key: cord-0785223-8ehfuq94 authors: de França, Natália Cecília; Campêlo, Guaracyane Lima; Santos de França, João Mário; Vale, Eleydiane Gomes; Badagnan, Thaísa França title: Socioeconomic inequalities and health status associated with the Covid-19 diagnosis and related symptoms, during the first wave of infections in Brazil: A decomposition analysis date: 2021-10-01 journal: EconomiA DOI: 10.1016/j.econ.2021.09.002 sha: 203eddb878a5ee415e188c12114eb45f1ca07dd2 doc_id: 785223 cord_uid: 8ehfuq94 Recent studies have shown that Covid-19 affects different population groups asymmetrically. This work quantifies the socioeconomic inequality, during the first wave of infections in Brazil, using available data on test diagnosis and related symptoms from the National Survey of Households - PNAD - Covid-19 /IBGE. We use the concentration curve and the concentration index, as well as a decomposition analysis to verify the factors that most influence the inequalities found in the health variables specified. The concentration index was positive for the incidence rate, indicating a greater concentration of diagnoses, number of tests, among groups with higher income levels. When considering symptoms similar to Covid-19 infection, inequality is practically non-existent. A pre-existing disease has a larger contribution to the concentration of Covid-19 in the presence of correlated symptoms, than in its diagnosis, among people with higher income. However, the results of the decomposition show that if the inequalities were explained only by race, and the fact of living in the most vulnerable regions of the country (North and Northeast), there would be a concentration of Covid-19 among the poorest. The distribution of health and disease in a society is related to its socioeconomic, cultural and environmental conditions, what can cause health disparities (Dahlgren and Whitehead, 1991) . The Covid-19 pandemic devastates many health systems and the global economy with dire consequences for individual and national well-being, specially affecting ethnic minorities, and lower socioeconomic status (SES) groups, as poverty and income inequality tend to increase the infection rate. (Liao and De Maio, 2021; Brown and Ravallion 2020) . A large international body of evidence that deals with the problem of health inequities and the outbreak of the SARS-CoV-2 shows an existence of social inequalities in the incidence of COVID-19 by age group, gender, geographic location and income. Such studies find that populations classified as non-white, residing in locations of low socioeconomic status, less educated and those living in poverty were associated with a greater number of confirmed cases and deaths from the disease. ( Shahbazi and Khazaei, 2020; Abedi et al., 2020; Khanijahani 2021; Marí-Dell'Olmo et al., 2021; Gutiérrez, Inguanzo and Orbe, 2021; Davillas and Jones, 2021 ; Mena et al, 2021; Cifuentes et al., 2021 ; Nwosu and Oyenubi , 2021) . Latin America currently has some of the world's highest mortality rates from the new coronavirus and faces a humanitarian crisis fueled by the multiple inequalities as social, economic, educational, ethnic-racial, labor and gender, among others. Lower SES is directly related to the mortality rate of the pandemic, and existing socioeconomic inequalities affects the course of the infection rate, with a disproportionate adverse burden on states and municipalities with high vulnerability. (Mena et al., 2021; Rocha et al., 2021) In this context, special effort need to be done in order to understand how disparities in health and its socioeconomic determining factors affect the ones with the double burden of a Covid-19 infection and a pre-existing noncommunicable diseases (NCD) (Henry, 2020) . Unfortunately, the national literature still lacks addressing the uneven impact of the new coronavirus by a comprehensive socioeconomic analysis with focus on the effects of income inequality over the number of Covid-19 tests realized and appearance of related symptoms, given pre-existing NCD. Most findings determined a disproportionate impact of the pandemic on black and poor populations. These studies focused on specific indicators of poverty, income and the prevalence of antibodies (Baqui et al., 2020; Hallal et al., 2020; Tavares and Betti, 2021) or contributed to the investigation over socioeconomic aspects, infection incidence, risk of hospitalization, and mortality rate. (Demenech et al., 2020; Martins-Filho et al., 2021; Li et al.,2021) We aim to contribute to the literature analyzing the factors which determine socioeconomic inequality in the diagnosis and related symptoms to a Covid-19, during the first wave infection in Brazil. Therefore we use available data of November 2020, considering this the last month of the first wave, from the National Survey of Households -PNAD -Covid-19 /IBGE. The methodology adopted uses the concentration curve (CC) and the concentration index (CI). The identification of the participation of the different elements