[Classification approach to prognostication of the severity of bronchial asthma]

To define the severity of bronchial asthma, to predict its subsequent course, and to prescribe adequate glucocorticosteroid therapy, a mathematical model based on the cluster analysis was used. The authors' algorithm of the classification of bronchial asthma was based on the cluster analysis was used. The authors' algorithm of the classification of bronchial asthma was based on the mean distance value between the objects characterizing patients' status and hyperspatial signs. Each patient was described using a set of 15 clinical, functional, and immunological indices which had a meaning of signs describing the objects. Three clusters corresponding to the groups identified a priori up to two patients were formed to make the classification by the algorithm proposed. The findings make it possible to efficiently analyze the baseline indices in a particular patient from his entrance into or highest relationship to a definite cluster. This may define the severity of bronchial obstruction and predict its progression.