Open-access Intersectionality in the mortality of Brazilians aged 50 years or older: analysis of ELSI-Brasil

Interseccionalidad en la mortalidad entre brasileños de 50 años o más: un análisis del ELSI-Brasil

Abstract

The objective was to analyze the factors associated with all-cause mortality and cause-specific mortality groups, including garbage codes (GC), from an intersectional perspective (operationalized by the combination of sex, race/skin color, education, and income), in a cohort of Brazilians aged 50 years or older. This is a prospective cohort study with data from 8,918 participants of the Brazilian Longitudinal Study of Aging (ELSI-Brasil), followed between 2015 and 2021. Intersectionality was operationalized in 32 strata. Mortality heterogeneity was quantified by the Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA). The association with time until death was analyzed using traditional and frailty Cox proportional hazards models, adjusted for sociodemographic, health, and behavioral variables. There were 817 deaths. The MAIHDA model without adjustment variables indicated that 7% of the mortality variation was attributable to differences between intersectional strata. In the final model adjusted with frailty, the adjustment variables explained part of the risk, but the residual analysis showed that strata composed of women with low education maintained a high residual mortality risk. In contrast, high-income non-white men presented a lower risk than expected. Intersectionality is a structural determinant of mortality. The intersection of axes of oppression generates risks that are not reducible to the sum of their parts, demanding policies that act on the social structures that produce inequity.

Keywords:
Aging; Mortality; Intersectionality


Resumo

O objetivo foi analisar os fatores associados à mortalidade geral e por grupos de causas específicas, incluindo códigos garbage (CG), sob a perspectiva da interseccionalidade (operacionalizada pela combinação de sexo, raça/cor da pele, escolaridade e renda), em uma coorte de brasileiros com 50 anos ou mais. Estudo de coorte prospectivo com dados de 8.918 participantes do Estudo Longitudinal da Saúde dos Idosos Brasileiros (ELSI-Brasil), acompanhados entre 2015 e 2021. A interseccionalidade foi operacionalizada em 32 estratos. A heterogeneidade na mortalidade foi quantificada pela Análise Multinível de Heterogeneidade Individual e Acurácia Discriminatória (MAIHDA). A associação com o tempo até o óbito foi analisada por meio de modelos de riscos proporcionais de Cox tradicional e com fragilidade, ajustados por variáveis sociodemográficas, de saúde e comportamentais. Ocorreram 817 óbitos. O modelo MAIHDA sem variáveis de ajuste indicou que 7% da variação na mortalidade foram atribuíveis às diferenças entre os estratos interseccionais. No modelo final ajustado com fragilidade, as variáveis de ajuste explicaram parte do risco, mas a análise dos resíduos revelou que estratos compostos por mulheres com baixa escolaridade mantiveram um risco de mortalidade residual elevado. Em contraste, homens não brancos de alta renda apresentaram risco inferior ao esperado. A interseccionalidade é um determinante estrutural da mortalidade. A sobreposição de eixos de opressão gera riscos que não são redutíveis à soma de suas partes, demandando políticas que atuem sobre as estruturas sociais que produzem iniquidade.

Palavras-chave:
Envelhecimento; Mortalidade; Interseccionalidade


Resumen

Este estudio tuvo como objetivo analizar los factores asociados a la mortalidad general y la mortalidad por grupos de causas específicas, incluidos los códigos garbage (CG), desde la perspectiva de la interseccionalidad (operacionalizada mediante la combinación de sexo, raza/color de la piel, escolaridad e ingresos), en una cohorte de brasileños de 50 años o más. Estudio de cohorte prospectivo con datos de 8.918 participantes del Estudio Brasileño Longitudinal del Envejecimiento (ELSI-Brasil), seguidos entre 2015 y 2021. La interseccionalidad se operacionalizó en 32 estratos. La heterogeneidad en la mortalidad se cuantificó mediante el Análisis Multinivel de Heterogeneidad Individual y Precisión Discriminatoria (MAIHDA). La asociación con el tiempo hasta la muerte se analizó a través de modelos de riesgos proporcionales de Cox tradicional y con fragilidad, ajustados por variables sociodemográficas, de salud y de comportamiento. Se registraron 817 muertes. El modelo MAIHDA sin variables de ajuste indicó que el 7% de la variación en la mortalidad era atribuible a las diferencias entre los estratos interseccionales. En el modelo final ajustado con fragilidad, las variables de ajuste explicaron parte del riesgo, sin embargo, el análisis de los residuos reveló que los estratos compuestos por mujeres con baja escolaridad mantuvieron un alto riesgo de mortalidad residual. En cambio, los hombres no blancos de altos ingresos presentaron un riesgo inferior al esperado. La interseccionalidad es un determinante estructural de la mortalidad. La superposición de ejes de opresión genera riesgos que no se pueden reducir a la suma de sus partes, lo que exige políticas que actúen sobre las estructuras sociales que producen inequidad.

