Racial inequities and aging : analysis of the 2010 cohort of the Health , Welfare and Aging Study ( SABE )

REV BRAS EPIDEMIOL 2018; 21(SUPPL 2): E180004.SUPL.2 RESUMO: Introdução: Entender as disparidades raciais no Brasil tem sido algo bastante complexo e pouco investigado em alguns segmentos populacionais, como na população idosa. Objetivo: Objetivou-se apresentar de forma descritiva uma análise comparativa, numa perspectiva racial, do perfil sociodemográfico, das condições de saúde e de uso de serviços de saúde dos idosos da cidade de São Paulo, SP. Métodos: Trata-se de uma análise transversal do Estudo Saúde, Bem-Estar e Envelhecimento (SABE). Para o presente trabalho, foram considerados 1.345 idosos da coorte de 2010. Selecionaram-se os dados referentes aos idosos de cor preta, parda e branca. Abordaram-se os dados em três eixos essenciais: sociodemográficos, condições de saúde e uso e acesso a serviços de saúde. A medida epidemiológica de associação escolhida foi a razão de prevalência (RP), para expressar as diferenças entre os grupos. Resultados e conclusão: Os resultados evidenciaram um cenário mais favorável para o envelhecimento dos idosos de cor branca em comparação com aqueles de cor parda ou preta, no tocante aos indicadores sociodemográficos, às condições de saúde ou de uso e ao acesso a serviços de saúde.

carried out since 2000, collecting data about several aspects of the lives of the non-institutionalized elders living in the city of São Paulo 9,10 .Data collection, conducted by trained interviewers, took place in three moments: 2000, 2006, and 2010, with the reposition, in the last two cohorts, of elders aged from 60 to 64 years old.For this study, 1,345 elders were considered in the 2010 cohort.
The dependent variable of the study was "self-reported skin color" 11 , and this analysis used data referring to black, brown, and white elders, according to the following variable: • sociodemographic indicators: sex (male and female), age (60-64 years old, 65-69 years old, 70-74 years old, 75-79 years old, 80 years old or more), marital status (divorced/ widower, married, single), number of living children (none, one or two children, three or more children), sufficient income for daily expenses (yes, no), knows how to read and write a message (yes, no), schooling (in years); • way of life: lives alone (yes, no), religion (catholic, evangelical, other/none), leaves home (yes, no), invites people over (yes, no), is currently working (yes, no), reasons to work (finances, keep an occupation), and having lived in the rural area for at least five years until the age of 15 (yes, no); • health conditions: health self-assessment (very good/good, fair, bad/very bad), selfreport of diseases -osteoporosis (yes, no), arterial hypertension (yes, no), diabetes (yes, no), stroke (yes, no), cognitive decline (yes, no), psychiatric or nervous problem (yes, no), and depression (yes, no).The geriatric depression scale was used to evaluate depressive symptomatology 12 , and for the mini mental state examination 13 ; • health behaviors: smoking history (currently smoking, has smoked, never smoked), risk for alcohol dependence -Michigan Alcoholism Screening Test, Geriatric Version (S-MAST-G) 14 -(yes, no) and conditions to care for their own health (usually, once in a while/rarely, never); • use and access to health services: having an additional health insurance plan (yes, no), amount paid by the health insurance (in Reais (BRL): USD 1 = BRL 3,39), difficulty to use health services (yes, no), having complementary examinations (yes, had them all; no, but they are all scheduled; no, just some), need for urgency care (yes, no), place of emergency care (Unified Health System -SUS, health insurance plan, private).
For the statistical data analysis, bivariate analyses were conducted between the dependent variable "race/skin color" and the independent variables through contingency tables, always considering the values proportional to each racial category.Since this is a complex sample, the Rao-Scott χ 2 test was used for the bivariate analyses of the qualitative variables, and the calculation of the mean and its respective 95% confidence interval (95%CI) was used when the independent variable was quantitative.
To express differences between groups, the epidemiological measurement of the association chosen was the prevalence ratio (PR), together with percentage values.For all of the analyses, the coefficients of sample weight were used, as well as the 5% significance level (a = 0.05).The statistical software used was Stata (version 11.0).

