Pathways that explain racial differences on edentulism among older adults: 2019 Brazil National Health Survey

Abstract This study aimed to evaluate the pathways that explain the association between race/skin color and edentulism in elderly Brazilians. This was a cross-sectional study using data from participants aged 60 years or older from the 2019 Brazilian National Health Survey, a nationally representative population-based sample. Data were obtained by a structured interview and participants were classified as edentulous if they reported having lost all natural teeth. Information on race, socioeconomic level, behavioral aspects, psychosocial aspects, and access to dental care was collected by interviewers using a questionnaire. The pathways between race/skin color and edentulism were analyzed using structural equation modeling. The final sample of the study included 22,357 participants. Most participants were white (51.5%; 95% confidence interval [CI]: 50.3–52.6), and 36.8% (95%CI: 35.7–37.9) were edentulous. Race/skin color was indirectly associated with edentulism via enabling factors. These findings suggest that socioeconomic inequalities are key in explaining racial inequalities in edentulism among Brazilian older adults.


Introduction
Reducing health inequalities and implementing health policies is a social responsibility because chronic diseases negatively impact the quality of life and well-being of older people, especially those who are poorer, disadvantaged, and socially marginalized. 1According to the 2019 Global Burden of Diseases Study, there were approximately 294 million edentulous persons aged 50 years or older wordwide. 2dentulism is associated with an important disease burden, and its consequences include both underweight and overweight 3 and esthetic and functional problems that negatively interfere in people's lives. 46][7] Race/skin color is defined in these studies as a social rather than a biological category, referring to social groups that generally share common characteristics. 8How individuals classify their race/skin color may reflect the way they perceive themselves in front of others and in certain contexts. 9The evidence suggests that the improvement of oral Declaration of Interests: The authors certify that they have no commercial or associative interest that represents a conflict of interest in connection with the manuscript.
][7] More than two decades ago, Andersen and Davidson (1997) proposed a theoretical model in which this association would be explained by primary determinants of oral health that include socioeconomic and psychosocial variables, access to dental care, and oral health behaviors. 10However, the mechanisms underlying the association between race/skin color and edentulism have not been explored.This study aimed to evaluate the pathways that explain the association between race/skin color and edentulism in older Brazilians using Andersen and Davidson's model.The conceptual hypothesis is that non-white older Brazilians have a higher prevalence of edentulism than their counterparts and that this association is mediated by primary determinants, particularly enabling factors, and oral health behaviors.

Methodology Study design and sample
This cross-sectional study analyzed data from the 2019 Brazilian National Health Survey (Pesquisa Nacional de Saúde -PNS), a nationally representative population-based study.The 2019 PNS is the second edition of this survey -the first one was carried out in 2013.The PNS used a cluster sampling plan with three stages: the census tracts in the Primary Sampling Units (UPA, in Portuguese) were the first stage, the household was the second stage, and within each permanent private household, a person aged 15 years or older (randomly selected) was the third stage of the selection of the PNS target population.More information on sampling, including the methods for sample size calculation, can be found in a specific publication on the subject. 11The present study included data from participants aged 60 years or older.

Data collection
Data were collected by a structured interview carried out by trained interviewers at the participants' homes.A household questionnaire and an individual questionnaire were used.
The study outcome, edentulism, was defined as absence of natural teeth. 12Participants were asked about the number of missing teeth and classified as edentulous (coded 1) if they reported they had lost all natural teeth and as dentate (coded 0) if they had lost fewer than 32 teeth (complete dentition).
Self-reported race was recorded through the question about race/skin color: 'What race are you?'.According to the Brazilian Institute of Geography and Statistics, the response options were: White, Brown, Black, Yellow, or Indigenous. 13This variable was dichotomized into 'White' (0) or 'Non-White' (1) (Browns and Blacks).Yellow and Indigenous people were excluded from the sample, as has been done elsewhere. 14The final study sample consisted of 22,357 participants.
Sel f-perceived need for dental ca re was evaluated through oral health perception, which was dichotomized as 'poor' (0) (regular, poor or very poor) and 'good' (1) (very good or good), based on a study that investigated self-rated oral health. 16ral health behaviors were included in the model, as they can influence the association between race/ skin color and edentulism.The variables were tooth brushing and tobacco smoking.The assessment of tooth brushing was based on answers to the question 'How often do you use a toothbrush for oral hygiene?'The variable was categorized into 'Not every day' (0), 'Once a day' (1), 'Twice a day' (2) and 'Three or more times a day' (3).Smoking was categorized into 'No' (0) and 'Yes' (1).Participants who exhibited any frequency of consumption ('yes, daily'; and 'yes, less than daily') were assigned to the 'yes' category.
Dental visit was assessed by the question: 'When was the last time you saw a dentist?' (up to 1 year; more than 1 year to 2 years; more than 2 years to 3 years; more than 3 years; never saw a dentist).This variable was categorized as 'More than 1 year' (0) (for those who did not use dental services in the year prior to the interview) or 'In the last year' (1) (for those who used dental services up to 1 year).

