老年患者中央导管相关血流感染风险预测评分模型构建与验证

目的构建老年患者中央导管相关血流感染(CLABSI)风险预测评分模型,为筛选高危人群,有效预防与控制老年患者血流感染提供依据。方法依据纳入排除标准,收集2015年1月1日—2017年12月31日住院期间曾留置中央导管的老年患者病例资料,按照7∶3的比例随机分为建模组和验证组(随机种子为20180708),对建模组数据进行危险因素识别,构建logistic回归模型,根据β值赋予各危险因素相应的分值,建立感染风险评分模型,利用受试者工作特征(ROC)曲线评价模型的预测准确度;依据建立的感染风险评分模型对验证组病例进行评分,利用ROC曲线评价模型的预测准确度。利用R软件构建决策曲线。结果 logistic回归分析结果表明:本次住院手术次数≥3次、住ICU日数≥2 d、中心静脉置管日数≥7 d、使用抗菌药物等是老年患者发生CLABSI的独立危险因素;风险评分模型中相应的分值分别为3、4、4、9分,得分13～17分为高风险人群;评分模型在建模组数据中ROC曲线下面积(AUC)为0.74;依据验证组患者风险得分情况绘制ROC曲线,曲线下面积(AUC)为0.70。决策曲线显示,在阈值0.01～0.05区间内风险评分模型的净获益较高。结论建立的风险评分模型具有较好的判别效度和应用价值,可用于老年患者CLABSI的易感高危人群识别,做到早期预防与控制。

Establishment and validation of risk prediction scoring model of central line-associated bloodstream infection in elderly patients

Objective To establish the risk prediction scoring model of central line-associated bloodstream infection (CLABSI) in elderly patients, provide basis for screening high-risk population, and effectively prevent and control bloodstream infection(BSI) in elderly patients. Methods According to the inclusion and exclusion criteria, data of elderly patients who underwent central catheterization during hospitalization from January 1, 2015 to December 31, 2017 were collected, patients were randomly divided into modeling group and verification group according to the ratio of 7:3 (random seed was 20180708), risk factors of data of modeling group were distinguished, logistic regression model was constructed. The corresponding score of each risk factor was assigned according to β value, infection risk scoring model was established, prediction accuracy of model was evaluated by receiver operating characteristic (ROC) curve; according to the established infection risk scoring model, cases in validation group were scored, prediction accuracy of model was evaluated by ROC curve. Decision curve was constructed using R software. Results Logistic regression analysis showed that the number of operation ≥ 3 times, length of stay in intensive care unit(ICU) ≥ 2 days, duration of indwelling central venous catheterization ≥ 7 days, and use of antimicrobial agents were independent risk factors for CLABSI in elderly patients; the corresponding points in risk prediction scoring model were 3, 4, 4, and 9 respectively, and the score of 13-17 points were high-risk population of CLABSI; the area of ROC curve (AUC) was 0.74 in the modeling group; ROC curve was plotted based on the risk score of patients in validation group, AUC was 0.70. The decision curve showed that the net benefit of the risk scoring model was higher in the high risk thershold of 0.01-0.05. Conclusion The established risk prediction scoring model has good discriminant validity and application value, and can be used in the identification of susceptible high risk population of CLABSI in elderly patients, so as to achieve early prevention and control.