Artigos

Conteúdo (article):

Associations between race/color and leisure time physical activity among Brazilians

Abstract: The objective of this study was to investigate the association between race/color and leisure-time physical activity (LTPA) in Brazilians. The sample was 87,852 adults who participated in the 2019 National Health Survey. The dependent variables were meeting the recommended LTPA and the main LTPAs of walking, weight training. The independent variable was self-reported race/color. Blacks (black and brown races) formed one group, and whites and Asiatics formed another group. Prevalence ratios (PRs) and confidence intervals (95% CIs) were calculated for men and women in general and stratified by income and education level. In the overall crude analysis, the prevalence of the recommended engagement in LTPA was higher in white/Asiatic men, but the association was inverted in the adjusted analysis, being higher in black men (PR = 1.07; 95% CI: 1.03–1.12). According to the adjusted analysis, a higher prevalence of soccer as the main activity was observed among black men (PR = 1.28; 95% CI: 1.20–1.37). In women, no associations were observed. In the analyses stratified by income and education, the associations indicated greater LTPA among black people, especially soccer. Race/color had a moderate relationship with LTPA, varying according to the type and level of income and education.
Keywords: Physical Activity. Racial Groups. Sociodemographic factors. Socioeconomic disparities in health.

INTRODUCTION
Physical activity is considered an important health behavior because of its relationship with the prevention of chronic noncommunicable diseases.1 Physical activity guidelines for adults worldwide2 and in Brazil3 provide recommendations to address these and other benefits, especially for the population’s leisure time, as this domain is usually the domain most closely related to the subjects’ choices, cultural aspects, and enjoyment, among others.4
Even if physical activity is considered an inalienable right5 and a determinant and conditioning factor of health according to the Brazilian Constitution6, about 70% of the adult population in the country did not meet the World Health Organization recommendations for their free time in 20197. This shows that often, access to this practice is still more of a privilege than a right. In addition, it is a complex behavior and is marked by strong social inequalities.4,8,9
Specifically, despite improvements in social indicators in recent decades, and despite the inequalities in race/color, the Brazilian population self-declared as black or brown, who make up 54% of the population, still has worse social indicators than whites10,11. Such disparities are not natural and can be attributed in part to the more than three centuries of enslavement, followed by an abolition devoid of reparatory policies for the enslaved group and its descendants. This context resulted in the “cycle of cumulative disadvantage”, which is still observed today.12 Furthermore, the inequities resulting from racial discrimination and racism, in their individual, institutional and structural manifestations, add to these disparities. These factors may influence exposure to certain health risk factors.13 This highlights the importance of investigating race/color as a health determinant14. Considering the diversity of classification systems existing in different countries, which reflects the specificities of each context, it is clear that in Brazil the population is classified by race/color by the Brazilian Institute of Geography and Statistics (IBGE), the country\'s official statistics body, through five categories: branco (white), preto (black), pardo (brown), amarelo (Asian origin) and indigena (indigenous). In the country, negros (blacks) is the sum of pretos (blacks) and pardos (browns).
In line with this, epidemiological studies that investigated the association between leisure-time physical activity (LTPA) and race/color based on national surveys with representative samples of the Brazilian adult population15–17 have indicated a lower prevalence of LTPA in black and also brown women, as well as in blacks of both sexes, than in whites. Notably, in these studies, race/color was used only as one of the independent variables, and/or LTPA was not the main dependent variable investigated.
Regarding other health indicators, it was observed that smoking, consumption of meat with visible fat, consumption of fruits and vegetables, and excess body weight, blacks and/or browns had worse indicators, both in crude analyses, as in the adjusted analyses17,18. On the other hand, in the case of hypertension, the crude analysis showed a higher prevalence among browns, while the adjusted analysis indicated a lower prevalence among browns.19
In this context, sociodemographic and economic characteristics (age group, education level, income, sex), alone or in combination, are highly important determinants of health. These factors should be included in analyses that seek to verify the effect of race/color on health behaviors17. Considering the various disadvantages that exist for ethnic-racial groups and their relationships in the determination of LTPA, studies on race/color specificity associated with LTPA are necessary and relevant20. In this sense, a better understanding of this relationship can help in the planning and execution of public actions and policies aimed at eliminating any inequities.
Thus, the objective of the present study was to investigate the associations between race/color and LTPA among men and women, in general and stratified by level of education and per capita family income.

