Aim. To make a comparative analysis of the data available in the accounting medical documents drawn up at a multidisciplinary hospital on the level and structure of alcohol-related mortality (ARM) and to evaluate the efficiency of its accounting. Materials and methods. Accounting medical documents, such as 453 inpatient cards (Form 003/у), 453 postmortem protocols (cards) (Form 013/н-80), and 453 death certificates (Form 106/у-08), were chosen as the basis for the study. The data of the final clinical and postmortem diagnoses in the patients who had died at hospital and their primary cause of death were comparatively analyzed. Results. According to Form 003/у, ARM was 5.5%; the detection rate of alcohol-related disease (ARD) was 11% (95% confidence interval (CI), 8.3 to 14.3%); according to Form 013/н-80, ARM was 7.1% (95% CI, 4.9 to 9.8%) and the detection rate of ARD was 12.6% (95% CI, 9.7 to 16%). The consistency of the diagnoses of ARD as a main cause of death, made by hospital unit physicians and pathologists, is estimated as the mean - the Cohen's kappa coefficient (κ) is 0.570) (p<0.001). Conclusion. The results of the investigation suggest that there are 3 types of ARM, which differ in its level and structure: ARM in the assessments of hospital unit physicians; that in the assessments of pathologists, and that according to the death certificates drawn up. The consistency index for the diagnosis of ARD as a main cause of death indicates that the hospital unit physicians only determine the etiology of alcohol-related cause of death, without identifying it specifically.