for the disparities is made through a decomposition analysis based on the method of Wagstaff, Doorslaer and Watanabe (2003) . The empirical results obtained corroborate the literature that emphasizes the race, education, the income-poverty ratio, and employment as the main contributing factors for socioeconomic inequalities during Covid-19 pandemic. In relation to the regional context, residing in the Northeast and North contributes to the concentration of the diagnosis of Covid-19 in the portion of the population most vulnerable socioeconomically, as these regions present a lower average household income per capita (HIPC), a higher percentage of beneficiaries for emergency aid and a reduced possibility of home office. A pre-existing disease (PED) contributes to the concentration of Covid-19 (both diagnosis and correlated symptoms) among people with greater purchasing power, given the positive sign of the contribution of this variable. In this sense, if socioeconomic inequalities in J o u r n a l P r e -p r o o f these PED were eliminated, there would be a reduction in inequalities with respect to the coronavirus. We follow in the next section presenting shortly the methods and offering a data description, along the third session the reader faces the research results, the forth section brings the final remarks, and ends this paper. In order to verify socioeconomic inequality in relation to Covid-19 in Brazil, this study uses the CC and the CI. The CC is a graph that helps to assess the degree of inequality and refers to the distribution of a variable in a certain population group related to another group. For the construction of the figure, the cumulative percentage of the measure referring to Covid-19 (y-axis) is plotted on a Cartesian axis versus the cumulative percentage of the population ranked in ascending order by household income per capita (x-axis). This papers uses the derivation of the CC according to Kakwani (1977) . Concentration curves can occupy any position on the graph. If the curve is below the equality line, as shown in the example in Figure 1 , there is a concentration of the variable of interest in the richest part of the population. When the curve is above the 45° line, the concentration is among the poorest. Furthermore, the farther the CC is from the equality line, the greater the degree of inequality in the analyzed variable. From the CC, the concentration index is obtained as a measure that indicates the degree of inequality in the variable of interest. Its algebraic formalization consists of the areas above and below the equality line delimited by the concentration curve. This index can assume values in the [-1 , 1] interval. The CI will be negative when the concentration curve is above the equality line, showing that Covid-19 is concentrated among the poorest. The values of this measure will be positive when the WC is below the 45º line, indicating a concentration of the infection among the richest part of the population. If the concentration curve coincides with the diagonal, the CI will be null, that is, the closer to the line of equality the concentration curve is, the smaller will be the index. However, as highlighted by Khaled et al (2018) , values very close to zero deserve some caution, as they may be an indication that CC crosses the equality line (and not a low level of inequality). In cases where the variable of interest is a binary variable, the conventional CI assumes values in the range between (μ-1) and (1-μ), where μ is the mean of the variable of interest. In this sense, Wagstaff (2005) suggested a normalization so that the measure assumes values in the [-1 to 1] interval. The normalized index, (CI) can be obtained as follows: One of the objectives of our work is to provide information about the determinants of socioeconomic inequality in Covid-19 in Brazil. Thus, the CI was broken down into contributions made by factors correlated with Covid-19 and with income. The method of Wagstaff, Doorslaer and Watanabe (2003) was implemented, investigating the extent to which inequality in a given variable affects inequality in relation to the infection. This decomposition is not intended to reveal causal effects between variables, but rather to generate relevant information about the factors that are simultaneously correlated with income and the variable of interest (Bilger et al., 2017) . The IC can be decomposed into two parts. Firstly, the deterministic components, given by the weighted sum of the concentration indices of the explanatory variables, which indicate the degree of social inequality in the determining factors. The elasticity of the variable of interest in relation to its independent variable can be estimated in this first part. A positive elasticity shows a direct relationship between the measure referring to Covid-19 and the explanatory variable. Secondly, the residual component and captures the part of the inequality not explained by the independent variables. This work uses data from the