Palabras-clave:
Envejecimiento; Mortalidad; Interseccionalidad


Introduction

Population aging is a global demographic phenomenon that, in Brazil, is occurring rapidly and in a context of deep and persistent social inequalities, posing unique challenges to public health 1. Current scientific knowledge indicates that the country is experiencing a complex and non-linear epidemiological transition, characterized by a triple burden of diseases: the growing prevalence of chronic non-communicable diseases (NCDs); the persistence of an unresolved agenda of infectious and nutritional diseases; and the maintenance of high rates of mortality from external causes 2,3.

It is known that older adult mortality is not homogeneously distributed, being influenced by a complex interrelation of socioeconomic, demographic, and health factors accumulated throughout life 4. However, a persistent gap in the literature is the tendency to analyze these factors in isolation, which is insufficient to capture the complexity of health inequities. This study is motivated by the following research question: how does the intersection of multiple axes of social oppression (gender, race/skin color, social class) influence all-cause and cause-specific mortality risk in older adults in Brazil?

To answer this question, this study is based on the theoretical model of intersectionality. Named by Kimberlé Crenshaw 5 in 1989, this framework, discussed by Brazilian intellectuals such as Lélia Gonzalez 6 and Sueli Carneiro 7, argues that oppressions do not act additively, but rather reinforce one another, producing inequities that are greater than the sum of their parts.

International studies show that individuals from more recent cohorts, such as baby boomers, present a higher multimorbidity burden than previous generations, with inequalities accentuated by race and gender 8. In Brazil, Black older adults, especially women, are more likely to have worse intrinsic capacity - physical and mental decline - when compared to white men 9, demonstrating how intersectionality influences aging. Another gap that justifies this investigation is the methodological gap: the high prevalence of poorly defined causes of death or garbage codes (GC) in Brazilian official records compromises the reliability of analyses and policy planning, underestimating the disease burden in vulnerable groups 10,11.

Considering the above, the objective of this study was to analyze the factors associated with all-cause mortality and cause-specific mortality groups, including GCs, considering the role of intersectionality in risk inequalities and these associations, based on a nationally representative cohort of Brazilians aged 50 years and older, between 2015 and 2021.

Materials and methods

Study design and population

This is a prospective cohort study that used data from the Brazilian Longitudinal Study of Aging (ELSI-Brasil), a comprehensive investigation into the aging process in the country. ELSI-Brasil employs a complex sampling design, with stratification of municipalities, and selection of census sectors and households to ensure national representativeness. Full details on the ELSI-Brasil methodology were previously published 12.

Participants were recruited at baseline of the study, between 2015 and 2016. The follow-up for the outcome of death was carried out until the second wave of data collection (2019-2021), with the observation period ended on December 31, 2021. Mortality data, including date and underlying cause of death, were obtained through probabilistic linkage of the ELSI-Brasil database with the Brazilian Mortality Information System (SIM, acronym in Portuguese) of the Brazilian Ministry of Health. The linkage procedure used personal identification variables (name, date of birth, and mother’s name) and included a manual review step for handling homonyms and confirming records, following the institutional protocols previously described 4,12. This procedure ensured the reliability of the data and the confidentiality of the participants.

The eligibility criteria for this analysis were: having participated in the ELSI-Brasil baseline and possessing complete information for the variables used in definition the intersectionality construct. The initial sample of ELSI-Brasil was 9,412 participants. After the exclusion of 494 individuals due to missing data in the exposure variables, the final analytical sample was composed of 8,918 participants.

Outcome variables

The primary outcome was time until all-cause death (overall mortality). Follow-up time was calculated in days, from the date of the baseline interview to the date of death or the final follow-up date for censored participants. Additionally, deaths were classified into four mutually exclusive groups for specific cause analyses, based on the 10th revision of the International Classification of Diseases (ICD-10). The first group, “Infectious-parasitic diseases, malnutrition, and reproductive problems”, included chapters I, X, XI, XII, and XIV. The second, “NCDs, neoplasms, and their risk factors”, covered chapters II, III, IV, VI, IX, and XIII. The third group, “External causes and mental/behavioral disorders”, was composed of chapters V and XX. Finally, the fourth, “Garbage Codes (GC)”, including chapter XVIII (Poorly defined causes) and other codes considered of little use for public health, according to the list from the study Global Burden of Disease (GBD) 13.