METHODS
This was an epidemiological cross-sectional study on secondary data from the 2019 National Health Survey (PNS), a nationwide household survey developed by the Ministry of Health in partnership with the Brazilian Institute of Geography and Statistics. The target population of the PNS is residents in permanent private households belonging to the geographical coverage area of the study. More information on the PNS sampling process is available7. The PNS sample consisted of 88,531 Brazilian adults aged 18 years or older from all 26 Brazilian states and the Federal District who responded to the individual questionnaire. A total of 670 individuals who declared themselves indigenous (due to the specificities of this population) and nine individuals without information regarding race/color were excluded from the study, leaving 87,852 respondents for inclusion. The PNS was approved by the National Research Ethics Committee (Opinion 3,529,376) and complied with Resolution 196/96 of the National Health Council.
The dependent variables were four indicators of LTPA: 1. recommended engagement in LTPA; 2. walking as the main LTPA; 3. weight training as the main LTPA; and 4. football as the main LTPA.
Regarding the recommended engagement in LTPA, individuals who achieved the World Health Organization recommendation 2 (≥150 minutes of moderate physical activity or ≥75 minutes of vigorous activity per week) were classified as active. Regarding the other indicators, individuals who reported these activities as their main activities in the last 3 months were classified as walking, weight training, or soccer practitioners. These activities were chosen because they were the most prevalent.
The independent variable race/color was collected by the self-declaration of participants from five response options (white, black, brown, Asiatic, indigenous), which were dichotomized into two groups: white/Asiatic and black/brown (individuals who declared themselves to be black orbrown). Considering that this variable is not based on biological markers but rather on a form of social classification21 that allows the verification of inequalities through social indicators, the grouping in the black/brown category does not claim homogeneity of the group in their experiences related to racial discrimination or socioeconomic indicators but rather is based on evidence that indicates that the white/Asiatic social group has advantages that distance them from the others, causing the differences between blacks and browns to be minimal compared to the differences from most privileged group22. It is also important to note that the option for the terminology “race/color” is based on the use of the terms officially used by the Brazilian Institute of Geography and Statistics, the Ministry of Health through its Ordinance No. February 2017, and the National Health Survey 2019 (2020). Regarding the theoretical-conceptual debate, the predominant idea is that the category “color” in Brazil is not restricted to “skin color”, as it encompasses other physical traits, such as hair texture, the nose, and lips23.
The confounding variables and their respective categories were age group (18-24, 25-39, 40-59, 60 years of age or older), per capita family income (0 to ½, ½ to 1, 1 to 2, 2 to 4.5, more than 4.5 minimum wages), education (0-8 years, 9-12 years, >12 years) and macroregion (Northeast, North, Midwest, Southeast, South).
To characterize the sample, a descriptive analysis was performed showing the frequencies of individuals who were white/Asiatic or black/browns according to sociodemographic characteristics (sex, age group, education, per capita family income) and the distributions of the dependent variables. Poisson regression was used to verify the associations (crude and adjusted) between race/color and indicators of LTPA from the calculation of the prevalence ratio (PR) and 95% confidence interval (95% CI). The analyses were first stratified by sex and then by years of education and income, separately for men and women. The income variable used in stratification was categorized into <2, 2-3, and >3 minimum wages, and education level was categorized into 0-8, 9-12, and >12 years. Interaction tests were also performed to see whether the effect of race/color on the four dependent variables of the study varied according to the levels of education and income, with p < 0.20 indicating a different effect between groups (see results in the supplementary table). For comparison purposes, white/Asiatic race group was the reference category. The analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 22.

Results
Most of the sample was women (52.6% and 53.9% for the black/brown and white/Asiatic group, respectively). The most prevalent age group in both groups was 40-59 years (35.7% among whites/Asiatics and 35.0% among blacks/browns); however, in the stratum 60 years or older, there was a higher percentage of whites/Asiatics (25. 3%) compared to blacks/browns (18.6%). In both groups, the stratum of education most frequently observed was 0-8 years (51.7% and 40.6% for blacks/browns and whites/Asiatics, respectively), but the proportion of those who had >12 years of education was almost double among whites/Asiatics (28.7%) compared to blacks/brown (14.6%). Similarly, for per capita income, the proportion of black/brown people in the stratum from 0 to ½ minimum wage was more than double that of white people. Conversely, the proportion of whites in the ≥4.5 minimum wages stratum was 4 times greater than that in the black/brown stratum.
Regarding LTPA, 31.6% of white/Asiatic individuals and 28.8% of black/brown individuals were classified as active. Among whites, 15.3%, 8.7%, and 5.2% reported walking, weight training, and soccer as the main physical activities, respectively. Among black/brown people, these values were 14.0%, 5.9%, and 8.0%, respectively (Table 1).