Statistical analysis
Data were analyzed using the STATA 14.0 software (Stata Corporation, College Station, USA) and Mplus version 6.12.All analyses were performed considering the sample weight due to the complex sample.Preliminary analyses were done to describe the characteristics of the sample.The pathways between race/skin color and edentulism were analyzed through structural equation modeling (SEM) and adjusted for sex and age.Participants with missing data were excluded from analysis.SEM was used to obtain a measurement model for a latent variable (enabling factors) and a structural model to estimates the magnitude of the effects among the pathways.The maximum likelihood estimator for complex samples with robust standard error was used.In the first stage, the latent variables were specified separately using confirmatory factor analytic models, considering all standardized factor loadings above 0.3.The second stage involved fitting the path analytic models, which included the latent constructs, to jointly estimate the standardized associations with edentulism.The results are provided as standardized coefficients (SC) and p-values.Fit indices and factorial loads guided the adjustment for the parsimonious model.Fit indices used were: root mean square error of approximation (RMSEA), which should have values ≤ 0.05 with their respective 90% CI; comparative fit index (CFI) with values ≥0.9; Tucker-Lewis index (TLI) also with values ≥0.9; and standardized root mean square residual (SRMR) with values ≤ 0.08. 17

Ethics
The 2019 PNS data are available online for public access and use at the Brazilian Institute of Geography and Statistics (IBGE) official website.The study was approved by the National Research Ethics Commission

Results
Among the participants, 36.8% were edentulous.Table 1 shows the characteristics of the sample, the prevalence of edentulism and the prevalence of sample characteristics of white and non-white individuals.Most participants were white (51.5%) and with a lower prevalence of edentulism.In relation to enabling factors, most individuals had a per capita family income higher than 1 MW and 46.4% had incomplete elementary education.Most participants also had an occupation and did not own a car.Moreover, most of the sample reported good self-perceived oral health, brushed their teeth three or more times a day, were non-smokers, and had not visited the dentist in the year prior to the survey.The prevalence of edentulism was higher among the less privileged individuals, including black individuals (Table 1).
Table 2 shows the standardized estimated effects of indicators in the initial and final structural models.The initial and final models presented good fit values, confirming the theoretical model used.The parameters of the parsimonious model were: RMSEA = 0.008 (90%CI: 0.004-0.011),CFI = 0.998, TLI = 0.994, and SRMR = 0.007 (Table 2).
Table 3 shows the effects of the variables on edentulism in the parsimonious SEM model.Race/ skin color had only an indirect effect (SC: 0.081) on edentulism and the enabling factors pathway (SC: 0.051) was the one that best explained this association.Also, paths considering enabling factors through dental attendance, individual perceived need and tobacco smoking had a weak association with edentulism (Table 3).