Intersectional analytical strategy and adjustment variables

The intersectional perspective was adopted as a central analytical strategy, being operationalized by the combination of four social markers to form 32 intersectional strata: biological sex (female, male), self-reported race/skin color (white, non-white), education (categorized into four levels: no formal education, 1 to 4 years, 5 to 10, and 11 years or more), and household income per capita (grouped into “up to BRL 1,190.00” and “greater than BRL 1,190.00”).

The dichotomy of race/skin color into “White” and “non-White” (Brown, Mixed-race, Indigenous, and Yellow) is justified based on the concept of structural racism, which identifies whiteness as the normative reference of privilege 7,14. Although Indigenous persons (n = 42) and Yellow persons (n = 5) have distinct origins, the small sample size would prevent reliable isolated estimates, so they were grouped in the racial subordination pole.

The adjustment variables were selected based on the literature as potential confounders. These included sociodemographic and contextual characteristics, such as birth cohort (grouped into four periods from 1910 to 1966), age (continuous), geographic macrorregion (North, Northeast, South, Central-West, Southeast), place of residence (urban, rural), marital status (unmarried, married), and living alone (yes, no). Factors related to access to and use of healthcare services, such as having health insurance (yes, no) and household coverage by the Family Health Strategy - FHS (yes, no), were also considered.

Finally, health and behavioral condition variables were included: the number of self-reported morbidities (constructed based on the count of diagnoses of hypertension, back problem, high cholesterol, arthritis, depression, diabetes, heart problem, chronic obstructive pulmonary disease - COPD -, stroke, cancer, asthma, kidney failure, Parkinson’s, and Alzheimer’s), the occurrence of falls in the last year (yes, no), the presence of functional limitation in at least one of the basic activities of daily living (BADL) - dressing, eating, bathing, using the toilet, moving around (walking within a room), and getting out of bed -, alcohol use (yes, no), smoking habit (ex-smoker, current smoker, never smoked), and the level of physical activity practice (sufficient - if at least 150 minutes of physical activity per week was performed -; insufficient - if less than 150 minutes of physical activity per week was performed).

Statistical analysis

Initially, we performed a descriptive analysis of sample characteristics and mortality proportions. The associations between the variables and mortality were tested using the chi-square test with Rao-Scott correction 15 and Mood’s median test for continuous age 16, incorporating sample weights and design effects. In addition, the Bayesian information criterion (BIC) was calculated to rank the significance of the variables and to determine the order in which they were included in the multiple models 17. To quantify heterogeneity in mortality attributable to social position, we adjusted a null model of Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA), i.e., without adjustment variables. This is a multilevel logistic regression model in which individuals (level 1) are nested within their respective intersectional strata (level 2). The intraclass correlation coefficient (ICC) was calculated based on this model to estimate the proportion of total variation in mortality attributable to differences between strata 18.

To analyze time-to-event data, Cox proportional hazards models were used, with measures of association estimated via hazard ratios (HR) and their respective 95% confidence intervals (95%CI). The central strategy of the analysis was the use of Cox frailty models were used, with measures of association estimated via frailty hazard ratios (FHR). In this statistical technique, “frailty” refers to a random effect assigned to the construct of intersectionality, formed by the strata of social position (combination of sex, race/skin color, education, and income), in order to capture the unobserved heterogeneity that influences mortality risk 18. This approach extends the traditional Cox model and provides a more precise estimation of the effects of covariates and the contextual effect of intersectionality 19.

The modeling strategy followed three steps: (1) univariate Cox analysis for all adjustment variables; (2) selection of variables with p < 0.20 for inclusion in the multiple model; (3) adjustment of the Cox model with final multiple frailty, considering a significance level of 5% (p < 0.05). The assumption of proportional hazards was verified graphically and through the Schoenfeld residuals analysis. The variable “physical activity practice” violated this assumption and was therefore used to stratify the final overall mortality model. The same analytical strategy was applied to each of the four groups of causes of death. However, for analyses by groups of specific causes, standard Cox models were used. The analyses were performed in R software (http://www.r-project.org), version 4.5.2. ELSI-Brasil was approved by the Research Ethics Committee of the Oswaldo Cruz Foundation, Minas Gerais (CAAE 34649814.3.0000.5091).

