Introduction
Worldwide, food insecurity has emerged as a leading public health crisis with recent estimates showing approximately 750 million people, or 1 in 10, globally experience severe levels of food insecurity(1). As defined by the Food and Agriculture Organization of the United Nations, food insecurity exists when people have a lack of ‘physical, social, and economic access to sufficient, safe, and nutritious food that meets their dietary needs and food preferences’(Reference Jaron and Galal2,3) . Factors that increase the risk for food insecurity include low levels of education, limited social networks and limited economic development, occurring at both the individual, population and country level(3–5). Indigenous communities, defined as communities that are culturally separate from the dominant society, experience an increased risk for food insecurity compared to the majority population(Reference Restrepo-Arango, Gutierrez-Builes and Rios-Osorio6,Reference Power7) . This increased risk for food insecurity may be attributed to indigenous communities experiencing marginalisation and limited access to needed resources as a result of upholding a unique social identity while maintaining political and economic structures outside of mainstream society(Reference Power7–Reference Gracey and King10).
Latin America has the second-highest prevalence of food insecurity in the world, with 51 % of the population reporting any level of food insecurity(Reference Jones11). They also have a large indigenous population, with about 45 million indigenous peoples living in Latin America, making up almost 8·5 % of the population in this region(12). Panama has the third-highest population of indigenous peoples in the region, with 12 % of the population reporting indigenous status(12). The Kuna Indians are one of Latin America and Panama’s largest indigenous communities residing on both the mainland and off the coast in the nearby San Blas Islands(Reference Vakis and Lindert13). With a significant history of marginalisation, the Kuna experience high levels of poverty and poorer health compared to the general Panamanian population due to isolation and limited access to resources(Reference Vakis and Lindert13). However, though research on food insecurity has been conducted in other indigenous communities, such as the Inuit communities of Canada and the aboriginal communities of Australia, limited research has been completed with the Kuna Indians of Panama(3,Reference Frongillo, Nguyen and Smith4,Reference Huet, Rosol and Egeland14) .
Recent work conducted with the Kuna Indians of Panama found an extremely high prevalence of mental health conditions, including depression and suicidal ideation(Reference Walker, Campbell and Dawson15). Conversations with local community leaders suggested social factors, such as food insecurity may play a role in this finding as the community noted changes in access to food over the past decade. Prior evidence collected through research with other indigenous and non-indigenous groups suggests this relationship may indeed exist(Reference Weaver and Hadley16–Reference Jalal, Frongillo and Warren24). However, little work has been conducted in Latin America and no published work on the relationship between social factors, such as food insecurity, and mental or physical health exists for the Kuna Indians. Since prior work found that indigenous communities may experience food insecurity differently, an investigation into the prevalence of food insecurity in the Kuna Indians and its possible association with mental health outcomes is highly warranted. Therefore, this study aimed to evaluate food insecurity in an indigenous population in Panamá and understand its association with mental health outcomes.
Methods
Research approvals
The current research was developed in collaboration with the Panamá Ministry of Health, the Kuna Congress and Indigenous Health International, an approved 501(c)(3) non-profit organisation. Founded in 2012, the mission of Indigenous Health International is to meet the health care needs of indigenous communities around the world. Indigenous Health International supports indigenous communities in Panamá by conducting need assessments, delivering health care and conducting research to improve health services. Data used in this study was collected as an initial step in a long-term relationship with the Kuna community to guide future programs. Indigenous Health International received approval from the Panamá Ministry of Health, Director of Indigenous Health in Panamá, the national Kuna Congress and local Kuna community leaderships prior to initiation of any study procedures, and received approval to publish results of the initial data collection after discussion with the community. In addition, the current research protocol, research procedures and study-related documentation was approved by the Western Institutional Review Board, an accredited organisation that provides human subjects and regulatory compliance across 70 countries.
Population
This study was conducted in partnership with community members in the Ustupu/Ogobsucum islands, located within the indigenous zone off the coast of Panama (see Fig. 1). Approximately 3500 adults live in the two communities. The communities share a health clinic run by the Panamanian Ministry of Health and an airport allowing access to the islands. Ustupu and Ogobsucum are two of the Kuna Indian communities of Panama, the second largest Indigenous population in the country. A report published by the World Bank on poverty in indigenous populations in Panama noted that Kuna is the second largest indigenous group in Panama, and the majority live in San Blas Island indigenous zone off the coast of Panama(Reference Vakis and Lindert13). The remote location requires boat or air travel to access the population, which lives in villages characterised by strong social cohesion, political organisation and leadership responsible for maintaining cultural traditions(Reference Vakis and Lindert13). Agriculture, seafood farming and artisan activities are the primary activities and sources of livelihood(Reference Vakis and Lindert13).
