判别分析在巨细胞病毒性肺炎筛查中的应用  

Application of discriminant analysis in the screening of cytomegalovirus pneumonia

作　　者：曾昭成[1] 于长春[2] 孙志伟[1] 宋明媚[1] 陈小桥[1] 程丽萍[1] 

机构地区：[1]中国人民解放军第一○一医院儿科,无锡214044 [2]中国人民解放军第一○一医院信息科,无锡214044

出　　处：《国际儿科学杂志》2013年第1期104-106,共3页International Journal of Pediatrics

摘　　要：目的为巨细胞病毒性肺炎筛查寻找可靠且简便易行的判断准则，以减少巨细胞病毒性肺炎的漏诊和误治。方法应用判别分析方法对收集的56例巨细胞病毒性肺炎和42例普通病毒导致的喘息性支气管炎病例数据进行处理，建立诊断结果预测模型。结果经过变量筛选最后只有年龄、淋巴细胞计数和血小板计数三个指标进入模型，所建立的巨细胞病毒性肺炎筛查模型的灵敏度为80．36％，特异度为80．95％，误诊率为19．05％，漏诊率为19．64％，诊断符合率为80．61％。结论采用判别分析建立的巨细胞病毒性肺炎诊断预测模型简洁易懂，可操作性强，预测准确率较高，能为医务人员提供有力的筛查手段。Objective To look for a reliable and convenient judgement criteria for the screening of cy-tomegalovirus pneumonia in order to reduce misdiagnosis and resulted mistherapy. Methods Process collected data on fifty-six cytomegalovirus pneumonia and forty-two common viruses induced asthmatic bronchitis cases by use of discriminant analysis to construct prediction model of diagnosis result. Results Only three indexes in- cluding age,lymph count and platelet count were selected into the model via sift. The performance of the estab- lished screening model showed as follows： sensitivity was 80. 36%, specificity was 80.95%, misdiagnosis rate was 19. 05%, false negative rate was 19. 64%, diagnostic accordance rate was 80. 61%. Conclusion Being concise and of strong maneuverability and high accuracy in prediction, cytomegalovirus pneumonia diagnosis model constructed through discriminant analysis can provide powerful screening means for medical staff.

关 键 词：判别分析 巨细胞病毒性肺炎 疾病筛查 预测模型 

分 类 号：R72[医药卫生—儿科]