Household food insecurity, dental caries and oral-health-related quality of life in Brazilian Indigenous adults

Resumo Insegurança alimentar é um fenômeno complexo que afeta a saúde e o bem-estar de famílias vulneráveis. Este estudo objetivou investigar a relação entre insegurança alimentar domiciliar, cárie dentária, qualidade de vida relacionada à saúde bucal, e determinantes sociais de saúde entre indígenas adultos. Foi conduzido um estudo transversal com adultos Kaingang entre 35-44 anos da Terra Indígena Guarita, Brasil. Insegurança alimentar foi avaliada pela escala EBIA. Cárie dentária foi avaliada pelo índice CPOD. Participantes responderam o questionário OHIP-14 e uma entrevista estruturada. Foram realizadas análises descritivas e multivariadas usando modelos de regressão de Poisson. A amostra incluiu 107 adultos Kaingang de 97 domicílios. Aproximadamente 95% dos participantes viviam em famílias com insegurança alimentar. Insegurança alimentar grave esteve presente em 58% dos domicílios. O fenômeno foi associado ao Bolsa Família, densidade familiar e percepção dos impactos da saúde bucal na qualidade de vida. O alto número de famílias afetadas pela insegurança alimentar revela a vulnerabilidade social do povo Kaingang. Insegurança alimentar em adultos Kaingang está associada à percepção da saúde bucal e determinantes sociais da saúde. Palavras-chave Privação de alimentos, População indígenas, Segurança alimentar, Cárie dentária, Qualidade de vida relacionada à saúde bucal Abstract Food insecurity is a complex phenomenon that affects the health and wellbeing of vulnerable families. The aim of this study was to investigate the relationship between household food insecurity, dental caries, oral health-related quality of life, and social determinants of health among Indigenous adults. A cross-sectional study was conducted among Kaingang adults aged 3544 years old from the Guarita Indigenous Land, Brazil. Food insecurity was assessed through the EBIA scale. Dental caries was assessed using the DMFT index. Participants answered the OHIP14 questionnaire and a structured interview. Descriptive and multivariate analyzes using Poisson regression models were performed. The final sample included 107 adults from 97 households. Approximately 95% lived in food insecure families. Severe food insecurity was present in 58% of the households. The phenomenon was associated to the Bolsa Família benefit, household size, and greater perception of oral health impacts on quality of life. The high number of families affected by food insecurity reveals the social vulnerability of the Kaingang people. Food insecurity in Kaingangs adults is associated to oral health perception and social determinants of health.


Introduction
Food access is essential to human life and consequently, food security is considered an important social determinant of health 1 . Food and Agriculture Organization of the United Nations 2 defines food security as a "situation that exists when all people, at all times, have physical, social and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life" 2 . Based on this definition, four dimensions of food security can be identified: food availability, economic and physical access to food, food utilization and stability over time. Food insecurity is a phenomenon that has been increasing worldwide in recent years and affects the most vulnerable populations, such as Indigenous peoples 2 .
The guarantee of the human right to adequate food and freedom from hunger incorporates aspects of food security and food sovereignty, the fight against hunger, poverty, and social inequalities, and the realization of other human rights 3,4 . Despite international efforts to meet the Millennium Development Goal of eradicating extreme poverty and hunger, the first increase in the world hunger was registered in 2017 after more than a decade of continuous decline 2,5 . In Brazil, social protection policies conducted in recent decades allowed the country to be withdrawn in 2014 from the so-called Hunger Map, the list of nations presenting more than 5% of their population with calorie intake deficit 6 . Nevertheless, the prospect of Brazil's return to the Hunger Map due to austerity measures challenges the solidity of the advances accomplished globally and in the Latin America 7 . Vulnerable subgroups and Indigenous peoples seem particularly sensitive to these setbacks.
Multiple determinants of historical, social, economic and environmental order affect the food security status of Indigenous peoples 8 . The increasing erosion of Indigenous food systems, associated with the decline of biodiversity and the reduction of traditional territories, threatens the existence of alternative models of ethnodevelopment and Indigenous food sovereignty 9,10 . The food transition observed in many Indigenous populations has dramatically affected sociodiversity, food security levels, and health conditions [11][12][13][14][15][16][17] .
The relationship between food security, food sovereignty and health equity has a particularly relevant role for Indigenous communities due to the cultural importance of food practices and the high burden of diseases attributed to changes in food habits 18 . Household food insecurity in Indigenous communities has been associated with increased frequency of smoking, reduced sense of belonging to the community, occurrence of mental disorders, chronic diseases, child growth deficit, and recent hospitalization 19,20 .
Studies conducted in Brazil, Canada and the United States with non-Indigenous populations have observed that individuals in situation of food insecurity present higher frequencies of untreated dental caries, use of dental prostheses, history of dental pain, and negative self-perception of oral health [21][22][23][24][25] . So far, the association between food insecurity and oral health in Indigenous populations has not been investigated, nor between household food insecurity and oral health-related quality of life (OHRQoL). In addition, exploring the relationship between dental outcomes and food security promotes a more comprehensive perspective of the oral health of Indigenous peoples, in line with their holistic understanding of health.
The aim of this study was to investigate the relationship between household food insecurity, dental caries, oral health-related quality of life, and social determinants of health among Kaingang Indigenous adults. We hypothesize that Indigenous adults living in food insecure households present worse oral health status due to social vulnerability. methods This is an exploratory, population-based, cross-sectional study which is part of a larger mixed-methods project conducted between January and August 2017 with Kaingang Indigenous adults in the South of Brazil.