Results

The final sample of 8,918 participants had a median follow-up period of 4.8 years (interquartile range: 4.1-5.4 years). During follow-up, there were 817 deaths. The distribution of causes of death showed that chronic NCDs and neoplasms (group 2) were responsible for 41.8% of deaths, followed by GC (33%), infectious and parasitic diseases (group 1) with 21.2%, and external causes (group 3) with 4.1%. The application of the MAIHDA model without adjustment variables, which assesses mortality variability solely among intersectional strata, showed that 7% (ICC = 0.07) of the total variation in the risk of death was attributable to differences among the 32 groups. This result indicates moderate but significant heterogeneity, suggesting that social position is a relevant predictor of mortality.

The univariate analysis (Table 1) showed a social gradient, in which factors such as male sex, low education, advanced age, not being married, having functional limitation, a higher number of morbidities, smoking, and insufficient physical activity were associated with a higher risk of death. However, the comparison between the standard Cox model (HR) and the Cox model with frailty (FHR), which considers intersectional heterogeneity, showed important nuances. Contextual variables, such as living in a rural area or in the North macrorregion, did not show a significant association in the standard model, but presented a statistically significant association as protective factors in the frailty model (FHR = 0.82; 95% confidence interval - 95%CI: 0.72-0.93 for rural area), although with overlap of the confidence intervals between the estimates (HR and FHR). Conversely, coverage by the FHS, observed as a protective factor in the standard model, lost its statistical significance after adjustment for intersectional frailty.

Table 1
Incidences and hazard ratios (HR) * from Cox proportional hazards and frailty models (FHR) * estimated for all-cause mortality, according to exposure and adjustment variables. Brazilian Longitudinal Study of Aging (ELSI-Brasil, 2015-2021).

The analysis of residual effects for the multiple frailty model provided the quantification of residual variability in overall mortality risk among strata after controlling for adjustment factors (Table 2). The variance of the random effect (σ² = 0.08; p < 0.001) showed the persistence of groups with mortality risk above and below the sample average during the follow-up period, regardless of health conditions and individual behaviors. The strata that presented a significantly lower than expected mortality risk were: non-white men, high income, 5 to 10 years of education; and non-White men, high income, no formal education. In contrast, strata with risk significantly above the expected average included women with 1 to 4 years of education: non-White, low income; white, high income; and non-White, high income. Advanced age (FHR = 1.06 per year; 95%CI: 1.05-1.07), having functional limitation in BADLs (FHR = 1.99; 95%CI: 1.69-2.34), unmarried individuals (FHR = 1.48; 95%CI: 1.26-1.73), and being an ex-smoker (FHR = 1.36; 95%CI: 1.15-1.61) or current smoker (FHR = 1.93; 95%CI: 1.57-2.39) remained as independent risk factors. The absence of morbidities was a protective factor (FHR = 0.71; 95%CI: 0.55-0.91).

Table 2
Estimates of risk of all-cause mortality according to selected variables: Cox model with frailty (FHR) * adjusted for the construct of intersectionality. Brazilian Longitudinal Study of Aging (ELSI-Brasil, 2015-2021).

The univariate analysis by groups of causes of death showed distinct patterns. For group 1 (infectious diseases), older birth cohorts (HR = 9.27 for 1910-1939), living in the Northeast (HR = 1.39) or Central-West (HR = 1.56), not being married (HR = 2.02), and having functional limitation (HR = 2.95) were strong risk predictors (Table 3). For group 2 (chronic NCDs), in addition to social gradient by education, functional limitation (HR = 2.74) and smoking (HR = 1.57 for smokers) were prominent. In group 3 (external causes), the male sex, residing in rural areas (HR = 2.47), and especially smoking (HR = 4.14 for smokers) were prominent as the main risk factors. Finally, for deaths from GCs, functional limitation presented the highest risk among all those analyzed (HR = 4.07), along with advanced age and low education, indicating a profile of clinical and social vulnerability associated with the inaccuracy of the diagnosis of the cause of death.

Table 3
Mortality incidence and hazard ratios (HR) * from Cox proportional hazards and frailty models (FHR) * by grouping of the 10th revision of the International Classification of Diseases (ICD-10) and classification as garbage code (GC), according to exposure and adjustment variables. Brazilian Longitudinal Study of Aging (ELSI-Brasil, 2015-2021).

In the final model adjusted by cause groups (Table 4), age was a consistent risk factor for groups 1 (infectious diseases) and GC. Not being married increased the risk of death from group 1 causes and, more sharply, from external causes (group 3) (FHR = 2.14). Functional limitation was associated with mortality from groups 1 and 2, but not from group 3. Smoking emerged as a transversal risk factor, significantly increasing all-cause mortality, noting a more than tripled risk of death from external causes (group 3) for both former smokers (FHR = 3.29) and current smokers (FHR = 3.65). Furthermore, residing in rural areas was associated with a 2.5 times higher risk of death from external causes, and the Central-West macro-region presented a significantly lower risk of registration of death from GC compared to the Southeast.

