EFFECTS OF INEQUALITY IN THE INCOME DISTRIBUTION ON THE INFANT MORTALITY INEQUALITIES IN BRAZIL

Tuesday, 19 August 2014: 11:15 AM
Ballroom C (Dena'ina Center)
Celia L Szwarcwald, PhD , Fundação Oswaldo Cruz, Rio de Janeiro, Brazil
Paulo R Souza Jr, PhD , Fundação Oswaldo Cruz, Rio de Janeiro, Brazil
Giseli N Damacena, PhD , Fundação Oswaldo Cruz, Rio de Janeiro, Brazil

INTRODUCTION: In the last 20 years, Brazil has undergone several changes in terms of socioeconomic development and health care. In the years 2000,  conditional cash transfer programs to the poorest resulted in changes in the income distribution. In this paper, we analyze the effects of these changes on infant mortality rate (IMR) inequalities among Brazilian municipalities from 2000 to 2010.

METHODS: Firstly, a log-log regression model was fitted to the 2000 IMR municipal data to express the relationship between IMR and income. To analyze the effects of changes in income distribution from 2000 to 2010 on traditional measures of health inequalities, we used the same log-log model for year 2010 and calculated the expected municipal rates. In the second part of the analysis, we used the same rationale, but two additional variables were used to characterize socio-demographic changes: proportion of mothers with incomplete fundamental education and the fertility rate.

RESULTS: In 2000, the IMR decreased from 47.9 to 17.3 per 1000 LB from the first to the last income decimal while the expected 2010 mean rates varied from 34.6 to 11.9 per 1000 LB, giving the false impression of a much lower degree of IMR inequality, although the relation between income and IMR remained the same. For the multivariate model, results are even more pronounced. All the inequality measures are much lower when the 2010 observed IMR values are compared to the expected ones, after controlling for income, education, and fertility rate.

CONCLUSIONS: The results evidenced that the traditional summary measures of health inequalities are affected by changes in income distribution. Therefore, if the social gradient is to be interpreted meaningfully in a health policy context, the narrowing of income-related health inequalities should be examined after removing the effects of changes in the inequality in the income distribution over time.