study design
In order to obtain a representative sample and identify eligible participants, a census-based sampling strategy was employed. Selection criteria adopted was individuals ethnically self-identified as Kaingang, aged from 35 to 44 years old, living in all villages of the Guarita Indigenous Land. The World Health Organization recommends targeting the 35-44 age bracket to assess the oral health status of adults 26 . All households were visited and individuals who met the inclusion criteria were invited to take part in the research by signing a written informed consent form. Individuals who lacked capacity to give informed consent were excluded. In addition to their native language, all participants were Portuguese speakers. The 35-44 years old local population is estimated at approximately 300 individuals.
Household Food Security (HFS) was assessed through the Brazilian Food Insecurity Scale (EBIA), a Portuguese version validated for the Brazilian population based on an instrument created by the United States Department of Agriculture 27 . A focus group was conducted with 9 participants previously selected by the Indigenous leader to analyze whether the Kaingang community understand the concepts assessed by the EBIA instrument. In order to reflect the diversity within the Kaingang population, researchers asked the Indigenous leaders to invite a heterogeneous group of locals, comprising young adults, midlife adults and elders of both genders involved in different occupations and social roles (such as agriculture, food preparation, handcrafting, and health workers). The EBIA scale comprises a 14item questionnaire on the family experience of inadequate access to food during the previous three months. Affirmative responses for each question were recorded as 1 point, resulting in a score ranging from 0 to 14 (Chart 1). In cases that more than one participant from the same household was included in the study, only the participant involved in the food preparation answered the EBIA questionnaire. The resulting score was then attributed to the respective household and, consequently, to the other family members.
Oral examinations were performed by a single examiner (GHS) following the World Health Organization guidelines to oral health surveys 26 . Prior to the study, the examiner calibration process comprised 20 hours of theoretical discussions, clinical examinations, and calibration exercises. A sub-sample of 10 participants was reexamined, with an interval of 5 to 7 days between the two assessments, in order to obtain the intra-examiner agreement coefficient (Kap-pa=0.817). Conditions of dental crowns were assessed through the decayed, missing or filled teeth index (DMFT). Participants also answered the validated Brazilian short version of the Oral Health Impact Profile (OHIP-14) questionnaire, which is intended to assess the OHRQoL, i.e., the impacts of oral health conditions on the individual's quality of life 28 . A structured interview was used to collect sociodemographic data related to social determinants of health. The OHIP-14 questionnaire and the sociodemographic interview were administered individually to each participant by the researcher. Data was processed and analyzed by an experienced biostatistician who was blind to the identity of the participants (EMC).

variables
Participants were classified regarding whether they live in a household which receives the Bolsa Família benefit ("beneficiary" or "non-beneficiary"). The Bolsa Família program is a federal policy created in 2003 aiming to combat poverty, improve health outcomes, and increase food security through direct income transfers to families in situation of social vulnerability 29 .
Village location was classified as Tenente Portela or Redentora/Erval Seco. These subareas belong to different health regions within the public healthcare network, which may interfere in the access and utilization of health services. For instance, while Indigenous people living in the villages located in Tenente Portela have access to a local Center of Dental Specialties (Centro de Especialidades Odontológicas), the Indigenous communities from Redentora and Erval Seco are referred for specialized dental care in another Center located more than 100 kilometers away from the Indigenous Land.
Oral health clinical outcomes, household size, age, and years of formal schooling were dichotomized based on the mean values observed. OHRQoL was the only factor analyzed as a continuous variable (score from 0 to 56). Household monthly income was classified based on half of the 2017 Brazilian minimum wage converted in US Dollars (mean exchange rate in August 2017: R$3.15=US$1.00).
The dependent variable "food insecurity" was analyzed both quantitatively, through the final score obtained by the EBIA questionnaire, and adopting severe food insecurity as a cut-off point. For the latter, participants were classified either as living in households "not severe food insecure" (scores from 0 to 9), or in households "severe food insecure" (scores from 10 to 14). Since the classification of the food insecurity level for families with no children is different (range from 0 to 8), households with no children were classified as "not severe food insecure" when presenting scores from 0 to 6, and as "severe food insecure" when presenting scores from 7 to 8. This approach follows the EBIA classification system (Chart 1).

