Effects of the high-inequality of income on the breast cancer mortality in Brazil

As well as breast cancer mortality, the income inequality in Brazil is different between Federative units, including between units of the same region. To assess the effects of the high-inequality of income on breast cancer mortality in Brazilian Federative Units, in the 2010 year. This is an ecologic study. Deaths from breast cancer in Brazilian women according to Federative units were obtained from the Department of Informatics of the Unified Health System. Mortality by breast cancer was estimated per 100,000 women and age-standardized by the direct method according to World Health Organization population. Income inequality was measured by the Gini index obtained from the United Nations Development Programme. The High-inequality of income was classified by the third tercile of the distribution from the Gini index of the Federative units. Univariate analysis was performed according to data normality. Linear regressions were performed by the stepwise backward method. The confidence level was 5%. Stata® (Stata Corp, LC) 11.0. was used. The High-inequality of income was associated with worse social and demographic indicators. The age-standardized breast cancer mortality was larger in the high-inequality of income Federative units. In the adjusted analysis, these Federative units presented a mean of 2 more deaths (ranging from the 0.4 to 3.7 deaths, r² = 0.79; p = 0.018) by breast cancer per 100,000 women when compared to the Federative units without high-inequality of income. In the Brazilian Federative units, the high-inequality of income was associated with age-standardized breast cancer mortality more.


Results
In Fig. 1 we present the spatial distribution of income inequality -assessed by the Gini index -, and breast cancer mortality in Brazil in the 2010 year. The Federative units with high-inequality of income (Gini ≥ 0.62) are also the ones with the worst socioeconomic and development indicators. This is reflected by the lower rate of aging (average difference of −1.5 (CI 95% −0.3; −2.8); p = 0.01), highest percentage of poverty (median difference of 18% more, ranging from 10.8 to 25.6%, p < 0.001), greater proportion of women under the age of 18 and who have children (average difference of 0.9, ranging from 0.3 to 1.6% more women per 100,000 women, p = 0.008) and lower Human Development Index for longevity (average difference of −0.03 (CI 95% −0.003; −0.05; p = 0.02)) and Human Development Index to education (average difference of −0.05 (CI 95% −0.1; −0.01); p = 0.02) found in the Federative units where high-inequality of income is present ( Table 2).
Based on these differences, we analyzed the impact of high-inequality of income on breast cancer mortality. In model 1, which analyzed the mean difference in age-standardized breast cancer mortality according to the high-inequality of income adjusted for income per capita, no statistically significant differences were observed (mean difference of −1.5 (CI 95% −3.8; 1.5); r² = 0.38; p = 0.001).
On the other hand, when we adjusted the age-standardized breast cancer mortality by aging index and HDI longevity (model 2) -variables present after exclusion in the statistical model -, we observed that Federative units with high-inequality of income show 2 more deaths (ranging to 0.4 to 3.7 death per 100,000 women, r² = 0.79; p < 0.001) by breast cancer when compared to Federative units with low/moderate income inequality (Fig. 2).