Table 4
Fragility hazard ratio (FHR) * and their respective 95% confidence intervals (95%CI) from Cox proportional hazards models adjusted by grouping of the 10th revision of the International Classification of Diseases (ICD-10) and classification as garbage code (GC). Brazilian Longitudinal Study of Aging (ELSI-Brasil, 2015-2021).

Discussion

This study showed that social position, constructed at the intersection of multiple axes of oppression, is a fundamental determinant of mortality in older adults in Brazil. The application of an analytical approach that integrates intersectionality theory with survival models enabled quantifying how mortality risks are unequally distributed, producing vulnerabilities that cannot be understood by the sum of their components in isolation. Therefore, this work aligns with a political act of restoring denied humanities 20, making visible, through data, the realities that the hegemonic narrative has historically silenced 14,21. The use of whiteness as a reference group enables highlighting how structural racism hierarchizes the risk of death, establishing a universal reference of protection and privilege 14,22.

In parallel, it is important to note that the potential confounding effect of the COVID-19 pandemic was considered. However, preliminary analyses did not indicate a significant impact on the results, as the vast majority of observed deaths (over 98%) occurred in the pre-pandemic period, ensuring the reliability of the findings in relation to mortality patterns established before the health crisis. Thus, the proportional distribution of deaths showed an incomplete epidemiological transition, with chronic NCDs and neoplasms accounting for the largest share (41.8%), but with a significant persistence of infectious-parasitic diseases (21.2%) and external causes (4.1%), consistently with other analyses of the national situation 23,24.

The MAIHDA model without adjustment variables showed that 7% of the total variation in the probability of death was attributable to differences between the intersectional strata. In Public Health studies, such between-group variation of this magnitude indicates that social position is not a secondary risk factor, but a central predictor of life/death outcomes 25. The persistence of high residual risks for women with low education, even after adjusting for morbidities, reinforces the hypothesis of the “matriarchy of misery” 7, in which the inscription of inequality on the body manifests through the confluence of oppressions of gender and class. On the other hand, the paradoxical finding of protection in non-white men with intermediate education suggests a critical threshold at which the accumulation of social capital enables a more effective navigation in an often hostile healthcare system 26.

However, the interpretation of these findings, especially those related to race/skin color, requires caution as to the survival bias. By establishing the initial follow-up point at 50 years of age, the research does not consider the early mortality that predominantly affects non-White individuals 27. From the perspective of necropolitics, structural racism operates as a mechanism of exclusion that subtracts a considerable portion of the non-White race/skin color population even before the data collection is initiated. This can, consequently, attenuate the magnitude of health inequities in the outcomes 14,22.

The estimates of the Cox models confirm a documented social gradient of mortality 28,29, but the analysis by specific causes provides a granular view of how inequities are inscribed in the body. Male overmortality, for example, was imperatively manifested in external causes. This finding reflects not only exposure to occupational risks and violence, but also a social construction of masculinity that neglects self-care and prevention 30. Similarly, the accentuated risk of death from infectious diseases in the oldest birth cohort (1910-1939) is the empirical translation of the Life Course Epidemiology, reflecting a history of policies that neglected sanitation and nutrition for the majority of the population in past decades 31,32.

Physical frailty, represented by functional limitation in BADLs, emerged as a potent predictor of mortality, almost doubling the risk of death in the adjusted model. This outcome is not purely biological, but a product of life trajectories marked by inequity, being more common in groups with lower education and income, and aggravated by failures of the healthcare system in providing access to rehabilitation therapies 33,34. The strong association of functional limitation with mortality classified as GCs also suggests that the clinical complexity of frail individuals challenges the healthcare system’s capacity to assign a basic cause of death, resulting in unspecific classification 35.

The analysis of lifestyle variables shows the association between individual behaviors and structural health determinants. Smoking has been shown to be a persistent risk factor, with current smokers having an almost doubled risk of all-cause mortality. The attenuation of this effect by integrating the frailty model, however, indicates that a portion of the risk attributed to smoking is, in reality, mediated by the social vulnerability inherent in groups in which tobacco consumption is more prevalent 36.