statistical analysis
Prevalence rates were calculated for each level of food security/insecurity. Bivariate analyzes were performed using the Mann-Whitney U test, Spearman's correlation, and robust Poisson regression, according to the nature of the outcome and the associated factor. Poisson regression models with robust variance were used for multivariate analyzes. Associations were considered statistically significant at the 5% probability level. All analyzes were performed using the STATA 14.0 software (StataCorp, College Station, Texas, USA).

ethical standards
The research project was previously presented and approved by the local indigenous leaders. All subjects gave their informed consent for inclusion before they participated in the study.
This study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the Dentistry School of the University of São Paulo and by the Brazilian National Research Ethics Committee.

Results
The focus group discussion revealed that Kaingang adults understand the main concepts related to food insecurity assessed by the Household Food Insecurity Scale (EBIA). Few minor modifications were applied to the questionnaire, such as replacing "quality food" for "good food".
A total of 107 participants from 97 households were included in the study, approximately 36% of the total population aged 35-44 years. The sample presented a mean age of 39.3 (SD 3.32). Average years of schooling was 4.8 (SD 3.85), and mean household monthly income was U$135.50. Nevertheless, 19 participants presented missing data on household monthly income, as they did not know or declined to answer the question. Mean number of people per household was 5.45 (SD 1.62). Only four families did not include any children. Mean number of affirmative responses to EBIA was 8.90 (SD 4.81, Min 0, Max 14). According to the EBIA scores, approximately 95.3% of the participants live in food insecure families. The distribution of individuals according to their household food insecurity status is presented in Table 1.
In the bivariate analysis, participants with up to 4 years of schooling (p=0.007), beneficiaries of the Bolsa Família program (p=0.001) and those living in households with more than 5 members (p=0.041) had significantly higher food insecurity scores than their corresponding peers (  (Figure 1). The Bolsa Família benefit remained associated with higher scores of food insecurity after adjustment for the other variables (PR 1.47; 95%CI: 1.03-2.08). Households with more than 5 members (PR 1.26; 95%CI: 1.02-1.57) and higher oral health impact scores on quality of life (PR 1.02; 95%CI: 1.01-1.03) were also associated with higher scores of the EBIA instrument in the multivariate analysis. Prevalence of severe food insecurity was 55% higher among participants living in households with more than 5 members (p=0.048). Higher oral health impacts on quality of life were also associated with the presence of severe food insecurity (PR 1.02; 95%CI: 1.01-1.04). Clinical outcomes related to dental caries were not associated to household food insecurity (Table 3).