National Household Sample Survey -PNAD COVID19, which began to be carried out in May 2020 by the Brazilian Institute of Geography and Statistics -IBGE. The research aims to investigate health aspects, more precisely the symptoms of the new Coronavirus, as well as the socioeconomic impacts of the pandemic in the Brazilian territory. In this sense, the questionnaire includes questions about demography, health, work, income and assistance benefits. PNAD COVID19 relies on telephone interviews in approximately 48 thousand households per week. Since the beginning of the household survey, in May 2020, information has been collected on symptoms related to Covid-19. However, data on the performance of tests for the diagnosis of the disease began to be made available from July 2020. Thus, the analysis of this study refer to the months between July and November 2020, which we will define here, for the objective of this work, as the first wave of the infections. Regarding the treatment of the database, the following were excluded: (i) people under 14 years of age; (ii) pensioner, domestic worker and relative of a domestic worker; and (iii) observations that had some information missing for the variables used in the analyses. The choice to consider only individuals aged 14 years or older is due to the fact that this is the minimum age allowed to work in Brazil (Constitutional Amendment No. 20/1998 ). In the construction of the variable income-poverty, the threshold of a monthly per capita household income of R$178.00 was used. This amount is one of the criteria for registering beneficiaries in the Bolsa Família Program (a federal government initiative, created in 2003 as a way to alleviate poverty). In order to verify the factors that most contribute to inequality in relation to the infection, the dependent variables, namely diagnosis and symptoms related to Covid-19, and independent variables, which include demographic and socioeconomic factors, are defined in Table 1 . Job 1, if during the week before the survey, the person worked (formally or informally) for at least 1 hour; 0, otherwise Pre-Existing disease (PED) 1, if the person previously was diagnosed with diabetes, and/or hypertension, and/or asthma/bronchitis/emphysema/chronic respiratory and/or lung disease, and/or heart disease, and/or depression, and/or cancer; 0 otherwise The total sample considered (people aged 14 or over) represents 80.6% of the Brazilian population, both in November 2020 (Table 2 ) and in July of the same year, which is the period considered here as first wave of infections. Table 1 shows that, among the population considered, 2 572 597 people had tested positive for the new Coronavirus in July 2020, which represents 1.51%. The volume of diagnoses more than doubled in the period analyzed, reaching a total of 6 067 510 people in November 2020, which indicates 3.56% of the population assessed. In turn, there was a drop in the share of individuals who reported some type of symptom related to the flu syndrome. This contingent was 12 539 790 people in July 2020 (7.37% of the population) and went to 7 016 844 in November of the year in question (4.11%). Note that data for November on the number of tests is cumulative for the whole period of the analysis, while data for November on the presence of related Covid-19 symptoms refer only to the last week prior the phone call interview in November. Source: Own elaboration based on PNAD COVID19 Survey. * Due to lack of information, we define first wave as the period from July to November 2020 With low testing capacity, the country has a high percentage of positive results in the total number of tests performed, when compared to the incidence of the disease in the general population. In July, of the total tests performed, 21.17% were positive, while this item corresponds to 23.08% in November 2020 ( Figure 2 ). This may indicate that the search for the diagnosis of Covid-19 is primarily done by symptomatic people. In addition, it is important to emphasize that at the beginning of the pandemic, the tests were aimed at professionals who work on the front lines in combating the new pathogen. In this context, it is extremely important that mass testing be carried out on the Brazilian population, as a way of showing the real situation experienced in this pandemic. According to the data available, in July 2020 only about 7% of the population considered had undergone some type of test for the diagnosis of the new flu syndrome, rising to 15.41% in November of the same year. J o u r n a l P r e -p r o o f Source: Own elaboration based on PNAD COVID19 Survey. * Due to lack of information, we define first wave as the period from July to November 2020 When analyzing the performance of these tests by income quintile, the difference is alarming (Figure 3 ). Among the poorest 20% of the Brazilian population, only 8.75% underwent any test for the diagnosis of the new Coronavirus until the end of the first wave of infections. This value is almost three