RESULTS
Of the total 1,345 elders in the 2010 cohort, 1,263 were considered for this analysis.Yellows and indigenous people were excluded.The following was analyzed: 782 (62%) whites, 387 (30.9%) browns, and 94 (7.1%) blacks, who, after the weighting procedures, began to represent the population of 1,244,371 elders, out of whom 771,510 are white, 384,511 are brown, and 88,350 are black.

SOCIODEMOGRAPHIC CHARACTERISTICS
According to Table 1, the fact that the highest proportion of browns is concentrated in the age group between 60 and 60 years old (p < 0.001) stands out.About marital status, there were more single people among blacks (p = 0.050).Not having living children in 2010 was more common for black older people (10.70%; p < 0.001), while whites were those who reported having one or two living children (40.80%; p < 0.001).However, when the elders with three or more children were analyzed, a higher proportion was observed among browns and blacks (p < 0.001).
In socioeconomic aspects, there was a concentration of browns and blacks (p < 0.001) among those who reported not having enough income for daily expenses, which dialogs with schooling data, because these are the same groups that had the highest proportions of illiterate people (p < 0.001) and the lowest mean of schooling years (p < 0.05).The mean schooling period for blacks (3.90 years; 3.01 -4.79 95%CI) and browns (4.09 years; 3.52 -4.66 95%CI) was much lower if compared to the white elders (6.10 years; 5.49 -6.72 95%CI).The highest proportion of those who did not know how to read or write a message (23.20%; p < 0.001) was among the black elders.

WAY OF LIFE
According to Table 2, it was possible to observe the prevalence of evangelicals among browns and blacks, and Catholics (p < 0.001) among whites.Concerning the social life of the elders, expressive proportions of black people said, respectively, they never go out to public places (64.80%; p = 0.005) and never invite people over (42.10%; p = 0.018).
The black (38.60%) and brown elders (38.30%) were those who mentioned working at the time of the survey (p = 0.01), even though the reason -the financial need -is the same in the three groups (p = 0.63).Regarding the previous life in a rural area, the brown elders were those who more often claimed to have lived in such areas for at least five years until they were 15 years old (p = 0.02).

HEALTH CONDITIONS: SELF-REPORT AND SELF-EVALUATION OF HEALTH
As in Table 3, most elders assessed their own health as very good or good (50.30%, in average), and this proportion decreases considering the black elderly (41.80%), even though without statistical significance (p = 0.38).Osteoporosis was one of the chronic diseases with   to whites (7.10%; p < 0.001).It is worth to mention, however, that self-reported psychiatric problems and depression, respectively, were less prevalent among blacks -4.10% (p = 0.049) and 13.80% (p = 0.013) -in comparison to whites (10.50% and 20.40%).

CHARACTERISTICS OF HEALTH BEHAVIOR
According to Table 4, regarding health behaviors, the proportion of black elders (52.30%; p = 0.007) who stopped smoking was remarkable, whereas most browns (11%) and blacks (12%) claimed to never care for their own health, when compared to whites (6.30%; p = 0.03).The probable risk for alcohol dependence was higher for browns (22.70%) and blacks (10.70%), when compared to whites (6.60%; p = 0.001).

USE AND ACCESS TO HEALTH SERVICES
According to Table 5, fewer black (25.70%) and brown elders (30.20%) had private health insurance plans (p < 0.001), and they paid the lowest mean values for these plans (blacks: R$ 119.84; browns: R$ 172.33; whites: R$ 314.29).As to the difficulty of using health services, there was no significant difference between racial groups, even though the following stands out: 25% of the elders report difficulty to access them, and a significant number of white people have all of the additional examinations requested (p = 0.011).When asked about undergoing an urgency service, browns (25%) were the ones who used this type of health service more often, whereas blacks were the ones that used it the least (16.30%), even if without statistical significance (p = 0.13).Regarding the place of urgency care, 48% of whites have a health insurance plan, whereas approximately 80% of the blacks and browns use SUS (p < 0.001).