Discussion
In this study, food insecurity affected a significant proportion of households with Indigenous Kaingang adults aged between 35 and 44 years. More than half of the participants presented the most serious form of food insecurity, which includes the manifestation of hunger among the family members. The phenomenon was associated to the Bolsa Família cash transfer benefit, households with more than 5 members, and adults with higher impacts of oral health on quality of life.
Few studies have reported the household food insecurity situation experienced by Indigenous peoples. Surveys carried out with the Teréna and Guarani-Kaiowá ethnic groups, both from the State of Mato Grosso do Sul in Brazil, observed food insecurity in 76% and 100% of the households investigated, respectively 30,31 . Canadian  Fitted values Score researchers, using a questionnaire equivalent to EBIA, found a 70% prevalence rate of household food insecurity in a First Nation community 32 . The severe food insecurity status observed among adults from the Guarita Indigenous Land (57.94%) presents a large discrepancy in relation to other Indigenous peoples (ranging from 17% to 28%), and the non-Indigenous population in Brazil (3.2%) [30][31][32][33] . These findings suggest that hunger represents a marked manifestation of health and social inequalities between indigenous and non-indigenous. The high frequency of severe food insecurity observed among the Indigenous from the Guarita Indigenous Land highlights the social vulnerability experienced by this population. Experience of hunger is, arguably, the most obvious and immediate form of violation of the human right to adequate food. Given the interdependence and indivisibility of human rights, it also represents a serious violation of rights to the territory, cultural identity, health, housing, education, water, justice and human dignity 31 . The existence of hunger in the Guarita Indigenous Land is directly associated with the complex land issue faced by native peoples in Brazil. Although the Federal Constitution recognizes the traditional rights of occupation of Indigenous communities over their territories, the current political and institutional dismantling of the National Indian Foundation (Fundação Nacional do Índio -FUNAI), an institution responsible for protecting and promoting Indigenous rights, has relativized such constitutional prerogative 34 . The reduction of ancestral territories, unjustified slowness in land demarcation processes, and  4,31 . Even though the Guarita Indigenous Land is legally recognized by the Brazilian Federation as an Indigenous territory, the illicit exploitation of the Kaingang land by non-Indigenous is a phenomenon that occurs for decades and deprive the Indigenous families from their traditional ways of subsistence 35 . Emerged among small producers from the Via Campesina movement around 1993, the concept of food sovereignty considers the democratic distribution of food production resources at all scales, with their fair application to different social and ecological contexts 36 . In relation to Indigenous peoples, food sovereignty also refers to the collective ability to make decisions about healthy and culturally appropriate food practices 37 . In this context, the Kaingang people from the Guarita Indigenous Land was the first Indigenous community to receive the food production label "Indígenas do Brasil" granted in 2015 by the Brazilian Ministry of Agrarian Development. The label indicates the origin of products grown or collected on Indigenous lands, adding market value to traditional food systems. Further studies are needed to understand the real effects of this type of measure on food security of Indigenous populations.
The high prevalence of severe food insecurity among Bolsa Família beneficiaries demonstrates the ability of the program in identifying households in situation of extreme poverty and greater need. On the other hand, evidence indicate that Bolsa Família, when isolated from other public policies, has limited effects on the household food security status, not being able to significantly change the context of families living in situation of moderate and severe food insecurity 38 . As a result, families living in contexts of greater vulnerability may remain food insecure even when receiving the cash transfer benefit, which partially explains the association of food insecurity and the Bolsa Família program reported in this study.
The effects of the Bolsa Família program on the dietary pattern of beneficiaries seem to be heterogeneous. Researchers have observed positive impacts of the cash transfer benefit on families' food purchases and diet quality, including increased household expenditures on food, higher caloric intake, and higher consumption of fresh or minimally processed items 39,40 . Other studies indicate that the increased purchasing power of vulnerable families is associated with unhealthy food choices, higher consumption of soft drinks and sugar, and lower intake of fruits and vegetables [41][42][43] .
On general, the Bolsa Família program seems to be effective in increasing the variety and quantities of the food items consumed through the promotion of the families's access to food markets, although households frequently adopt a high-calorie and low-nutrient diet 44 . The unhealthy food choices made by groups in situation of food insecurity might not be simply a matter of lack of information, but rather an actual strategy employed by families to cope with the food-resource inadequacy 45 . Families experiencing financial constraints are likely to reduce diet costs by prioritizing the purchase of energy-dense foods, which often leads to lower overall diet quality 46 .
The high intake of sugar-rich foods has been long established as the major dietary factor in the development of dental caries 47 . Furthermore, micro and macronutrients deficiencies potentially compromise the oral health in individuals with poor dietary patterns 48 . Due to the cumulative burden of dental caries through the lifespan and the lifelong effects of sugars on the dentition, individuals living in a context of food insecurity might experience high prevalence rates of carious lesions and tooth loss 22 . Studies conducted in Canada and the United States observed that adults in situation of food insecurity present higher frequencies of use of dental prostheses, experience of dental pain, and negative self-perception of oral health 24,25 . Nevertheless, household food insecurity was not associated to any clinical variable related to dental caries among Kaingang families. The positive association observed in this study between household food insecurity and the OHRQoL scores suggests a higher perception of oral disorders by individuals in a context of greater vulnerability.