The comparison between the standard and frailty Cox models demonstrated the limit of sectoral policies in the face of the strength of structural inequalities. The protective effect of the FHS, for example, was nullified in the frailty model. This result does not invalidate the merits of the FHS, but suggests that, for bodies burdened by multiple oppressions, access to primary health care may not be sufficient for overcoming the accumulated vulnerability 37,38. Contextual analysis also showed multiple “geographies of abandonment”, such as the 2.5 times higher risk of death from external causes in rural areas, challenging the narrative that concentrates violence in urban centers and indicating the higher lethality of events in areas with precarious health and transportation infrastructure 39,40.

However, it is emphasized that the observed differences between the standard Cox model and the frailty model suggest that the intersectional structure may absorb part of the inequalities previously attributed to contextual factors (such as area and macroregion of residence) that only emerge when controlling for the shared heterogeneity of the strata. However, the overlap of confidence intervals in the direct comparison between the different risk estimates for variables such as FHS and the North macrorregion demands caution, indicating that, although the inclusion of frailty alters statistical significance, the absolute magnitude of this shift should be interpreted from the perspective of accumulated structural vulnerability.

One of the findings with vast political implications is the alarming proportion of 33% of deaths classified as GC, corroborating previous analyses 10. This magnitude reveals a facet of data inequity that sustains structures of necropolitics 22, obscuring the true distribution of causes of death and leading to ineffective policies 11,41. The fact that the Central-West macrorregion presents a significantly lower risk of GC registration compared to the Southeast, even after adjustment, shows that improving data quality is a feasible political action, which depends on prioritizing strategies such as strengthening Death Verification Services and surveillance 42,43.

The strengths of the study include the use of a population-based cohort with national representativeness and the pioneering application of an analytical approach that integrates intersectionality theory with survival models. However, limitations should be considered. The main limitation is survival bias: by initiating follow-up at age 50 years, the cohort excludes the early mortality that disproportionately affects the non-white population, which likely underestimates the impact of structural racism 44. Other limitations include the use of self-reported data and the broad scope of the causes of death categories, which may conceal internal heterogeneities.

Final considerations

The intersectional matrix built on sex, race/skin color, education, and income is a fundamental determinant of mortality in older adults in Brazil. The intersection of axes of oppression generates risks of death that are not reducible to the sum of their parts. This study contributes to knowledge by demonstrating, through a robust quantitative approach, that mortality inequalities are structural and persist even after controlling for individual and behavioral risk factors. The results reinforce the need for intersectoral public policies that act on the social structures that produce and perpetuate health inequities, going beyond interventions focused solely on the individual. By naming and measuring the impact of intersectionality, this research contributes to the construction of a critical, feminist, and anti-racist epidemiology, which not only describes the world but provides the tools to transform it.

Data availability

The sources of information used in the study are indicated in the body of the article.

Acknowledgments

To Brazilian Coordination for the Improvement of Higher Education Personnel (CAPES) e ao Graduate Program in Public Health at the René Rachou Institute (IRR-Fiocruz Minas) for the financial supportpelo apoio financeiro.

Reviewers who authorize identification:

Daniela de Assumpção (0000-0003-1813-996X)