*INSERT TABLE 1 HERE*

In the analysis of the association of the indicators of LTPA with race/color, without any stratification, the crude analyses revealed, in men, that the prevalence of compliance with the WHO recommendations was higher in whites/Asiatics than in blacks/browns (PR = 0.93; 95% CI = 0.91–0.96). This association was inverted after adjustment for confounders, indicating a 7% greater probability of blacks/browns being active in their leisure time than whites/Asiatics (PR = 1.07; 95% CI: 1.03–1, 12). Among women, in the crude analysis, black/browns women were less active (PR = 0.88; 95% CI: 0.85–0.91), but in the adjusted analysis, this relationship went away (PR = 1.03; 95% CI: 0.99–1.08).
In complementary analyses, the impact of each adjustment variable individually and in pairs was tested for the impact of entering the model. In men, the inversion of the association occurred when schooling and per capita income were included alone and was even greater when both variables were included in the model together. For women, there was an inversion in the adjustment for per capita income and the adjustment for both income and education. The other adjustment variables considered in this study (age group and macroregion) did not change the results (compared to the results of the crude analysis).
In the crude analysis, there was a lower proportion of black/brown men (PR = 0.79; 95% CI: 0.75–0.83) with walking as the main LTPA, similar to weight training (PR = 0.61; 95% CI: 0.57–0.66). After adjustments, in both cases, the association was no longer observed. Women had a lower probability of weight training as the main LTPA (PR = 0.75; 95% CI: 0.70–0.81) in the crude analysis, but not in the adjusted analysis.
In both sexes, complementary tests of models with only one or two adjustment variables indicated that no significant changes in walking activity were found when one or two adjustment variables were included in the models. For weight training, education and income seemed to have greater impacts. In males, after adjustment for both of these variables, there was a loss of association. In females, the effect was stronger, as the association was lost when adjusting only for education and was inverted in favor of blacks/browns when adjusting only for income or adjusting for income and education combined. Regarding soccer as the main LTPA, a higher prevalence was observed in black/brown men, both in the crude analysis (PR = 1.65; 95% CI: 1.56–1.75) and in the adjusted analysis (PR = 1.28; 95% CI. %: 1.20–1.37). In women, this higher prevalence was only in the crude analysis (PR = 2.00; 95% CI: 1.56–2.55). Most of the complementary interaction analyses indicated that the effect of race/color on LTPA differed according to education level and family per capita income (p < 0.2).

*INSERT TABLE 2 HERE*

According to the analyses stratified by schooling and the adjusted analyses, black/brown men were more likely to be active in their leisure time and to report soccer as the main LTPA at all levels of education and after adjustment for confounders. In addition, black/brown men with >12 years of schooling were 13% more likely to report walking as the main LTPA than white/Asiatic men were (PRadjusted = 1.13; 95% CI: 1.02–1.26). Regarding sex, adjusted analyses revealed that, compared to white/Asiatic women, black/brown women were more active among groups with 9-12 years of school (PRadjusted = 1.08; 95% CI: 1.02–1.14) and >12 years of schooling (PRadjusted = 1.08; 95% CI: 1.03–1.13). Among the other practices, black/brown women with 0-8 years of schooling were more likely to report soccer as the main LTPA (PRadjusted = 1.97; 95% CI: 1.11–3.51), while black/brown women aged >12 years reported more walking as the main LTPA (PRadjusted = 1.22; 95% CI: 1.11–1.32) (Table 3).

*INSERT TABLE 3 HERE*

When the analyses were stratified by per capita family income, in the stratum of up to two minimum wages, black/brown men were more likely to be active in their leisure time (PRadjusted = 1.09; 95% CI: 1.04–1.14). Additionally, among men, the proportion of black/brown people who reported soccer as the main LTPA was higher in all income strata. For females, adjusted analyses revealed an association between income and race/color only in the 2-3-minimum-wage stratum, where the probability of black/brown women reporting walking as the main LTPA was 15% greater than that of white/Asiatic women (PRadjusted = 1.15; 95% CI: 1.00–1.32) (Table 4).

