Objectives. To compare the prognostic value of doctors’ and patients’ assessments of the severity of depression with consideration of treatment outcomes. Materials and methods. The study was organized as a muticenter investigation of 107 patients with depression. Assessments of mental status were preformed using the HAM-D, SHAPS, CGI-S, CGI-I, PGI-S, PGI-I, and VAS. Results. Differences were found between psychiatrists and patients in their assessments of the severity of depression. In addition, doctors’ assessments showed contradictions between the results on the HAM-D and CGI-S scales. Assessments of changes by patients and doctors became more similar as the severity of depression decreased and patients’ wellbeing improved on the background of treatment with agomelatine (Valdoxan). Conclusions. Agomelatine (Valdoxan) was effective and well tolerated in the treatment of depression of any degree of severity. The contradictions in assessment of the initial severity of depression by doctors and patients using different psychometric scales may eliminate the prognostic value of these assessments in relation to treatment outcomes. © 2018, Springer Science+Business Media, LLC, part of Springer Nature.