times lower than that observed among the richest 20%, namely 25.62%. Unfortunately, another facet of inequality experienced by the poor, who have less access to health services. Thus, it is likely an underreporting of cases of the disease among the poorest. * Due to lack of information, we define first wave as the period from July to November 2020 When implementing the decomposition of inequality in Covid-19, socioeconomic and demographic factors of the Brazilian population were considered. One of these variables concerns the person's race, being classified as white and non-white (black, yellow, brown and indigenous). Regarding the non-white population, it should be noted that it is composed almost exclusively of blacks and browns, representing a total of 98.07% 1 . Thus, when mention is made of non-whites, the focus will be on the black population 2 . Table 2 presents the descriptive statistics of the independent variables used in this study, considering the total sample and disaggregating according to people who had a positive diagnosis for Covid-19 or who had symptoms related to the disease. Overall, for all subsamples analyzed, the relative participation of men and women is quite similar; the same is happening between whites and blacks; and most reside in urban areas. As for the educational level, the relative participation of people with higher education is higher among those who tested positive for Covid-19. Similarly, the income-poverty ratio was higher among this group of individuals. This may be associated with the high price of the tests, making it difficult to diagnose the disease among the most vulnerable people. Another interesting result is the higher percentage of people with PED among those who reported some type of symptom related to . Finally, it is worth noting that the South and Southeast, the two richest regions in the country, represent 57.42% of the sample considered. In turn, evaluating the Covid-19 diagnosis, this participation drops to 47.95%. This is an indication that the new Coronavirus affects more, in relative terms, less developed regions. The results found are in accordance with the specialized literature. (Abedi et al., 2021; Clouston, 2021; Baqui et al., 2020) . 1 Data relate information from November, 2020. 2 As stated in the Racial Equality Statute (Law 12 288/2010), the black population is made up of people who declare themselves black and brown, according to the color or race item used by the IBGE, or who adopt a similar selfdefinition. Covid Source: Own elaboration based on PNAD COVID19 Survey. Figures 4 shows the distribution, by income quintile, of people diagnosed positive for Covid-19 (cumulated value for the first wave of infections), and with symptoms related to the flulike illness (reported in November 2020, related to one week prior the phone interview). The prevalence of positive tests for the new Coronavirus grows with income. In turn, the share of individuals who reported some symptom is more even among the income distribution quintiles, being slightly higher among the richest. *Due to lack of information, we define first wave as the period from July to November 2020 The results in Table 6 highlight that the concentration index is statistically significant both for the diagnosis and for the symptoms related to the new Coronavirus in the analyzed period. Both indices indicate a pro-rich inequality, that is, there is a higher prevalence of positive tests for Covid-19 and people with some type of symptom among the groups with greater purchasing power. Such evidence corroborates the results obtained by Davillas and Jones (2021) . The extent of this inequality is much more accentuated when considering the people who tested positive than when evaluating those who reported some type of symptom (the CI of this last variable is very close to zero, indicating that there is almost no inequality in its distribution) . The positive sign of the CI for the diagnosis of the new Coronavirus is in agreement with the results of Shahbazi and Khazaei (2020) , which show a concentration of the incidence of Covid-19 in countries with a higher human development index (in general, these places have higher income level). According to the authors, this may be associated with better health systems in wealthier nations, with a better structure for mass testing and early detection of the disease, even in asymptomatic cases. They also highlight that in countries with a low HDI, poor access to diagnostic services means that there are low rates of disease incidence. In the Brazilian context, the high cost of testing for Covid-19, combined with the lack of mass testing carried out by the government, contributes to the underreporting of cases of the disease among the poorest people. Figure 5 shows the concentration curve for the diagnosis of Covid-19 and flu-related symptoms in November 2020. Confirming the result of the previous section, the incidence of the new Corona virus is concentrated in the richest part of the