Padrão da mortalidade brasileira: estimativas a partir do nível municipal

Abstract

Know the level and pattern of mortality is important to understand population dynamics as well as to plan public policies for the health and welfare of the population. One of the ways to get information about mortality is the Mortality Information System (SIM). Despite showing significant advances in recent years, the quality of the SIM, still prevents the direct use of data to generate indicators of mortality, which creates the need to use methods of correction for underreporting. The methods, however, are mostly not suitable for correcting the data on deaths of children and young people, and neither of sparsely populated areas. Given this need, this paper proposes to estimate the level and pattern of mortality in Brazilian municipalities in 2010, through two approaches to correct for underreporting. The first, which is applied to young and adult mortality, combines the adjusted method of extinct generations, proposed by Hill, You, and Choi (2009) with the empirical Bayes estimator proposed by James Stein (Marshall, 1991). The second methodological approach, the multiple regression analysis, in order to estimate the odds of child mortality (under 14 years) by municipalities, with bases in adult already estimated probabilities is used. In search of estimates of the level and pattern of mortality in Brazilian municipalities, these methodologies were applied to all municipalities for the year 2010, by age group and sex. The spatial pattern found for the quality of reported deaths, shows that the South and Southeast regions have mortality data of better quality in the country, while the North and Northeast have the lowest coverage of the death records. Regardless of gender, the odds of higher juvenile death occur in the municipalities of North and Northeast regions, while the probability of young adult death (15-60 years) the municipalities are the most likely of South and Southeast . The methodologies used in this work suggest a division of Brazilian male mortality in six regions with different patterns of mortality for women has suggested four patterns.