*INSERT TABLE 4 HERE*

DISCUSSION
In general, the results of the present study suggested some differences between race/color and indicators of LTPA, which were almost always favorable to black/brown people, when controlling for some possible confounders, especially income and education. Black/brown men were more active in their leisure time than white men at all levels of education and among those with lower per capita family income. Regarding the practices reported as the main LTPAs, soccer was more cited by black/brown men in the crude analysis and at all levels of education and income and by women only among those with less education. According to the adjusted analyses, no association between race and color was observed with weight training as the main LTPA. Regarding walking, an association was only observed in the group with higher education, and black/brown men and women cited this as the main LTPA more than whites with the same level of education.
Notably, we used four indicators of the LTPA, each with some important specificities. The first indicator, meeting the WHO LTPA recommendation, is an indicator that encompasses all types of physical activity practiced by survey respondents in their free time and uses the most traditional WHO criterion (per week, moderate intensity ≥ 150 minutes or ≥ 75 minutes of vigorous intensity). In contrast, the three other indicators (walking, weight training, and soccer) were gathered by asking which LTPA the subjects engaged in most frequently. Thus, we must be careful when interpreting these indicators as “favorite” physical activities. It is likely that, in some cases, the activity referred to as the main LTPA is not necessarily the activity that the person most enjoys or wants to practice but rather the activity with greater availability and/or ease, since each activity has particularities in relation to access and practice.
Regarding specifically the indicator that allows a more specific idea of the level of LTPA (because it includes duration and weekly frequency), significant differences between the crude and adjusted analyses were observed in the nonstratified analysis. In the crude analysis, black/brown men and women were less active than other races. However, in the adjusted analysis, this association ceased to exist among women and was reversed in black/brown men, i.e., black/brown men became more active. In their crude analyses, similar studies also revealed that in both sexes, whites were more active than blacks or browns, but after adjustment, the differences among women ceased to exist15,16,24.
Considering all the LTPA indicators used in the present study, the differences in the crude analyses compared to the adjusted analyses are an important point to be considered. A total of 56 crude associations were tested, of which almost half (27; 48.2%) did not show a significant association, while in 18 (32.2%) blacks/browns were more active and/or had a higher prevalence of a specific practice, and in 11 (19.6%) Caucasians did. In the adjusted analysis, of the 27 results without associations in the crude analysis, four became significant. Of the 18 in favor of blacks/browns, 13 stayed significant, and of the 11 in favor of whites, none stayed significant and one was even reversed, becoming favorable to blacks/browns.
The results obtained in this study contrast with those found in the literature on race/color and health behaviors, which requires further investigation of the specificities of LTPA associated with race/color for the understanding of inequalities and inequities. It is worth mentioning, that a study on hypertension with data from the 2019 PNS, which divided race/color categories into white, black, and brown, found that brown people of both sexes had a lower prevalence than whites in the crude analysis but that this association was inverted after adjustment19, similar to some results of our study. It was not the objective of the present study to delve into methodological issues related to data analysis, but we consider these findings important indicators of the care required in the treatment of statistical analyses, especially when they involve the issue of race/color. It is not uncommon for some studies to perform or present only adjusted data for those variables that were found to be associated in the crude analysis, but our findings demonstrate that it is even possible to invert some results when doing crude vs. adjusted analyses.
Furthermore, as mentioned in the Results section, it is important to highlight the influence of education and per capita family income on the relationship between race/color and LTPA. Seeking a deeper understanding of this aspect, we performed complementary analyses that considered only these isolated variables and observed great strength of these variables for inversion of the associations. However, adjustments for income and education should be analyzed sparingly. These factors are part of the determination path (causal path) in the evaluated associations. The worse health outcomes observed in the black/brown population may come mainly from two factors: (1) socioeconomic factors, considering that black/brown people tend to be more vulnerable to these factors, and (2) racial segregation and racism, or both. If we remove the effect of socioeconomic factors, which was very strong in this case and where it was also subject to intervention, racial segregation and some residue will remain. It was expected in this study that, controlling for part of this causal pathway, blacks/browns would still be less active, which in general was not confirmed because some associations observed in the crude analysis ceased to exist after the adjustments, or were even inverted. The findings of this study show that the LTPA may have some different specificities in relation to other health indicators, where even after adjustments, black/brown people continued to have worse indicators than white people17,18.
In addition, these results reinforce the importance of the search for racial equity through the indicators of income and education in this sense since the disadvantages observed for black/brown people were greatly influenced by these factors in the analyses. There is also a need for new studies to confirm these findings, as well as articles that seek a deeper understanding of the impact of racism on physical activity, given that health behaviors can be influenced by the issue of racial discrimination25,26.
