There were studied the possibilities of predicting the effectiveness of treatment with roflumilast in the complex therapy of patients with chronic obstructive pulmonary disease (COPD) with low and high risk by assessing the clinical and functional parameters (lung function, severity of respiratory complaints, frequency of exacerbations in the previous year) and serum indicators of systemic inflammation (C-reactive protein, TNF-α, IL-6, IL-8, fibrinogen). One-year study included 60 patients with COPD, among whom there were 22 patients with low risk (group 1) and 38 patients with high risk (group 2), according to the multi-faceted classification of disease. Each group was divided into 2 subgroups depending on the response to treatment: subgroup A included the patients whose therapy was effective and subgroup B consisted of the patients whose therapy was not effective. The criteria for the effectiveness of the treatment were: group 1 – the absence of exacerbations during a year or 1 episode that did not require hospitalization, the initial test CAT (COPD Assessment Test) – 10 points and lower, dyspnea on the mMRC scale no more than 2 points, an increase in FEV1 by 11% or more; in patients of group 2 the number of exacerbations was 0 or no more than 2 and they did not require hospitalization, dyspnea on the mMRC scale was no more than 2 points, there was no decrease in FEV1. The effectiveness of treatment was evaluated in points (4 points meant effective treatment, below 4 points – not effective treatment). It was found out that in group 1 in subgroup A after treatment the results of CAT (p<0.001), IL-8 (p<0.001), TNF-α (p<0.001) were significantly lower than in group B (p<0.001). In group 2 after treatment in subgroup A test CAT (p<0.001), CRP (p<0.001) and IL-6 (p<0.01) were significantly lower than in subgroup B. The revealed regularities formed the basis for the creation of mathematical model for predicting the effective application of roflumilast by discriminant analysis for different phenotypes of the disease which allow clinicians to solve the problem of a personalized approach to the selection of patients with COPD with low and high risk.