Marcos Gonzaga , Universidade Federal do Rio Grande do Norte (UFRN)
Flávio Henrique M. de A. Freire, Universidade Federal do Rio Grande do Norte
Everton E. C. Lima, IFCH - UNICAMP
Bernardo L. Queiroz, Universidade Federal de Minas Gerais
José Monteiro, IPC-IG
The study of mortality level and trends in developed countries is limited by the quality of the vital registration system. The most common problems faced in these countries are incomplete coverage of vital registration systems and errors in age declaration for both population and death counts. he problem is more complex when it is studying small areas and sub-national populations. In recent years, a series of methodological alternatives were developed to provide mortality and life expectancy estimates for small areas. We aim to study the quality of mortality data for small areas in Brazil and estimate levels and structure of mortality combining demographic methods with Bayesian statistics. From these estimates, we produce a series of life tables for Brazilian municipalities from 1980 to 2010. We use data from the Ministry of Health Database – Mortality Information System – and population data (censuses) from 1980, 1991, 2000 and 2010. Data are available by age and sex for all municipalities of the country. The first step of the analysis uses traditional demographic to evaluate the quality of infant, child and adult mortality in Brazil. In the second step, we use the method suggested by De Beer and the variation proposed by Gonzaga and Schmertman to produce single age mortality estimates. Finally, we use a Bayesian model to estimate quality of data and life tables for all municipalities in the country.
Presented in Session 30. Life Tables and Applications