Referências

  • 1 Abud T, Kounidas G, Martin KR, Werth M, Cooper K, Myint PK. Determinants of healthy ageing: a systematic review of contemporary literature. Aging Clin Exp Res 2022; 34:1215-23.
  • 2 Martins TCF, Silva JHCM, Máximo GC, Guimarães RM. Transição da morbimortalidade no Brasil: um desafio aos 30 anos de SUS. Ciênc Saúde Colet 2021; 26:4483-96.
  • 3 Mendes EV. As redes de atenção à saúde. Ciênc Saúde Colet 2010; 15:2297-305.
  • 4 Macinko J, Beltrán-Sánchez H, Mambrini JVM, Lima-Costa MF. Socioeconomic, disease burden, physical functioning, psychosocial, and environmental factors associated with mortality among older adults: the Brazilian Longitudinal Study of Ageing (ELSI-Brazil). J Aging Health 2024; 36:25-34.
  • 5 Crenshaw K. Demarginalizing the intersection of race and sex: a black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. In: Maschke KJ, editor. Feminist legal theories. New York: Routledge; 2013. p. 23-51.
  • 6 Gonzalez L. Racismo e sexismo na cultura brasileira. Revista Ciências Sociais Hoje 1984; 2:223-44.
  • 7 Carneiro S. Escritos de uma vida. São Paulo: Pólen Editorial; 2019.
  • 8 Bishop N, Haas S, Quiñones A. Emerging cohort disparities in multimorbidity in US adults entering early older adulthood. Innov Aging 2023; 7 Suppl 1:231.
  • 9 Plácido J, Marinho V, Ferreira JV, Teixeira IA, Castro-Costa E, Deslandes AC. Association among race/color, gender, and intrinsic capacity: results from the ELSI-Brazil study. Rev Saúde Pública 2023; 57:29.
  • 10 Machado EL, Mambrini JVM, Silva TC, Braga LS, França GVA, Lima-Costa MF. Determinantes de Códigos Garbage como causa da morte em coorte nacional representativa de adultos mais velhos: ELSI-Brasil. Ciênc Saúde Colet 2025; 30:e18572023.
  • 11 Naghavi M, Richards N, Chowdhury H, Eynstone-Hinkins J, França E, Hegnauer M, et al. Improving the quality of cause of death data for public health policy: are all 'garbage' codes equally problematic? BMC Med 2020; 18:55.
  • 12 Lima-Costa MF, Andrade FB, Souza PRB, Neri AL, Duarte YAO, Castro-Costa E, et al. The Brazilian longitudinal study of aging (ELSI-Brazil): objectives and design. Am J Epidemiol 2018; 187:1345-53.
  • 13 Johnson SC, Cunningham M, Dippenaar IN, Sharara F, Wool EE, Agesa KM, et al. Public health utility of cause of death data: applying empirical algorithms to improve data quality. BMC Med Inform Decis Mak 2021; 21:175.
  • 14 Almeida S, Ribeiro D. Racismo estrutural. São Paulo: Pólen Editorial; 2019. (Feminismos Plurais).
  • 15 Rao JNK, Scott AJ. On chi-squared tests for multiway contingency tables with cell proportions estimated from survey data. Ann Stat 1984; 12:46-60.
  • 16 Lumley T, Scott AJ. Two-sample rank tests under complex sampling. Biometrika 2013; 100:831-42.
  • 17 Sakamoto Y, Ishiguro M, Kitagawa G. Akaike information criterion statistics. Dordrecht: D. Reidel; 1986.
  • 18 Evans CR, Leckie G, Subramanian SV, Bell A, Merlo J. A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA). SSM Popul Health 2024; 26:101664.
  • 19 Carvalho MS, Andreozzi VL, Codeço CT, Campos DP, Barbosa MTS, Shimakura SE. Análise de sobrevivência: teoria e aplicações em saúde. Rio de Janeiro: Editora Fiocruz; 2021.
  • 20 Xavier G. Catálogo intelectuais negras visíveis. Rio de Janeiro: Malê Edições; 2025.
  • 21 Gonzalez L, Hasenbalg C. Lugar de negro. São Paulo: Editora Schwarcz/Companhia das Letras; 2022.
  • 22 Mbembe A. Necropolítica: biopoder, soberania, estado de exceção, política da morte. São Paulo: N-1 Edições; 2017.
  • 23 Azeredo Passos VM, Champs APS, Teixeira R, Lima-Costa MFF, Kirkwood R, Veras R, et al. The burden of disease among Brazilian older adults and the challenge for health policies: results of the Global Burden of Disease Study 2017. Popul Health Metr 2020; 18 Suppl 1:14.
  • 24 Ferrari AJ, Santomauro DF, Aali A, Abate YH, Abbafati C, Abbastabar H, et al. Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet 2024; 403:2133-61.
  • 25 Diez Roux AV. Social epidemiology: past, present, and future. Annu Rev Public Health 2022; 43:79-98.
  • 26 Farina MP, Klopack ET, Andrade FCD. The "long arm of childhood" on mortality in older adulthood in the United States and Brazil: examining the role of educational attainment and differences by gender. Am J Epidemiol 2026; 195:424-33.
  • 27 Malta DC, Minayo MCS, Cardoso LSM, Veloso GA, Teixeira RA, Pinto IV, et al. Mortality among Brazilian adolescents and young adults between 1990 to 2019: an analysis of the Global Burden of Disease study. Ciênc Saúde Colet 2021; 26:4069-86.