Another aspect worth mentioning is that the race/color classification and/or analysis were not identical in the aforementioned studies and that this lack of standardization was observed in a review study that investigated how the race/color variable had been classified in studies on physical activity. In its analysis of 20 articles, nine forms of stratification and 10 different nomenclatures were identified. On the other hand, studies on data from national surveys generally categorize race/color in a similar way, including categories present in the Brazilian Institute of Geography and Statistics classification20.
Regarding the two variables used for stratification, it is worth mentioning that more associations were observed in the stratification by education than by per capita family income. This result may be related to the accuracy of the information, since it is easier for most people to accurately report their level of education than their income because the latter may vary monthly, for example.
Our findings show possibilities by stratifying analyses and looking for results in certain social characteristics (e.g., LTPA in young, highly educated black/brown women), which sometimes allow us to observe associations that were hidden when analyzed without stratification. The many effect changes/interactions attributed to race/color in different education levels or different income levels on the indicators of LTPA (see complementary material) indicated that stratification was relevant for the purpose of this study.
Health disparities between subgroups are an important issue 27 and have gained attention from recent intersectional analyses28, which are tools that analyze the impact of combinations of disadvantages on dependent variables. On this last aspect, Mielke et al. (2022)28 investigated LTPA in intersection analyses and reported that only 10% of black/brown women with low education and income performed sufficient LTPA, while in white men with high income and education, the prevalence was approximately 48%, showing that LTPA varies largely on these factors.
Regarding the indicators of the main LTPA, greater differences in prevalence were found for soccer, especially among men, black/brown people reporting this activity more as their main activity. This result can be explained, at least partially, by the influence of the cultural and social aspects of football. At the beginning of organized Brazilian football, it was thought of as an elitist and exclusionary game, but after 1920, popular clubs began to integrate blacks into their squads. A major step forward was when Vasco da Gama, a club from Rio de Janeiro, was Rio de Janeiro champion in 1923, the year that it included blacks/browns on the team. Even with a blind eye to their actions, the racism and discrimination suffered, blacks/browns began to play a central role in Brazilian football30,31,32.
Soccer plays an important role in promoting physical activity in part of the Brazilian population, especially males. However, a previous study revealed that people who cited soccer as the main LTPA had a lower weekly frequency of practice than people who cited walking, weight training, and cycling, among others. Nevertheless, the average time of practice per day was higher in soccer than in the other activities33. In complementary analyses, which excluded individuals who reported soccer as the main LTPA, the association for LTPA ceased to exist in general and in all groups, except for the lowest income group. This analysis indicated that many of the observed increases in LTPA scores in black/brown people (in the adjusted analyses) seem to be explained by the greater time spent on soccer in this group.
With regard to the main LTPA, it is interesting to note the higher prevalence of walking among black/brown people with more years of education, especially women. In this sense, it would be important to verify whether this result is related to a lack of access or time due to longer working hours and/or more time spent on household chores compared to white women34. It is also worth reflecting on whether this practice has been taking place in informal spaces of physical activity, given that walking is sometimes accessible and can be practiced in parks, villages, and sidewalks without necessarily following professional guidance.
A limitation of this study is the measurement of physical activity through a questionnaire. The reported information may be overestimated, as LTPA is a socially desirable behavior.35 This type of strategy is widely accepted and used in this type of study, as it is still unfeasible to objectively measure physical activity in large samples, especially in population-based studies. It is also noteworthy that we do not have information related to people’s perceptions of their personal experiences of experiencing situations of prejudice, for example, which could help to better understand the relationship between race/color and the practice of LTPA. In this regard, in addition to the need for new studies as discussed above, we suggest that such studies seek to use qualitative methods, which can be important to better study some elements that usually cannot be studied in quantitative studies. In addition, longitudinal studies that investigate the impact of racial prejudice and discrimination may be important.
On the other hand, the originality of the work should be highlighted, especially because a recent review20 found no studies that specifically investigated this subject in an in-depth and specific way, since articles that usually include race/color analyze it together with other variables, such as sex, income, age group, education, and marital status. Thus, this is possibly one of the first studies to explore the relationship between race/color and LTPA in Brazilians.
The data of the present study indicate that in the Brazilian context, there seems to be a moderate relationship between race/color and the practice of LTPA. Our findings reinforce that much of this relationship is influenced by social indicators such as income and education. Among the activities indicated as most habitual, soccer was cited most often by blacks/browns (which even seems to help explain part of the result that blacks/browns had a higher prevalence of general practice of LTPA). Regarding walking and weight training, few associations were observed, suggesting that there is little difference regarding race/color in these activities.

