ROC-analysis in determination of diagnostic importance of ventilation parameters in bronchial asthma patients

Abstract

ROC-analysis in determination of diagnostic importance of ventilation parameters in bronchial asthma patientsThe aim of this study was the ROC-analysis of diagnostic importance of the lung ventilation parameters: FVC, FEV₁, FEV_i/FVC, flow parameters PEF, FEF_25,50,75, FEF_25-75. The probabilistic approach was considered at this study as an alternative way against a widespread method of comparison of means to evaluate the diagnostic importance of the lung function parameters.
We examined 284 bronchial asthma (BA) patients and 101 persons as a comparative group. The BA patients were divided into 4 groups according to severity of their ventilation disorders (the dividing criterion was FEV₁% .
To assess informative values of the ventilation parameters in the BA patients we plotted ROC-curves, calculated their areas; we also used a "detectability" index. Then a correlative analysis was performed between this index and areas of the ROC-curves. There was a close positive correlation between these values. This fact allowed to recommend this method as an express-method to asses the diagnostic importance of the parameters instead of a more laborious method of the ROC-curves. According to the results of the ROC-analysis and index-based assessment the diagnostic values of the ventilation parameters in the BA patients were decreasingly graded: FEV₁, FEV₁/FVC, FEF₂₅, FEF₅₀, FEF_25-75, PEF, FVC, FEF₇₅.
The our data showed that the ROC-analysis of the diagnostic importance of the ventilation parameters in BA patients gives fuller and more demonstrative conception of diagnostic abilities of the parameters when compared with the method of comparison of means. This fact improves a reliability of diagnosis of the obstructive disorders.