[Experience in using regression analysis for assessing the quality of the diagnosis of hemorrhagic fever with renal syndrome].

A new approach has been developed to evaluate the adequacy of HFRS clinical diagnosis, by using regression analysis of large selections. The HFRS morbidity rate and the antibody prevalence rate in man were viewed as a function and an argument of the function, respectively. An empiric model for one of the active HFRS endemic area (Saratov Province), characterizing the correlation between the immunity and morbidity rates has been elaborated. It calculates the theoretical morbidity level that corresponds to determine an immune portion of the population in each administrative area. The theoretical (counted) values were compared to the annually registered morbidity rates; their similarity has been estimated. Estimation of the adequacy of HFRS diagnosis in the examined region could have been admitted as satisfactory. However, various faults in the clinical identification of HFRS cases were found to occur in 14 of 44 districts of this region. Hyperdiagnosis was made in the most active HFRS natural foci (broad-leaved forests), but in the endemic areas being characterizes by moderate or low activity (forest-steppe and steppe districts). The results of the authors' calculations were shown to be generally in agreement with the data from the Regional Epidemiological Service reports which confirmed the adequacy of the proposed material.