References
1. World Health Organization. Global action plan on physical activity 2018–2030: more active people for a healthier world [Internet]. Geneva: World Health Organization; 2018 [citado 15 de julho de 2023]. 101 p. Disponível em: https://apps.who.int/iris/handle/10665/272722
2. World Health Organization. Guidelines on physical activity and sedentary behaviour. Geneva: World Health Organization; 2020.
3. Brasil M da S. Guia de Atividade Física para a População Brasileira [Internet]. Brasília, DF: Ministério da Saúde; 2021. Disponível em: https://bvsms.saude.gov.br/bvs/publicacoes/guia_atividade_fisica_populacao_brasileira.pdf
4. Crochemore-Silva I, Knuth AG, Mielke GI, Loch MR. Promoção de atividade física e as políticas públicas no combate às desigualdades: reflexões a partir da Lei dos Cuidados Inversos e Hipótese da Equidade Inversa. Cad Saúde Pública. 2020;36(6):e00155119.
5. Programa das Nações Unidas para o Desenvolvimento (PNUD). Relatório de desenvolvimento humano nacional – movimento é vida: atividades físicas e esportivas para todas as pessoas [Internet]. Brasília: PNUD; 2017. Disponível em: http://www.each.usp.br/gepaf/wp-content/uploads/2017/10/PNUD_RNDH_completo.pdf
6. Brasil. Lei no 12.864, de 24 de setembro de 2013. Alteração do caput. 12.864, de 24 de setembro de 2013 2013.
7. Instituto Brasileiro de Geografia e Estatística (IBGE). Pesquisa Nacional de Saúde 2019: Percepção do estado de saúde, estilos de vida e doenças crônicas [Internet]. IBGE; 2020. Disponível em: https://biblioteca.ibge.gov.br/visualizacao/livros/liv101764.pdf
8. Botelho VH, Wendt A, Pinheiro EDS, Crochemore-Silva I. Desigualdades na prática esportiva e de atividade física nas macrorregiões do Brasil: PNAD, 2015. Rev Bras Atividade Física Saúde [Internet]. 29 de junho de 2021 [citado 15 de julho de 2023];26. Disponível em: https://rbafs.org.br/RBAFS/article/view/14531
9. da Silva ICM, Mielke GI, Bertoldi AD, Arrais PSD, Luiza VL, Mengue SS, et al. Overall and Leisure-Time Physical Activity Among Brazilian Adults: National Survey Based on the Global Physical Activity Questionnaire. J Phys Act Health. 1o de março de 2018;15(3):212–8.
10. Fontoura, N., Rezende, M., Lobato, A. L., et al. Retrato das Desigualdades de Gênero e Raça – 20 anos. Instituto de Pesquisa Econômica Aplicada; ONU Mulheres; 2017.
11. Instituto Brasileiro de Geografia e Estatística (IBGE). Pesquisa Nacional por Amostra de Domicílios Contínua: Primeiro Trimestre de 2021. Diretoria de Pesquisas, Coordenação de Trabalho e Rendimento, Indicadores IBGE; 2021.
12. Hasenbalg C. Discriminação e Desigualdades Raciais no Brasil. 1° ed. Rio de Janeiro: Graal; 1979.
13. Almeida SL. Racismo Estrutural. São Paulo: Carneiro S; Editora Jandaíra; 2020.
14. Almeida-Filho N. Ethnoepidemiology and mental health: insights from Latin America. Salud Colect. 11 de setembro de 2020;16:e2786.
15. Casas RCRL, Bernal RTI, Jorge ADO, Melo EMD, Malta DC. Fatores associados à prática de Atividade Física na população brasileira - Vigitel 2013. Saúde Em Debate. dezembro de 2018;42(spe4):134–44.
16. Malta D, Andrade S, Stopa S. Estilos de vida da população brasileira: resultados da Pesquisa Nacional de Saúde, 2013. Epidemiologia e Serviços de Saúde. 2015;24(2):217–26.
17. Malta DC, Moura LD, Bernal RTI. Differentials in risk factors for chronic non-communicable diseases from the race/color standpoint. Ciênc Saúde Coletiva. março de 2015;20(3):713–25.