  • 28 Szwarcwald CL, Almeida WS, Boccolini CS, Soares Filho AM, Malta DC. The unequal impact of the pandemic at subnational levels and educational attainment-related inequalities in COVID-19 mortality, Brazil, 2020-2021. Public Health 2024; 231:39-46.
  • 29 Silva Filho AM, Araújo EM, Souza IM, Luiz OC, Máximo G, Queiroz FA, et al. Anos potenciais de vida perdidos devido à COVID-19, segundo a raça/cor e gênero, no Brasil, entre 2020 e 2021. Ciênc Saúde Colet 2024; 29:e04702023.
  • 30 Santos JAF. Interações de classe social, raça e gênero na saúde no Brasil. Sociologia & Antropologia 2025; 15:e240021.
  • 31 Slopen N, Shonkoff JP. Employing life course epidemiology to catalyse larger impacts of early childhood policies on lifelong health. In: Kuh D, Susser E, Blodgett JM, Ben-Shlomo Y, editors. A life course approach to the epidemiology of chronic diseases and ageing. 3rd Ed. Oxford: Oxford University Press; 2025. p. 335-52.
  • 32 Wagner C, Carmeli C, Jackisch J, Kivimäki M, Van Der Linden BWA, Cullati S, et al. Life course epidemiology and public health. Lancet Public Health 2024; 9:e261-9.
  • 33 Boccolini PMM, Duarte CMR, Boccolini CS. Functional limitations and the use of rehabilitation therapies among brazilian adults with chronic diseases. Acta Fisiátr 2024; 31:102-10.
  • 34 Vieira MML, Borges VS, Oliveira EJP, Bof de Andrade F. Functional limitation in the older Brazilian adults: association with multimorbidity and socioeconomic conditions. PLOS One 2023; 18:e0294935.
  • 35 Toyoshima K, Araki A, Tamura Y, Ishikawa J, Iwakiri R, Takei T, et al. Prognostic value of the electronic Multimorbidity Frailty Index for mortality, change in basic activities of daily living, length of hospital stay and discharge home in older hospitalized patients. Geriatr Gerontol Int 2025; 25:1185-93.
  • 36 Malta DC, Gomes CS, Andrade FMD, Vasconcelos NM, Prates EJS, Pereira CA, et al. Smoking in Brazil: insights from the results of the household surveys. REME Rev Min Enferm 2023; 27:e-1518.
  • 37 Dias BA. Impact of the Family Health Strategy on reducing infant mortality in Brazilian rural communities: advanced analysis and future proposals. Int J Adv Res (Indore) 2024; 12:1295-9.
  • 38 Silva Júnior VB, Melo KAF, Freitas ASSL, Lima JPM, Higa F, Brito RCT, et al. Atenção primária à saúde: implicações para a promoção da equidade no SUS. Lumen Et Virtus 2024; 15:8262-75.
  • 39 Baesso MM, Regazzo JR, Soares CC, Zanolli IM, Zuin LFS, Tech ARB. A segurança como aliada na produtividade: avaliação de riscos de acidentes com máquinas agrícolas e a importância da prevenção. Brazilian Journal of Animal and Environmental Research 2024; 7:680-94.
  • 40 Bichofe AR, Jesus BLN, Brito FAE, Pereira JMC, Hoffmann N, Pereira SF, et al. Impacto do tempo de resposta no serviço de trauma agudo pré-hospitalar. Revista ft 2024; 29:21-2.
  • 41 Iburg KM, Mikkelsen L, Adair T, Lopez AD. Are cause of death data fit for purpose? evidence from 20 countries at different levels of socio-economic development. PLOS One 2020; 15:e0237539.
  • 42 Lima RB, Frederes A, Marinho MF, Cunha CC, Adair T, França EB. Investigation of garbage code deaths to improve the quality of cause-of-death in Brazil: results from a pilot study. Rev Bras Epidemiol 2019; 22 Suppl 3:e19004.supl.3.
  • 43 Gomes GMT, Braz INA, Jordão I, Jordão I. Levantamento de casos de morte sem assistência médica equivocadamente encaminhados ao IML de uma cidade do sul fluminense. Cadernos UniFOA 2024; 19:1-9.
  • 44 Wrigley-Field E. Three ways of looking at black-white mortality differences in the United States. Annu Rev Sociol 2025; 51:311-33.

Correspondence

L. Oliveira-Santos Instituto René Rachou, Fundação Oswaldo Cruz. Av. Augusto de Lima 1715, Belo Horizonte, MG 30190-002, Brasil luizaos.enf@gmail.com

Associate Editor

Evaluation coordinator: Priscila Maria Stolses Bergamo Francisco (0000-0001-7361-9961)

Publication Dates

  • Publication in this collection
    12 Oct 2026
  • Date of issue
    2026

History

  • Received
    18 Nov 2025
  • Reviewed
    20 May 2026
  • Accepted
    26 May 2026
location_on
Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz Rua Leopoldo Bulhões, 1480 , 21041-210 Rio de Janeiro RJ Brazil, Tel.:+55 21 2598-2511, Fax: +55 21 2598-2737 / +55 21 2598-2514 - Rio de Janeiro - RJ - Brazil
E-mail: cadernos@ensp.fiocruz.br
rss_feed Acompañe los números de esta revista en su lector de RSS
Ir para arriba