Discussion
Analyzing the differences in mortality rates for breast cancer among Federative units with high-inequality of income compared to Federative units with low/moderate inequality of income, we found that higher mortality due to breast cancer where there is a high-inequality of income.
However, the relationship between high-inequality of income and mortality due to breast cancer was observed after adjustment made by the Human Development Index and the aging index. The adjustment of the analyzes by confounding variables is one of the main steps necessary to understand the behavior of the variables before taking decisions to accept or reject hypotheses 12 and is know that the breast cancer mortality is associated with the aging  www.nature.com/scientificreports www.nature.com/scientificreports/ and with the Human Development Index 13 . As found in the present study, this change in the relationship between variables after adjustment was also observed in another study 14 about the income inequality.
Additionally, the HDI is related to the mortality due to breast cancer such as the improving of the health services quality and consequently, to increasing of the life expectancy 15 . Another important issue is that in middle-high income countries, the relative income is a main associated factor to the health.
In Brazil, the great part of the population lives with the elderly and was dependent on these elderlies and your income. Thus, the increase in longevity can be associated with the reduction of income inequality in Brazil 15,16 . In this country, there are public policies related to increasing of income such as the improvement of the minimum wage, adjustment in the security benefits laws and Brazilian social protection programs such as Bolsa família 17 to the population in general, but also policy directed to elderlies such as the retirement by age or by the fiscal contribution 16 .
The rapid demographic transition between 2001 and 2011 has increased income inequality, especially among the poorest 18 , such as inequality between regions 19 . On the other hand, the investment in health 20 and the increase of the minimum wage 21 were factors related to the decrease of the income inequality in the Federative units and federal district and that are related to better conditions of access to health in this country.
With high-inequality of income in Brazil, there are Federative units with lower economic power and few in the opposite situation. This can make it increase of cost of living of inhabitants and the need and of the state provide resources to the health. However, the resources and healthcare are unequal among Brazilian administrative regions, mainly for the breast cancer diagnosis and delay to the confirmation 22,23 . One of the main factors related to the increase in mortality is the late diagnosis of the cases, which makes treatment difficult and increases the lethality of the cases 23 .
Living in less developed regions and low educational level are factors that increase the risk of late diagnosis of breast cancer in Brazilian women 24 , a fact that also occurs in different regions of Europe 25 and Hong Kong 26 . With less access to health services, lower education and unhealthy living habits, people are more likely to develop chronic diseases such as breast cancer and to have the diagnosis of these chronic diseases when they are already in higher stages. When those people living in areas of high inequality are diagnosed with breast cancer, they are already at higher stages 13 , and associated with delays in the provision of health services 22 , the lethality of the cases is higher and shorter survival 23 .
In the present study, we found that high-income inequality is related to worse socioeconomic and developmental indicators. This is because the increase in income inequality is directly related to social determinants that influence health such as smoking habits, alcoholism, and low educational levels 1,27,28 . The increase in these indicators points to a scenario of less self-care, as well as lower access to the primary health service 29 . With the greatest income inequality, there is also exacerbation of the social determinants that are related to health, mainly to the life habits and health care, who are consequences of the social stressor and lack of resources related to the income inequality. For these reasons, it is common to find studies that report an association between income inequality and outcomes related to social behavior 30 , suicide rate 4 or breast cancer 6,8 .
In addition, in the last decade, increased investment in strategies for early diagnosis and treatment of breast cancer through public policies for decentralization of the single health system 8 and increased life expectancy 17 are also factors related to mortality and the reduction of income inequality, respectively. In a scenario of changes related to income inequality, it is important to understand that the change in income inequality may influence in a different way the current income of the populations 31 , and in this sense, this is a field of knowledge that still has much to be discovered. In addition, understanding how high inequality of income is related to other health outcomes can be an important tool to understand the real impact of this characteristic on the health of populations.
We do not understand the results found as if reducing income inequality would save lives. On the other hand, we find here that the places where there is high inequality of income are susceptible to worse socioeconomic conditions and consequently, lower conditions to have adequate health, which reflected in higher mortality due to breast cancer in the Brazilian Federative units with this characteristic in the 2010 year.
When we analyze the effects of high-inequality of income on breast cancer mortality in Brazilian Federative units in 2010, we found that there is higher mortality due to breast cancer in the Federative units, where there www.nature.com/scientificreports www.nature.com/scientificreports/ is a high-inequality of income. In this sense, it is necessary that there be adequate public policies for breast cancer for each reality at the state level, given the existing income inequalities, as well to understand the impact of high-inequality of income on the other health outcomes.

Methods
Study design. This is an ecologic study. Data sources. They were used as data sources used for other epidemiological studies, whose reliability and validity has already been described 6,32  Thus, were considered as high-inequality of income the Federative units with Gini index higher than 0.62 and as low/moderate inequality of income states with values less than or equal to 0.62.
Mortality from breast cancer -outcome variable: The deaths owing to breast cancer were obtained from the Department of Informatics of the Unified Health System (DATASUS) through the Mortality Information System (Sistema de Informações sobre Mortalidade, SIM), defined according to the Tenth International Classification of Diseases (ICD-10) 33 .
The population of women was obtained from the Brazilian Institute of Geography and Statistics (IBGE) using data from the 2010 Demographic Census. Crude mortality was calculated per 100,000 women. Crude rates were standardized by age using the direct method, using the standard population of the World Health Organization (WHO) 34 .
Sociodemographic characteristics -adjustment variables: The sociodemographic variables of the population according to Brazilian Federative units were obtained from the IBGE states, UNDP and DATASUS. The variables were presented in Table 1.
Statistical analysis. The Shapiro-Wilk test was used to assess the distribution of quantitative variables. For the variables with a normal distribution (Shapiro-Wilk, p ≥ 0.05) the T-test was used and for the variables without normal distribution (Shapiro-Wilk, p < 0.05), was used Mann-Whitney test.
Linear regression was used to analyze the association between high-inequality of income and mortality due to breast cancer. Two models were tested: I) high-inequality of income adjusted for per capita income and II) high-inequality of income adjusted by the other variables in the model after application of the stepwise backward strategy, with input selection criteria of 0.05 and withdrawal of the model of 0.10.
The significance level was 5%. The program used for the statistical analysis was Stata 11.0 ® (Stata Corp, L C).
Tabwin version 3.0 was used to create the maps.