BACKGROUND: In 819% of patients with atrial fibrillation (AF) with anticoagulant therapy (ACT), hemorrhagic complications occur, including due to excess doses of AC. At the same time, ACT is necessary for patients with AF, since anticoagulants effectively reduces the risk of ischemic stroke. To make a decision on the appointment of ACT, it is necessary to correlate the risks of ischemic stroke and bleeding, this requires knowledge of current clinical using ACT recommendations and instructions. Among patients admitted to hospital, 30% receive ACT, so increasing adherence to clinical recommendations for prescribing AC to patients with AF by doctors of various profiles is an urgent task. AIM: To analyze the adherence of physicians to recommendations for prescribing ACT before and after the introduction of decision support system (DSS) in patients with AF in a multi-specialty hospital. MATERIALS AND METHODS: A single-center non-randomized study with historical control to assess adherence to recommendations based on the analysis of medical prescriptions and the structure of drug errors in patients with AF in a multi-specialty hospital in Moscow before and after the introduction of DSS. Compliance with the recommendations of physicians was evaluated in the sections indications /contraindications to AC and dosage regimen of AC. The presence of deviations from the clinical recommendations /instructions for medical use of AC was regarded as management of the patient with non-compliance with recommendations. Physicians adherence level to recommendations was calculated as the ratio of cases of compliance with recommendations to the total number of cases. RESULTS: In the control and experimental groups, there was a significant increase in the proportion of POAC at discharge in comparison with admission to hospital: from 54.5 to 76.8% (p=0.0005) and from 63 to 85.7% (p=0.0002), respectively. However, only in the experimental group it was possible to significantly reduce the number of patients without a prescribed ACT (if there are indications) from 7.6 to 1% (p=0.04) in comparison with admission. During the study, it was possible to significantly increase physicians adherence level to the recommendations for the AC dosage regimen in patients with AF from 59% (44 discrepancies for 107 prescriptions) to 84.6% (16 discrepancies for 104 prescriptions); p0.005. Before the introduction of the DSS, the analysis of drug prescriptions revealed 56 drug errors (0.5 errors per patient), after the introduction of the DSS, the number of drug errors significantly decreased to 21 (0.2 errors per patient); p0.05. After the introduction of DSS, the number of sub-therapeutic doses of AC was reduced from 31 (27.7%) to 8 (7.6%); p0.05. CONCLUSION: The level of adherence to the recommendations for prescribing ACT to patients with AF in the hospital is high. The use of DSS increases the level of adherence to the recommendations on the AC dosage regimen in patients with AF, as well as eliminates errors in calculating the risk of ischemic stroke and systemic thromboembolic complications, and contributes to reducing the frequency of prescribing sub-therapeutic doses of AC.