OBJECTIVES : In the biggest diabetes epidemiological study NATION in Russia (2016), included 26620 adults, the prevalence of T2DM (HbA1c ≥6.5%) was found as 5,4%. The objective of this analysis is to develop a tool to access the probability of having DM2T for the Russian population using the NATION database.

METHODS : A factor analysis of the pooled NATION database revealed 19 potentially relevant factors correlated with diabetes. Different combinations of factors were tested and used to train the logistic regression to identify optimal parameters. Grouping thresholds were set so as to maximize the area under the ROC-curve based on 10 main parameters. Based on the obtained data, a scoring system was constructed to establish the risk of diabetes depending on the values of the logistic regression model results.

RESULTS : The following variables strongly correlated with diabetes: age, body mass index, waist circumference, at least one reading of elevated blood sugar (fasting), high systolic blood pressure, low physical exercise / activity, low consumption of fruits and vegetables (less than 3 times per week), visit to endocrinologist within the last 6 months, waist-to-hip ratio, alcohol consumption (more than 4 drinks within the last month), fast-food and sweet beverages consumption, hypertension, use of lipid-lowering drugs, certain comorbidities, DM2T in close relatives. A special scale for screening of DM2T was constructed (sensitivity is 81%, specificity – 75%).

CONCLUSIONS : The diabetes risk calculated designed based on NATION data can be noninvasive, fast and reliable tool for screening of Russian population for T2DM.