DISCUSSION
The results analyzed here, confirming the first hypothesis, showed a scenario that is more favorable to aging among white elders in comparison to browns or blacks when it comes to sociodemographic indicators of health conditions and use and access to health services.
The prevalence of white people was remarkable, especially among the older participants, as well as the important reduction of browns in this age group.In the context of São Paulo, except for indigenous people, indicators such as the lower mean age of death, lower aging index, and higher mean of potential lost living years in the group of brown people 15,16 clearly express the lower life expectancy in this racial group.Nationally, Oliveira et al., based on the National Household Sample Survey (PNAD), 2008 8 , reinforce these findings.
Regarding sex, despite the expected existence of more women in all racial groups, in the black group the exceeding number of women in relation to that of men is lower than in other racial groups.This finding is in agreement with the data from IBGE, which already indicated, in the population pyramid, the same tendency, especially in the older age groups 17 .
Considering marital status, the proportion of single black elders was clear -almost three times higher than the number of white and brown people.The condition of being single, allied to the highest proportion of people without children, may indicate a situation of vulnerability when it comes to the network of social support.This situation has also been observed in the United States, where black elders, compared to whites, reported living alone and having the need for emotional support 18 .It seems that the worst levels of social support (frequency of contacts, diversity of contacts, and help received and provided) are experienced by single elders.The explanation for this situation is that children are essential in providing support to parents at older ages, and single elders are most likely the ones who do not have children, thus having a disadvantage 19 .The way of life of black elders reinforces this unfavorable configuration, since they present the highest proportions of elders who live alone, who never invite people over and who never go out to public places.
On the other hand, in this study, the relevance of religion in the old age, because of the high proportion of those who mentioned having a religion, seems to consider the church as an important source of informal social support.The church provides spaces of socialization and mutual support, such as groups of prayer, volunteer work, charities, and other situations that promote groups.Other analyses in the context of SABE show that the importance attributed to religion is higher among those who live alone, in comparison to people who live with others, and that the elders reported having friendship connections formed in the religious corporation 20 .
Regarding the economic situation, black and brown elders were, proportionally, those who mentioned not having enough income for their daily expenses and the ones who worked the most at the tie.Worse schooling levels were found among blacks and browns, and they presented the highest proportions when it came to not knowing how to read/write a message, showing illiteracy, and the lowest mean values of schooling years.Both the results related to income and those of schooling are consistent in the entire national territory 8 .
Therefore, both the situation of the need of keeping working for economic reasons until older ages and the worse schooling may define a condition that has been considered for years as the main reason for inequalities in the health field.
Rural life in the first 15 years of age, and for at least five years, was more common among browns.Several authors have established that these conditions, called "previous" 21 , generate disadvantages for individuals living in rural areas or the suburbs of major urban centers, where health services and sanitary conditions are supposedly unfavorable if compared to the central urban regions.So, it is important to analyze if, among the brown people from São Paulo, the unfavorable conditions experienced in the rural zones in the earlier days could also determine worse life conditions maintained throughout the years in the city of São Paulo.This also leads to lower life expectancy in this racial group.
The worse health pattern of black and brown elders observed in this study is mainly related to the high prevalence of arterial hypertension, diabetes, and stroke.Cerebrovascular diseases, diabetes mellitus, and hypertension have been pointed out as the most important causes of mortality among black people in Brazil 22 and in the United States 18 .Among Brazilians, cerebrovascular diseases were mostly associated with poverty in early stages of life, and among North-Americans this type of condition was more prevalent among black people because of the constant situations of underdiagnosis and undertreatment to which they have been submitted 23 .
Mental health issues are another aspect of health conditions, in which black and brown populations are negatively different from the white population.Among blacks and browns, higher proportions of cognitive impairment were observed and assessed by a screening scale.Studies in this field in several countries suggest that the association between poverty and mental disorders is universal, regardless of the level of development in the country 23 .These investigations state that the association of income distribution with homicides, violent crimes, and deaths related to alcohol consumption reinforce the conception that income inequalities also have psychosocial effects.
