总数是CVD患者住院时间延长的独立危险因素(P<0.05)。基于此构建的预测模型ROC曲线下面积为0.821(95%CI=0.787~0.850),校准曲线预测值与实际值基本一致。结论:CVD患者中pDDIs发生率较高,基于pDDIs等因素构建的CVD患者住院时间延长风险预测模型具有较好的预测效能,可为临床决策提供相关参考。OBJECTIVE:To construct a risk prediction model of prolonged length of stay(LOS)in patients with cardiovascular disease(CVD)based on potential drug-drug interactions(pDDIs)and other factors.METHODS:The demographic data,related examinations,pDDIs and other information of CVD inpatients in the First Affiliated Hospital of Chongqing Medical University from Oct.2019 to Oct.2021 were collected,and univariate and multivariate logistic regression was used to analyze the risk factors for prolonged LOS in CVD patients,and a nomogram prediction model was constructed,the receiver operating characteristic(ROC)curve and calibration curve were used to evaluate the model.RESULTS:A total of 599 inpatients with CVD were included,of which 566 cases(94.49%)had at least one pDDIs.Multivariate Logistic regression analysis showed that contraindicated pDDIs,body mass index and total number of medications were independent risk factors for prolonged LOS in CVD patients(P<0.05).The area under the ROC curve of the model was 0.821(95%CI=0.787-0.850),and the predicted value of the calibration curve was basically consistent with the actual value.CONCLUSIONS:The incidence of pDDIs in CVD patients is high,and the risk prediction model of prolonged LOS in CVD patients based on pDDIs and other factors has effective predictive performance and can provide relevant reference for clinical decision-making.

关 键 词：潜在药物相互作用 危险因素 住院时间 预测模型 心血管疾病 

分 类 号：R969.2[医药卫生—药理学]