MINERAÇÃO DE DADOS DE AUTOPSIA PARA DETERMINAS AS CAUSAS DE MORTE NA DEPRESSÃO

Abstract

Depression is associated with increased mortality, but the causes of death related to depression, as well as the life expectancy of patients, are still poorly explored and controversial. Identifying possible diseases associated with death in patients with depression can help in public health policy decision making and culminate in more specific treatments, prevention strategies, and improved life expectancy for these patients. In studies on causes of death which autopsy examinations are analyzed, it is possible to acquire more accurate information about the diseases related to death, since the autopsy determines the precise cause of death. In this study, we evaluated the causes of death of 1,136 subjects, according to autopsy reports from the Death Verification Service of the Capital (SVOC-USP) in the metropolitan region of São Paulo. The diagnosis of depression of these subjects was made according to the Structured Clinical Interview for DSM-IV (SCID). Data mining based on the ICDs of causes related to death was performed, in which eleven Machine Learning algorithms were applied in order to search for patterns to determine the possible causes of death related to depression. In addition to major depression, eight other subgroups of depression were analyzed. Although this study perfomed a broad investigation in the general population and in specific groups of patients, the results obtained by the generated models do not indicate differences in patterns in the causes of death in individuals with and without depression. This result corroborates with previous studies in the literature where the evidences for all-cause and cause-specific causes of death and depression are not significant.