On the contrary, an unexpected result was mentioned regarding the prevalence of depression.It was significantly higher among white elders and lower among negros elders.Even though it is necessary to conduct a more careful analysis, such results seem to be more consistent with the propositions of Barros et al., who debate the matter of underdiagnosis and undertreatment provided to black elders 24 .
The literature analyzed did not show Brazilian studies carried out to explain ethnical and racial inequalities in terms of health, even though several indicators show significant variations in mortality and morbidity rates between racial groups 22 .Likewise, there are no explanations for some unfavorable behavioral characteristics that are more frequently observed among blacks and browns, such as smoking and alcohol dependence.However, with the clear association of these unfavorable behavioral characteristics and others, like the lack of care with health among groups of black and brown elders, it apparently is not difficult to suggest these racial categories as being predictors of worse health indicators.
Regarding the use and access to health services, it was possible to observe the unfavorable condition of black and brown people in comparison to white people, concerning the use of supplementary health, since the former reported using health insurance plans less; when they have it, they pay lower prices.It is interesting to notice that despite this visible difference in the type of public or private health service used by negro and white people, there are no differences between racial groups regarding difficulties to access health services.However, it important to mention that for 25% of the elders, in all racial groups, it is still difficult to access health services, both public and private.
Another remarkable aspect was the fact that there was no difference between racial groups regarding the access to the request of additional examinations and medication (data not shown in tables).However, there is the hypothesis that the difference lies on the time spent between the request of the examination and its performance, since a significant higher number of white people claimed to have had all of the requested examinations, unlike blacks and browns, who declared only having scheduled the examinations or only some of them.
On the one hand, it is possible to conclude that SUS has ensured access to health services in a relatively adequate way.On the other, mortality rates show a higher proportion of early deaths among blacks and browns, which points out to the need of studies that can qualify the services in order to understand the possible inequalities that may be occurring in the health services of SUS.A related study was conducted in the United States and identified that inconclusive diagnostic conditions, incomplete treatment and lack of control of factors associated with the occurrence of stroke were common for black people, and identified the lack of access to health services as one of the causes for it 23 .
This paper has worrisome results regarding inequalities between the racial segments that on until the old age.Further studies can measure and help to create strategies to eliminate/ ease changeable factors such as schooling, income, and diseases.Interventions focused on the search for equality can be beneficial for adjusting services that compensate the main conditions of vulnerability in these racial groups.
To conclude, it is necessary to point to the main limitation of the study, which is the fact that it was not designed to analyze the variable race/color.There was also a probable selection bias that occurred during the development of field work, causing possible relative sampling error in racial groups, because in the general population there are more white people than browns and blacks, proportionally 17 , than in the study sample.The hypothesis would be that white elders had fewer chances of being interviewed for living in condominiums, which usually makes it difficult for interviewers.On the other hand, brown and black people had more chances of being interviewed because their households are concentrated in areas where the access to interviewers in general is easier, so, there were more interviews with browns and blacks in comparison to whites.Still, comparative analyses with findings from other studies identified many consistent elements.Therefore, this study is favorable regarding its validity.
The project SABE was approved by the Ethics Committee of the Public Health School at Universidade de São Paulo, according to protocol number 2,044, in March, 2010.

Table 1 .
Demographic and economic characteristics of elders in the Study Health, Well-being and Aging (SABE Study) according to color, 2010.

Table 2 .
Characteristics about the way of life of the elders in the study Health, Well-Being and Aging (SABE Study) according to color, 2010.

Table 3 .
Self-report and self-evaluation of health and other health conditions among the elderly in the study Health, Well-being and Aging (SABE Study), 2010.

Table 4 .
Characteristics of health behavior of the elderly in the study Health, Well-being and Aging (SABE Study), 2010.

Table 5 .
Characteristics of the use of health services and access by elderly in the study Health, Well-being and Aging (SABE Study), 2010.