CAJ | 학술논문

目的：探讨Logistic回归在巨细胞病毒性肺炎临床筛查中应用的可行性。方法：采用单变量分析筛选出巨细胞病毒性肺炎组与普通喘息性支气管（肺）炎组差异有统计学意义的指标，然后利用Logistic回归建立预测模型，最后对样本进行回代，以评估模型预测的准确性。结果：两组差异有统计学意义的指标有年龄、血红蛋白（HB）、白细胞（WBC）、淋巴细胞数（L）、血小板（Plt）、IgG、IgM ，进入最终 Logistic预测模型的有血红蛋白（HB）、淋巴细胞数（L）、血小板（Plt），预测模型的灵敏度、特异度和诊断符合率分别为83．08％、76．09％、80．18％。结论：运用Logistic回归预测模型筛查巨细胞病毒性肺炎是可行的，可以为临床医务人员提供辅助诊断依据。
목적：탐토Logistic회귀재거세포병독성폐염림상사사중응용적가행성。방법：채용단변량분석사선출거세포병독성폐염조여보통천식성지기관（폐）염조차이유통계학의의적지표，연후이용Logistic회귀건립예측모형，최후대양본진행회대，이평고모형예측적준학성。결과：량조차이유통계학의의적지표유년령、혈홍단백（HB）、백세포（WBC）、림파세포수（L）、혈소판（Plt）、IgG、IgM ，진입최종 Logistic예측모형적유혈홍단백（HB）、림파세포수（L）、혈소판（Plt），예측모형적령민도、특이도화진단부합솔분별위83．08％、76．09％、80．18％。결론：운용Logistic회귀예측모형사사거세포병독성폐염시가행적，가이위림상의무인원제공보조진단의거。
Objective:To discuss the feasibility of applying Logistic regression to clinical screening of cy-tomegalovirus pneumonia .Methods :Sift indicators statistically significant between cytomegalovirus pneumo-nia and common panting bronchitis by univariate analysis and then build prediction model through Logistic re-gression ,finally assess the accuracy of the model by means of back substitution .Results:Indicators statisti-cally significant include :age ,HB ,WBC ,Lymph ,Plt ,IgG ,IgM ,among w hich HB ,Lymph and Plt enter the fi-nal Logistic prediction model .T he prediction sensitivity、specificity and diagnostic accordance for the sample go as follows respectively :83 .08% 、76 .09% 、80 .18%. Conclusion:It is feasible to screen cytomegalovirus pneumonia by application of logistic regression prediction model ,w hich can offer aided diagnostic basis for clinical staff .