population, given that the concentration curve is located below the equality line. And, as expected, social inequality practically disappears when the symptoms of the disease are considered (concentration curve very close to the 45º line). Source: Own elaboration based on PNAD COVID19 Survey. * Due to lack of information, we define first wave as the period from July to November 2020 In order to investigate the factors that contribute to socioeconomic inequality in the diagnosis of Covid-19 and in the symptoms related to the disease, a decomposition analysis was performed. Table 6 shows the concentration index for each explanatory variable, capturing socioeconomic inequality in the factor itself. The estimated coefficients, elasticity, as well as the absolute and relative contribution of each factor to the concentration indices of the analyzed health variables are also presented. A first point that deserves to be emphasized is the high concentration of people living in rural areas and of black people in the poorest part of the population (both variables had a negative concentration index). This confirms racial inequality in Brazil, which disadvantages black people (Campante, Crespo and Leite, 2004; Chadarevian, 2011) . Similarly, individuals with complete primary education are concentrated in the lower part of the income distribution (CI = -0.1844). On the other hand, as expected, there is a high concentration of people with complete higher education and people employed in the labor market among the richest sections of the population (the respective concentration indices were positive). With regard to social inequality between Brazilian regions, the data highlight the vulnerability of the North and Northeast. Both regions had negative CI, indicating a concentration of people residing in these locations in the lower part of the income distribution (poorest population). This fact is evidenced in the studies by Rocha et al (2021) and Baqui et al. (2020) . The factors that most contributed to socioeconomic inequalities in relation to the new coronavirus were the person's race, level of education (completed higher education), the incomepoverty ratio, and employment. According to the data in Table 6 , the person's race explains -39.96% of the inequality in symptoms related to Covid-19. Thus, if the inequalities related to the symptoms of the new coronavirus were determined only by race of Brazilians, there would be a concentration of Covid-19 in the poorest part of the population. This evidence is in line with studies that show a greater probability of poor and black people to die victims of Covid-19, due to the difficulty in accessing quality health services, the greater difficulty in maintaining social isolation and the higher prevalence of comorbidities. (Li et al., 2021; Khanijahani, 2021; Abedi et al., 2021; Liao and De Maio, 2021; Nwosu and Oyenubi (2021) , Baqui et al., 2020) . The data used in this work corroborate the greater social vulnerability among black people. As shown in Table 3 , the average per capita household income is much lower among blacks compared to whites (blacks' income is practically half that of whites) . In addition, among blacks and browns, there is a greater portion of those benefiting from emergency aid (which assists people in vulnerable situations). Finally, the remote work perspective is less plausible among the black portion of the population, making social isolation difficult among this population group. These factors make black people more exposed to Coronavirus infection. However, the high prices of diagnostic tests may be causing an underreporting of the disease in this group. Also according to Table 6 , the positive contributions of education and income weighted by the poverty line indicate that, in the absence of inequalities in these variables, inequalities in relation to Covid-19 would decrease. In turn, the employment variable affects inequality in these two variables differently. It has a positive effect on the CI for the diagnosis of Covid-19, showing that having an occupation contributes to a slight concentration of positive tests among the richest part of the population. In fact, having an occupation in the labor market contributes to the individual being more able to pay for the diagnostic test. On the other hand, if social inequality in the symptoms of the disease were explained only by occupation in the labor market, there would be a concentration among the poorest Brazilians. As shown in Table 4 , the possibility of remote work is much smaller in the portion of the low-income population, making it more difficult for them to maintain social distance, becoming more susceptible to being infected by the virus and developing symptoms of the disease. Assessing the impact of regions, if the only determinants of inequality in the incidence of Covid-19 were to reside in the Northeast and the North, there would be a concentration of the disease among the poor population (based on the results in Table 6 , the