[Male homicides in two regions of Brazil: analysis of the effect of age, period, and cohort].

The objective was to assess the effect of age, period, and birth cohort on the time trend in mortality from homicides in men in the states of the Northeast and Southeast regions of Brazil from 1980 to 2014. This was an ecological time trend study. PCA models were estimated using estimable functions in inference of the parameters. Mortality and population data were obtained from the Brazilian Health Informatics Department. State-by-state mortality rates from homicide were standardized by the direct method, with the world population as the standard, as proposed by the World Health Organization. The Northeast region recorded 317,711 deaths from homicides and the Southeast 544,640 deaths, corresponding, respectively, corresponding to mean standardized rates of 58.68 and 64.68 deaths per 100,000 men. The highest mean standardized mortality rates were observed in the states of Alagoas (157.74 deaths) and Pernambuco (109.58 deaths). All the states showed an increase in mortality up to the third decade of life, with a progressive reduction in the other age brackets. There was an upward trend in all the states of the Northeast and in Minas Gerais and Espírito Santo in the Southeast, while in the other states there was a downward trend. All the states showed an increase in the risk of death in the younger age brackets, except for the state of São Paulo, which showed the inverse profile. The current study's findings may correlate with the process of discontinuity in the cohort, in which members of wide cohorts found less opportunity for access to employment, income, and education, thus increasing the risk of involvement in crime and death from homicide.