摘要
Abstract
Objective To establish a Logistic regression model for investigating the diagnostic values in the de‐tection of C‐reactive protein (CRP) and serum amyloid A (SAA) and procalcitonin (PCT) in the identification of in‐fectious diseases .Methods 230 cases of virus infections ,245 cases of bacterial infections and 200 cases of healthy controls were enrolled in the study .The single factor and multiple factors analysis method were adopted to conduct the comparative analysis on the laboratory tests results ,the Logistic regression model of infectious disease diagnosis method was established and ROC curve was drawn according to the model .Results The differences of CRP ,SAA , PCT ,WBC and SAA/CRP ratio had statistical significance between the infection group and the control group (P<0 .05) .The levels of CRP ,SAA ,PCT ,WBC and NEU% in the bacterial infection group were obviously higher than those in the virus infection group and the control group ,but serum LY% and SAA/CRP ratio in the virus infection group were significantly higher than those in the bacterial infection group (P<0 .01) .The four indicators in the bac‐terial infection group were screened into the Logistic regression model:CRP ,SAA ,PCT and WBC ,in which the areas under ROC curve of CRP ,SAA and PCT were more than 0 .9 ;only SAA and WBC in the virus group were entered in‐to the Logistic regression models ,in which the area under ROC curve of SAA was 0 .859 .Conclusion The Logistic regression model based on the experimental diagnostic technology is helpful for clinical early differential diagnosis of infectious disease ,especially the joint detection of SAA and CRP or PCT is more meaningful .关键词
感染性疾病/C反应蛋白/血清淀粉样蛋白A/降钙素原/Logistic回归Key words
infectious diseases/C-reactive protein/serum amyloid protein A/procalcitonin/Logistic regression