two regions have a contribution negative for CI in the diagnosis of the new Coronavirus). As shown in Table 5 below, these two regions are more vulnerable in the sense of having a lower average per capita household income, a higher percentage of beneficiaries for emergency assistance and a lower possibility of remote work. These factors together further aggravate the situation in the North and Northeast regions in the context of the current pandemic. In turn, the impact of living in the North and Northeast regions for inequality in terms of symptoms related to the disease is positive, contributing to a concentration among the richest part of the population. This result may be influenced by the lower educational level observed in these regions (Table 5) . With less education, it is possible that people do not adequately report Covid-19related symptoms (either because of lack of knowledge or because of confusion with flu symptoms), so that there may be an underreporting among the poorest portion of the population. The variables age, gender and race have very discrete effects on the concentration index for the incidence of the new Coronavirus (Table 6) . With regard to symptoms related to the disease, if inequalities were explained only by age and gender (female), there would be a rapid concentration of people with some type of symptom in the poorest part of the population. The residuals in Table 6 show the unexplained sources of the inequalities. A PED contributes to the concentration of Covid-19 in the number of tests realized and presence of correlated symptoms, among people with higher income. This can be noted though the positive sign of the contribution for this variable. In this sense, if socioeconomic inequalities in these pre-existing diseases were eliminated, there would be a reduction in inequalities with respect to the coronavirus. In relative matters, the impact of the PED on the inequality decomposition is significantly higher in the case of the variable related symptoms to 20 .70%, than for the dependent variable Covid-19 diagnosis, just 1.62%. This article presents evidence on socioeconomic inequality in diagnosis and symptoms related to Covid-19 using concentration indices (CI). Additionally, we used the decomposition of the CI by Wagstaff, Doorslaer and Watanabe (2003) , which allows the generation of relevant information on the determining factors. Understanding the nature and main generators of these inequalities during the pandemic period is essential to design and implement effective policies aimed at tackling health disparities. The results found show an accelerated growth in the incidence of the new pathogen during the analyzed period. Furthermore, inequality in the diagnosis of Covid-19 is concentrated among the richest, corroborating empirical evidence in the literature for the first wave of infections in different countries. However, this higher concentration of the disease among the richest may be influenced by the high prices of the tests, making diagnosis difficult in the poorest population. As a way to get around this possible bias, social inequalities in relation to the symptoms of the flu syndrome were analyzed. The results indicated a very low inequality in the two months analyzed, considering the value of the CI very close to zero. Inequalities are mainly explained by race and socioeconomic factors (education, income weighted by the poverty line and employment). If inequalities in Covid-19 symptoms were determined only by race, there would be a concentration among the poorest population. This fact is consistent with the empirical studies reported that reveal the greater probability of poor and black people to die victims of Covid-19. Regarding the regional context, living in the Northeast and the North contributes to the concentration of the Covid-19 diagnosis in the most socioeconomically vulnerable portion of the population, in which it is highly exposed to the risk of contamination by the virus, mainly due to the difficulty in maintaining the social isolation and the drastic fall in income levels. A pre-existing disease contributes to the concentration of Covid-19 in the number of tests realized, but especially for the presence of correlated symptoms, among people with higher income. These conclusions are important in providing relevant information to public managers, as a way to better direct resources to the most vulnerable subgroups of the population. Identifying disproportionately affected population groups can help fill gaps in disparities and guide policy makers in designing and implementing specific interventions to mitigate the negative consequences of the pandemic. Efforts to contain the spread of the disease require the development and implementation of public health interventions that take into account equity and social justice. Even with the beginning of the immunization system, the importance of testing strategies on a large scale in the population for the detection of cases, the tracking of