Discussion
The hypothesis of this study was accepted.The association between race/skin color and edentulism was mediated predominantly by enabling factors.These findings suggest that racial inequalities related to oral health occur through primary determinants rather than oral health behaviors, but by primary determinants.The manifestation of multiple systems of oppression linked to social structures that emerged throughout Brazilian history, such as political marginalization and economic exploitation of racial minorities, are the major driving forces behind oral health inequalities. 14,18Racial issues are a complex and long-lasting social process that was shaped by slavery, and the reproduction of racial discrimination is a complex, socially patterned phenomenon that impairs dignity, well-being, and health. 18,19he indirect path that best explains the association between race/skin color and edentulism was the path with only enabling factors, which included limited financial resources.The materialist and the psychosocial theories are helpful in understanding how socioeconomic conditions influence health outcomes.The materialist theory implies that socioeconomic position and access to material and structural conditions affect health status.The psychosocial theory suggests that the perceptions of social stat us i n f luence health stat us. 20,21otwithstanding, compared to white people, racial minorities have lower income, lower levels of education, and less purchasing power. 22,23Thus, different enabling factors of racial minorities can generate inequalities in oral health, due to the different availability of economic resources and access to structural conditions, such as access to dental care and the psychosocial response to relative social status.
The findings of this study demonstrated that enabling factors negatively impacted perceived need by mediating the relationship between race/skin color and edentulism.The enabling factors can reflect the emotional and social meanings of inequalities, which can influence behaviors and coping strategies. 24urthermore, health beliefs and perceptions are rooted in social and cultural contexts and the response to the self-rated oral health question may be the product of a multitude of experiences. 5Therefore, some studies have observed a relationship between race and self-reported oral health outcomes, 23,25,26 and ethnic differences in perceived oral health may reflect clinical oral health problems. 5his study also showed that enabling factors impacted dental visits and this also mediated the relationship between race/skin color and edentulism.Some empirical evidence points to racial inequities in access to and quality of oral health care among older people.Data from the 2017 US National Center for Health Statistics indicated that older black people are more likely to have an unmet dental need than white people due to the cost of dental care. 27Another North American study -in which most analyzed individuals were older than 60 years -identified that the probability of African American patients receiving a tooth-preserving treatment (compared to tooth extraction) was lower than among white patients. 28 2003 study with older Brazilians found that 3.8% of white older people reported never having visited a dentist in their lives while this percentage was 7.8% for blacks. 29Our results corroborate these findings, as they indicate that the relationship between race/ skin color and edentulism is partly explained by the lower access to dental care among black people and that the socioeconomic inequalities are the starting The relation between race/skin color and edentulism was also mediated by enabling factors via tobacco smoking (variable from the "Oral Health Behavior" pathway).There is a lack of evidence on the importance  of racial inequalities in oral health behaviors.A study with data from the 2013 PNS found that blacks reported brushing their teeth less frequently. 30Results of the 2013 PNS also showed that the prevalence of tobacco smoking in black Brazilians was the highest among all categories of race/skin color. 31However, both studies were essentially descriptive and did not consider the possible confounding effect of crucial variables, such as income.It is well known that smoking is more prevalent among socioeconomically underprivileged classes.This behavior pattern can be a mechanism to cope with the stress generated by personal difficulties, including economic needs and living in an underserved environment. 32The findings support this association and the link to inequalities in edentulism.Differences of the impact of smoking exposure attributable to race/skin color on edentulism are explained by socioeconomic inequities, since race/skin color does not determine predisposition to unhealthy habits.This study has some limitations referring to data collected in the 2019 PNS.The cross-sectional design limits causal inferences, highlighting the need for prospective studies.In addition, this study was based entirely on self-reported data, meaning that recall bias may have occurred.However, this is not expected to be significant, as self-reported data are considered valid oral health measures. 33Also, populations living on the street and in nursing homes were excluded from the survey; these groups have little or no access to oral health services, which may indicate a worse scenario than that presented in this study.This study has some strengths, including the use of a national high-quality information source and a sample that is representative of Brazilian older adults living in private households.To the authors' knowledge, no study has assessed the pathways that explain the association between race/skin color and edentulism, which should be taken into account when planning and implementing oral health policies and programs.

Conclusion
In conclusion, the findings of this study suggest that the association between race/skin color and edentulism in Brazilian older individuals is mediated by socioeconomic variables.This study contributes to the understanding that racial inequities in oral health are associated with worse socioeconomic position in black older Brazilians.Prevention and rehabilitation actions that address edentulism among older Brazilians should take into account racial differences in the distribution of resources, with priority given to disadvantaged groups.

Figure 1 .
Figure 1.Hypothetical model from the association between race/skin color and edentulism.

Figure 2 .
Figure 2. Pathway analysis of the association between race/skin color and edentulism in Brazilian older adults.

Table 1 .
Sample characteristics and prevalence of edentulism by race/skin color (White and Non-white) in older adults from the 2019 National Health Survey (Pesquisa Nacional de Saúde -PNS) (n = 22,357).

Table 2 .
Standardized estimated effects of indicators in initial and final structural models.Adjusted for sex and age.RMSEA: root mean square error of approximation; CI: confidence interval; CFI: comparative fit index; TLI: Tucker-Lewis index; SRMR: standardized root mean square residual.£: correlated factors.

Table 3 .
Standardized coefficients for effects of race/skin color on edentulism.
Taking into account sample weight; Adjusted for sex and age.