The main purpose of this article is to present a Differential Item Functioning method of item analysis. It is designed to minimize the discriminatory effect of individual items in the accreditation of graduates of medical universities with different training programs. The one-parameter Item Response Theory model is used to align graduates' rights in accreditation. The measure of the difference in the location of the item characteristic curves constructed for different graduates' samples is presented using the difficulty estimates. For the development of the methodology, a number of research questions are posed, the solution of which made it possible to analyze the items bias for the "Polyclinic" discipline and the range of item difficulties from 1.5 to 2, 5 logits. The item bank was cleaned and an interpretation of decisions on items release and correction for the cases of different arrangement of their characteristic curves was presented.