骨髓增生异常综合征并发医院感染危险因素及其预测模型评价
Risk factors for hospital-acquired infections in patients with myelodysplastic syndrome and its predictive model evaluation
王舒 1王娟 2刘冰 1陈宛丽 1杨莹1
作者信息
- 1. 南阳市第二人民医院血液淋巴瘤科,河南南阳 473000
- 2. 河南省南阳市内乡县人民医院肿瘤科,河南南阳 474350
- 折叠
摘要
目的 探讨骨髓增生异常综合征(MDS)并发医院感染危险因素及其预测模型评价.方法 收集2021年4月-2023年2月南阳市第二人民医院收治的548例MDS患者为研究对象,根据患者医院感染情况分为感染组(102例)和未感染组(446例);分析MDS患者医院感染现况、临床资料,通过Logistic回归分析法归纳MDS患者医院感染的危险因素,构建列线图预测模型、Logistic回归模型,并绘制受试者工作特征(ROC)曲线分析其对MDS并发医院感染的预测价值.结果 侵入性操作、抗菌药物使用时间、染色体核型是MDS患者医院感染的危险因素(P＜0.05),骨髓原始细胞和中性粒细胞绝对计数(ANC)是MDS患者医院感染的保护因素(P＜0.05);将上述模型筛选出的5个因素纳入模型,构建MDS患者医院感染发生的列线图预测模型,经Hosmer-Lemeshow拟合度检验结果显示,x2=4.453,P=0.616,模型拟合的准确度较好;将上述因素纳入多因素Logistic预测模型:logit(P)=-0.864+侵入性操作×0.678+抗菌药物使用时间×0.735-骨髓原始细胞×0.826-ANC×1.129+染色体核型×2.222;绘制ROC曲线,结果显示,当logit(P)＞0.176时,曲线下面积(AUC)为0.911,诊断敏感性为83.33％、特异性为85.65％.结论 MDS患者医院感染与多种因素密切相关,构建预测模型预测MDS患者医院感染发生的价值较高,可早期筛查高危患者,并进行预防性干预.
Abstract
OBJECTIVE To investigate the risk factors for hospital-acquired infections in patients with myelodys-plastic syndrome(MDS)and its predictive model evaluation.METHODS Totally 548 patients with MDS admitted to Nanyang Second People's Hospital from Apr.2021 to Feb.2023 were selected as the research subjects,and were divided into the infected group(102 cases)and the non-infected group(446 cases)according to the occurrence of hospital-acquired infection.The status of hospital-acquired infection and clinical data in patients with MDS was an-alyzed,the risk factors of hospital-acquired infection in patients with MDS were analyzed by logistic regression a-nalysis,and the nomogram prediction model and logistic regression model were constructed,and receiver operat-ing characteristic curve(ROC)was plotted to analyze the predictive value of the model in MDS complicated with hospital-acquired infection.RESULTS Invasive procedures,duration of antibiotic use,and chromosomal karyotype were the risk factors for hospital-acquired infections in MDS patients(P＜0.05),while absolute counts of bone marrow progenitor cells and neutrophils(ANC)were protective factors for hospital-acquired infections in MDS pa-tients(P＜0.05);The 5 factors screened by the above model were included in the model to construct a nomogram model for the occurrence of hospital-acquired infections in MDS patients,and the Hosmer Lemeshow fit test showed that x2=4.453,P=0.616,and the accuracy of model fitting was good.The above factors were included in the multivariate logistic prediction model:logit(P)=-0.864+invasive operations × 0.678+antibacterial drug usage time × 0.735-Bone marrow primitive cells × 0.826-ANC × 1.129+chromosomal karyotype× 2.222.The ROC curve was plotted,and the results showed that when logit(P)＞0.176,the area under the curve(AUC)was 0.911,the diagnostic sensitivity was 83.33％,and the specificity was 85.65％.CONCLUSION The occurrence of hospital-acquired infection in patients with MDS was 18.61％was closely related to a variety of factors,and the value of the predictive model to predict hospital-acquired infection in patients with MDS was high,which allowed for early screening of high-risk patients and preventive interventions.
关键词
骨髓增生异常综合征/医院感染/现况/危险因素/模型/预测价值Key words
Myelodysplastic syndrome/Hospital-acquired infection/Status/Risk factor/Model/Predictive value引用本文复制引用
出版年
2024
NETL
NSTL
万方数据