ESTIMATIVAS DE MORTALIDADE INFANTO-JUVENIL PARA AS MESORREGIÕES DO BRASIL PARA O DECÊNIO 2000/2010
Mortality infant-juvenile, Mesoregions, Linear Regression Models, Underreporting, Model Log-Quadratic
The sharp decline in mortality rates and improvement in the quality of vital records in Brazil were processes widespread in recent decades, but with significant regional contrasts. Estimates of specific mortality rates in some regions of the country still represent a challenge for demographers. Among the limitations, highlight the high variability in the rates by age and incomplete coverage of vital records. The demographic methods for evaluation and correction of underreporting of deaths at early ages have limited applicability in sub-populations that experienced a rapid and intense destabilization of its age structure, especially in more disaggregated geographical levels. Thus, the objective of this study is to propose indirect estimates for mortality infant-juvenile, 0-14 years, based on adult mortality estimated for the mesoregions of Brazil in the decade 2000/2010. The study data sources are: Human Mortality Database (HMD), Census 2010, Mortality Information System (SIM) and adult mortality estimates for mesoregions of Brazil, estimated by the research project "Mortality Estimates and Construction Tables Life for Small Areas in Brazil, 1980-2010 "(470866/2014 MCTI / CNPQ / MEC / CAPES / 454,223 and / 2014-5: MCTI / CNPQ). Two methodological approaches were employed, both using regression models which presupposes the existence of a strong relationship of infant mortality with adult mortality. Both proposals show mortality patterns with higher levels of mortality infant-juvenile in mesoregions the North and Northeast. The results show that mortality rates estimated by Proposal 2 show, in general, greater heterogeneity among mesoregions the same state, especially among men. Overall, the results point to lower female mortality levels with more homogeneous patterns compared to male mortality. Among the methods employed, the proposed 2 is shows promise for indirect estimates of mortality infant-juvenile in areas with lower quality of vital records.