new variants, relaxation of the lockdown, and control of the virus outbreak is highlighted. Additionally, the continuity of social protection measures (emergency assistance) associated with the development of structural policies that promote the mitigation of regional inequalities are essential for combating and preventing future health crises and for tackling structural health determinants, such as poverty and adequate basic sanitation infrastructure, which limit the ability of marginalized populations to fight the disease. When people have the means to meet their basic vital needs, such as food and housing, they will be encouraged to practice healthy behaviors, such as physical distance, wearing masks and hygienic practices that will reduce their vulnerability to the disease. Some limitations of this study are noteworthy. There are different decomposition methodologies, so the results may differ depending on the method used. In addition, due to the lack of mass testing, it is possible that there is an underreporting of Covid-19 cases among the poorest part of the population, harming the results of the work. Future work is essential for monitoring individuals with post-Covid-19 syndrome and determining the long-term consequences of the infection. The authors have no competing interests to declare This research did not receive any specific grant from funding agencies in the public, commercial, or Not-for-profit sectors. **Due to lack of information, we define first wave as the period from July to November 2020. *** Not significant Racial, economic, and health inequality and COVID-19 infection in the United States Ethnic and regional variations in hospital mortality from COVID-19 in Brazil: a cross-sectional observational study. The Lancet Global Health Measuring socioeconomic inequality in obesity: looking beyond the obesity threshold Inequality and the coronavirus: Socioeconomic covariates of behavioral responses and viral outcomes across US counties (No. w27549) Desigualdade salarial entre raças no mercado de trabalho urbano brasileiro: aspectos regionais Para medir as desigualdades raciais no mercado de trabalho Socioeconomic inequalities associated with mortality for COVID-19 in Colombia: a cohort nationwide study Socioeconomic inequalities in the spread of coronavirus-19 in the United States: a examination of the emergence of social inequalities. Social science & medicine, v. 268 Policies and strategies to promote social equity in health. Background document to WHO-Strategy paper for Unmet health care need and income-Related horizontal equity in use of health care during the COVID-19 pandemic Desigualdade econômica e risco de infecção e morte por COVID-19 no Brasil. Revista Brasileira de Epidemiologia Distributional impact of COVID-19: regional inequalities in cases and deaths in Spain during the first wave SARS-CoV-2 antibody prevalence in Brazil: results from two successive nationwide serological household surveys. The Lancet Global Health COVID-19 in Latin America: a humanitarian crisis A framework for testing the equality between the health concentration curve and the 45-degree line Racial, ethnic, and socioeconomic disparities in confirmed COVID-19 cases and deaths in the United States: a county-level analysis as of Applications of Lorenz Curves in Economic Analysis Association of social and economic inequality with coronavirus disease 2019 incidence and mortality across US counties Higher risk of death from COVID-19 in low-income and non-White populations of São Paulo, Brazil Socioeconomic inequalities in COVID-19 in a european urban area: two waves, two patterns Socio-economic inequalities and COVID-19 incidence and mortality in Brazilian children: a nationwide register-based study. Public health, v. 190 Socioeconomic status determines COVID-19 incidence and related mortality in Income-related health inequalities associated with the coronavirus pandemic in South Africa: A decomposition analysis. International journal for equity in health Síntese de indicadores sociais: uma análise das condições de vida da população brasileira Pesquisa Nacional por Amostra de Domicílios: PNAD COVID19 (maio/2020) [Internet]. Brasil: Instituto Brasileiro de Geografia e Estatística Effect of socioeconomic inequalities and vulnerabilities on health-system preparedness and response to COVID-19 in Brazil: a comprehensive analysis. The Lancet Global Health Socio-economic inequality in global incidence and mortality rates from coronavirus disease 2019: an ecological study Impact of social determinants of health on the emerging COVID-19 pandemic in the United States. Frontiers in public health The pandemic of poverty, vulnerability, and COVID-19: evidence from a fuzzy multidimensional analysis of deprivations in Brazil The bounds of the concentration index when the variable of interest is binary, with an application to immunization inequality On decomposing the causes of health sector inequalities with an application to malnutrition inequalities in Vietnam