18. Meller FDO, Santos LPD, Miranda VIA, Tomasi CD, Soratto J, Quadra MR, et al. Desigualdades nos comportamentos de risco para doenças crônicas não transmissíveis: Vigitel, 2019. Cad Saúde Pública. 2022;38(6):e00273520.
19. Malta DC, Bernal RTI, Ribeiro EG, Moreira AD, Felisbino-Mendes MS, Velásquez-Meléndez JG. Hipertensão arterial e fatores associados: Pesquisa Nacional de Saúde, 2019. Rev Saúde Pública. 7 de dezembro de 2022;56:122.
20. Chiconato AG, Evedove AUD, Panta MADS, Loch MR. A variável raça/cor em estudos epidemiológicos brasileiros sobre atividade física (2015-2019): formas de estratificação e principais resultados. Pensar Prática [Internet]. 27 de maio de 2022 [citado 15 de julho de 2023];25. Disponível em: https://www.revistas.ufg.br/fef/article/view/69858
21. Guimarães A. Racismo e Anti-racismo no Brasil. São Paulo: Editora 34; 1996.
22. Batista LE, Monteiro RB, Medeiros RA. Iniquidades raciais e saúde: o ciclo da política de saúde da população negra. Saúde Em Debate. dezembro de 2013;37(99):681–90.
23. Guimarães ASA. Raça, cor, cor da pele e etnia. Cad Campo São Paulo - 1991. 30 de março de 2011;20(20):265–71.
24. Mielke GI, Stopa SR, Gomes CS, Silva AG da, Alves FTA, Vieira MLFP, et al. Leisure time physical activity among Brazilian adults: National Health Survey 2013 and 2019. Rev Bras Epidemiol Braz J Epidemiol. 2021;24(suppl 2):e210008.
25. Rodrigues YE, Fanton M, Novossat RS, Canuto R. Perceived racial discrimination and eating habits: a systematic review and conceptual models. Nutr Rev. 9 de junho de 2022;80(7):1769–86.
26. Fanton M, Rodrigues YE, Schuch I, Cunha CMDL, Pattussi MP, Canuto R. Direct and indirect associations of experience of racial discrimination, dietary patterns and obesity in adults from southern Brazil. Public Health Nutr. 1o de fevereiro de 2024;1–23.
27. Garcia GAF, SIlva EKPD, Giatti L, Barreto SM. The intersection race/skin color and gender, smoking and excessive alcohol consumption: cross sectional analysis of the Brazilian National Health Survey, 2013. Cad Saúde Pública. 2021;37(11):e00224220.
28. Mielke GI, Malta DC, Nunes BP, Cairney J. All are equal, but some are more equal than others: social determinants of leisure time physical activity through the lens of intersectionality. BMC Public Health. dezembro de 2022;22(1):36.
29. Knuth AG, Antunes P de C. Práticas corporais/atividades físicas demarcadas como privilégio e não escolha: análise à luz das desigualdades brasileiras. Saúde E Soc. 2021;30(2):e200363.
30. Filho M. O negro no futebol brasileiro. 5° ed. Rio de Janeiro: Mauad; 2003.
31. Levine R. Esporte e Sociedade: o caso do futebol brasileiro. Em: Futebol e Cultura: Coletânea de Estudos. Imprensa Oficial/Arquivo do Estado; 1982. p. 42–3.
32. Souza M. Gênero e raça: a nação construída pelo futebol brasileiro.
33. Wendt A, Carvalho WRGD, Silva ICM, Mielke GI. Preferências de atividade física em adultos brasileiros: resultados da Pesquisa Nacional de Saúde. Rev Bras Atividade Física Saúde. 9 de outubro de 2019;24:1–9.
34. Pinheiro, L, Fontoura N, Querino A. Retrato das Desigualdades de Gênero e Raça [Internet]. Instituto de Pesquisa Econômica Aplicada; UNIFEM; 2008. Disponível em: https://www.ipea.gov.br/retrato/pdf/terceiraedicao.pdf
35. De Rezende LFM, Azeredo CM, Canella DS, Claro RM, De Castro IRR, Levy RB, et al. Sociodemographic and behavioral factors associated with physical activity in Brazilian adolescents. BMC Public Health. dezembro de 2014;14(1):485.