In addition to contributing at some level to improve the food consumption quality, the Bolsa Família may positively impact the oral health status of families in situation of food insecurity by allowing the acquisition of oral hygiene kits 49 . Underserved families who receive the benefit also tend to experience improved access to health care services, especially vaccination, check-ups and growth monitoring 50 . Yet, oral health check-ups are not included as a conditionality of the program.
Cash-transfer programs contributes to increase families' autonomy and empower the most vulnerable subgroups of society 51 . The benefit also impacts the family structure by increasing the female decision-making power regarding contraception, which in turn may have positive implications for the household food security status 52 . In this study, severe food insecurity was associated with higher household size, similarly to the findings reported for the Teréna Indigenous people from Brazil 30 . A systematic review assessed the food security status in different socio-demographic contexts in Brazil and reported that number of members living in the household was one of the main factors associated to food insecurity, as well as income and type of housing 53 .
Few studies have assessed the impact of the Bolsa Família program on Indigenous populations. On general, the benefit is well accepted by the communities, perceived as a form of assistance specially targeting the children needs, and mostly used for direct expenditures on food or acquisition of resources for food production 54 . In many Indigenous communities the benefit has become the main source of income 55 .
In order to enhance the repercussions of Bolsa Família on HFS status is mandatory its expansion and association with public policies focused on common issues faced by Indigenous peoples such as lack of basic sanitation, food production, and food sovereignty. Our findings highlight the need to formulate culturally-appropriate policies addressing the serious situation of social vulnerability and violation of the human right to adequate food in Indigenous populations.
Our study describes the first investigation on the association between HFS and oral health outcomes in Indigenous peoples. The Kaingang communities constitute vulnerable, hard to reach populations with scant information available regarding their oral health profile. In this manner, this report contributes with original perspectives related to the phenomenon and the population investigated. Conducting studies with Indigenous peoples impose significant challenges to epidemiological research. Geographic and social characteristics of the Kaingang population such as the disperse distribution over a large territory influenced the final sample size of the study. The number of participants was also affected by a seasonal flow of Indigenous adult workers, specially men, during the period of the study, to outside of the Indigenous Land. Even though the final sample was relatively small, it is safe to assume that the findings of this study are representative of a segment of the Guarita's Kaingang community, particularly of families led by young adults. Due to the paucity of available epidemiological data for Indigenous peoples, the cautious discussion of the findings justifies their use in a critical analysis, despite eventual low external validity or precision 35 .
Although the calibration process undertaken by the examiner was an intensive training directly supervised by an experienced researcher and clinician, the interexaminer agreement was not estimated in this phase. Another limitation of this study was the inability to blind the examiner, as the data collection for each participant was conducted in sequence (oral examinations followed by interviews) by the same researcher, which may have exposed observations to some degree of bias. Subgroups susceptible to food insecurity, such as families composed exclusively of the elderly, were not evaluated. The cross-sectional design prevented the establishment of inferences of cause and effect. years. The instrument used for the measurement of household food insecurity does not consider cultural aspects of indigenous food practices such as the importance of food exchanges, collectivity, and environment.
To the best of our knowledge, this was the first assessment of the oral health-related quality of life in an Indigenous population from Brazil. Therefore, the use of the OHIP-14 questionnaire in this study was essentially exploratory. The instrument was not previously validated to the study population and, as a result, it may have not been fully understood by all participants. We used the version validated by Oliveira and Nadanovski 28 for the Brazilian population. All participants were Portuguese speakers and no interpreter was required during the interviews. This exploratory, population-based, cross-sectional, study conducted with Kaingang adults from the South of Brazil demonstrated that household food insecurity in Indigenous families is associated with worse oral health-related quality of life, higher household size and the benefit of the Bolsa Família program. The vulnerability of Kaingang Indigenous families is accentuated by the disproportionately high frequency of food insecure families and, more specifically, by the number of those living with severe food insecurity. The importance of these findings for public health underlies the need to address structural determinants related to housing, oral health care, health inequalities, social welfare policies, and food production through inter-sectorial, culturally-sensitive approaches. Further studies should assess the association of HFS and oral health in Indigenous communities with different sociocultural characteristics, as well as to attempt to develop explanatory pathways to the investigated association.

Collaborations
GH Soares participated in concept and design of the study, data collection, and writing the manuscript. JMS Mota participated in interpretation and relevant critical review of the manuscript's intellectual content. FL Mialhe participated in the interpretation of data, and relevant critical review of the manuscript's intellectual content. MGH Biazevic participated in the manuscript concept, interpretation of data, and relevant critical review of the manuscript's intellectual content. ME Araújo participated in the interpretation of data, and relevant critical review of the manuscript's intellectual content. E Michel-Crosato performed the fundraising, participated in the concept and design of the study, coordination and supervision of the data collection, statistical analysis, and writing and revising the manuscript.