预测医院感染发生概率的列线图模型研究
|
摘要
目的探索可用于预测医院感染发生的列线图模型。方法回顾性收集2015-2016年于宝鸡市中心医院的住院患者的临床信息,采用1∶2匹配的方式,得到发生医院感染的感染组1 523例,对照组3 046例。采用Logistic回归的方法得到相关危险因素的系数,并依此建立列线图模型。结果年龄、是否ICU患者、手术切口类型、糖尿病、脑梗死、入院检查耐药菌存在均为医院感染发生的独立预测指标(P<0.05)。对医院感染发生影响较大的因素包括入院时检查发现有耐药菌存在(OR=3.879,95%CI:2.014～5.985)、Ⅲ型手术切口(OR=3.105,95%CI:1.987～4.745)、糖尿病(OR=2.152,95%CI:1.221～3.598)。列线图模型构建成功,其预测效果的ROC曲线下面积为0.786(95%CI:0.432～0.894)。结论列线图模型可以应用在医院感染的预测上,应进一步优化模型,使其在临床得以推广性。
Objective The aim of this study was to explore the nomogram model that can be used to predict the occurrence of nosocomial infection. Methods A retrospective collection of clinical information was performed in hospitalized patients from 2015 to 2016. 1:2 matching method was adopted to obtain 1523 cases of nosocomial infections and 3046 cases in the control group. The coefficients of the related risk factors were calculated by logistic regression, then the nomogram model was established. Results Age, ICU patients, surgical incision type, diabetes, cerebral infarction, and admission examination of drug-resistant bacteria were all independent prognostic indicators of nosocomial infection(P<0.05). The factors with a great influence on nosocomial infection including the presence of drug-resistant bacteria were found at admission( OR=3.879, 95%CI: 2.014-5.985), type Ⅲ surgical incision(OR=3.105, 95%CI: 1.987-4.745), and diabetes(OR=2.152, 95%CI: 1.221-3.598).The nomogram model was established and the area under the ROC curve of the prediction effect was 0.786(95% CI: 0.432-0.894). Conclusion The nomogram model can be applied to the prediction of nosocomial infection, and the model should be further optimized so that it can be generalized in clinic.
Objective The aim of this study was to explore the nomogram model that can be used to predict the occurrence of nosocomial infection. Methods A retrospective collection of clinical information was performed in hospitalized patients from 2015 to 2016. 1:2 matching method was adopted to obtain 1523 cases of nosocomial infections and 3046 cases in the control group. The coefficients of the related risk factors were calculated by logistic regression, then the nomogram model was established. Results Age, ICU patients, surgical incision type, diabetes, cerebral infarction, and admission examination of drug-resistant bacteria were all independent prognostic indicators of nosocomial infection(P<0.05). The factors with a great influence on nosocomial infection including the presence of drug-resistant bacteria were found at admission( OR=3.879, 95%CI: 2.014-5.985), type Ⅲ surgical incision(OR=3.105, 95%CI: 1.987-4.745), and diabetes(OR=2.152, 95%CI: 1.221-3.598).The nomogram model was established and the area under the ROC curve of the prediction effect was 0.786(95% CI: 0.432-0.894). Conclusion The nomogram model can be applied to the prediction of nosocomial infection, and the model should be further optimized so that it can be generalized in clinic.
引文
[1]Wu X,Yan T,Liu Y,et al.Nosocomial infections among acute leukemia patients in China:An economic burden analysis[J].Ajic American Journal of Infection Control,2016,44(10):1123-1127.
[2]Sun B.Nosocomial infection in China:Management status and solutions[J].American Journal of Infection Control,2016,44(7):851-852.
[3]Brockman J A,Alanee S,Vickers A J,et al.Nomogram Predicting Prostate Cancer-specific Mortality for Men with Biochemical Recurrence After Radical Prostatectomy.[J].European Urology,2015,67(6):1160-1167.
[4]田晓文,梁小波,王振华,等.胃肠道间质瘤术后复发危险因素及列线图的应用价值[J].中华消化外科杂志,2017,16(1):71-76.
[5]崔方博,曹向明,李敏,等.非小细胞肺癌术后患者脑转移风险预测列线图的构建[J].医学研究生学报,2017,30(8):849-853.
[6]赵霞,王力红,张京利,等.老年患者术后医院感染的危险因素分析[J].中华医院感染学杂志,2017,27(8):1825-1828.
[7]Zuhal Yes,ilbag,Asl|Karadeniz,Ba?aran S,et al.Nosocomial Infections and Risk Factors in Intensive Care Unit of a University Hospital[J].Journal of Clinical&Experimental Investigations,2015,6(3):233-239.
[8]王蕾,鲍中英,苑晓冬.预防性抗生素应用对肝硬化并消化道出血患者医院感染的影响[J].标记免疫分析与临床,2015,22(7):617-619.
[9]李方龙,李德维,张一琰,等.预测前列腺穿刺阳性风险的列线图模型的建立[J].中华泌尿外科杂志,2016,37(8):616-620.
[10]Liang W,Zhang L,Jiang G,et al.Development and Validation of a Nomogram for Predicting Survival in Patients With Resected Non鈥揝mall-Cell Lung Cancer[J].2015.
[11]张春华,孙勇,马学先,等.手卫生成本与医院感染发病率的研究[J].中国感染控制杂志,2016,15(3):195-197.
[12]蒋红卫,张罗漫,孟虹.logistic回归模型共线性三种降维方法的模拟比较研究[J].中国卫生统计,2010,27(6):562-566.
[2]Sun B.Nosocomial infection in China:Management status and solutions[J].American Journal of Infection Control,2016,44(7):851-852.
[3]Brockman J A,Alanee S,Vickers A J,et al.Nomogram Predicting Prostate Cancer-specific Mortality for Men with Biochemical Recurrence After Radical Prostatectomy.[J].European Urology,2015,67(6):1160-1167.
[4]田晓文,梁小波,王振华,等.胃肠道间质瘤术后复发危险因素及列线图的应用价值[J].中华消化外科杂志,2017,16(1):71-76.
[5]崔方博,曹向明,李敏,等.非小细胞肺癌术后患者脑转移风险预测列线图的构建[J].医学研究生学报,2017,30(8):849-853.
[6]赵霞,王力红,张京利,等.老年患者术后医院感染的危险因素分析[J].中华医院感染学杂志,2017,27(8):1825-1828.
[7]Zuhal Yes,ilbag,Asl|Karadeniz,Ba?aran S,et al.Nosocomial Infections and Risk Factors in Intensive Care Unit of a University Hospital[J].Journal of Clinical&Experimental Investigations,2015,6(3):233-239.
[8]王蕾,鲍中英,苑晓冬.预防性抗生素应用对肝硬化并消化道出血患者医院感染的影响[J].标记免疫分析与临床,2015,22(7):617-619.
[9]李方龙,李德维,张一琰,等.预测前列腺穿刺阳性风险的列线图模型的建立[J].中华泌尿外科杂志,2016,37(8):616-620.
[10]Liang W,Zhang L,Jiang G,et al.Development and Validation of a Nomogram for Predicting Survival in Patients With Resected Non鈥揝mall-Cell Lung Cancer[J].2015.
[11]张春华,孙勇,马学先,等.手卫生成本与医院感染发病率的研究[J].中国感染控制杂志,2016,15(3):195-197.
[12]蒋红卫,张罗漫,孟虹.logistic回归模型共线性三种降维方法的模拟比较研究[J].中国卫生统计,2010,27(6):562-566.