Fig. 1 Map of Panama indicating San Blas region
All adults in the community (age 18 or older) were eligible to participate and provided verbal consent as approved by the Western Institutional Review Board to ensure they were aware of their participation in a research study. No incentive was offered, and medical care was provided separately from participation and completion of the survey.
Recruitment
No regular census exists for the community, so a sampling frame was not possible and convenience sampling was used. Recruitment was completed over a 1-week period in August 2013, so a variety of methods were used to invite all community members to participate if willing. First, community members trained by the research team approached individuals in the local health clinic and explained the study. Community members were trained to explain the nature of the project and that it was research, consent participants verbally and administer the questionnaire. Those selected were willing to work with the research team for the week and could speak Spanish and the local Kuna language. Second, local community leadership announced that survey administration would be conducted at the health clinic during evening meetings held daily and invited individuals to participate. Finally, word of mouth was used to inform community members of the ongoing study. Information on those who declined were not collected, however, few individuals who were approached directly to participate declined.
Data collection
Data were collected using a paper-based questionnaires. Questions were developed using measures from the ‘Panamanian prevalence of risk factors associated with cardiovascular disease in 18 year or older population survey’, a national survey used in Panamá to assess CVD risk factors in an adult population in 2010(Reference McDonald, Motta and Roa25), validated measures for psychosocial factors(Reference Kroenke, Spitzer and Williams26–Reference Cohen, Williamson, Spacapan and Oskamp28) and standardised health outcome measures. All surveys were translated into Spanish and back translated to ensure validity. Community members were trained to deliver the survey by reading questions word for word in Spanish or translating into the local spoken language, Kuna, if requested. Participants were able to complete the survey on their own in Spanish if preferred, and community members trained by the research team then checked through for completion and clarified any answers if necessary. Training included discussion of the overarching study, review of the survey and community members taking the survey themselves to ensure understanding of questions and response options. Scoring was completed by the research team after data were entered into a secure database and downloaded for analysis.
Food insecurity
To quantify food insecurity in the population using a validated scale, the short form US Department of Agriculture Food Security questionnaire(Reference Bickel, Nord and Price29), a six-item scale developed by the US Department of Agriculture for use by households with and without children, was used. Questions ask about whether respondents were worried food would run out, the frequency of food not lasting until more money was available, not being able to afford balanced meals, whether meal sizes were cut, or respondents were hungry, lost weight or ate less because of money. Questions were asked about both the respondent and members of their household and were not based on the level of income. Questions have been validated across multiple populations, including indigenous people in the USA. As questions do not specify food type, respondents can answer based on their personal food preferences. The scale has been validated outside the USA, including in a Caribbean community, finding similarly high Cronbach alpha of 0·87 and item response correlation coefficients ranging from 0·70 to 0·78(Reference Gulliford, Mahabir and Rocke30). Individuals that respond negative to all answers are assigned zero, with points for responses scored so higher scores indicate greater food insecurity. Individuals scoring zero or one were categorised as food secure, and individuals scoring two or more were categorised as food insecure, based on scoring outlined in the original report(Reference Bickel, Nord and Price29).
Measures of mental health
Three measures of mental health, capturing three separate mental health constructs, were included in the study. Prevalence of depression and serious psychological distress in this population have been reported elsewhere(Reference Walker, Campbell and Dawson15).
-
Depression measured by the nine-item Patient Health Questionnaire. The nine-item Patient Health Questionnaire is a valid and reliable measure across multiple populations with internal reliability of Cronbach’s alpha greater than 0·85(Reference Kroenke, Spitzer and Williams26). Participants are asked nine questions about the frequency of symptoms for depression, and responses could range from not at all to nearly every day. Higher scores indicate more depressive symptoms or more frequent symptoms of a similar kind. Based on a sensitivity and specificity of 88 % for major depressive disorder, scores greater than or equal to 10 are categorised as depression(Reference Kroenke, Spitzer and Williams26).
-
Serious psychological distress measured using the six-item K6 scale. The six-item scale was originally developed for population-level screening for non-specific psychological distress(Reference Kessler, Andrews and Colpe27) and is routinely used by the US National Health Interview Survey and the WHO Mental Health Survey(Reference Cohen and Zammitti31). Higher scores indicate more distress. Based on the validated scale, a cut-off of 13 or above is categorised as serious psychological distress.
-
Perceived stress measured using the perceived stress scale, a four-item scale that assesses the degree to which the respondent finds regular situations stressful. Responses range from never to very often, and questions ask about the frequency of feelings related to events in the previous month(Reference Cohen, Williamson, Spacapan and Oskamp28). The Cronbach alpha value is 0·69, and scores are highly correlated with stress, depression and anxiety(Reference Andreou, Alexopoulos and Lionis32). Higher scores indicate higher perceived stress.
Demographic variables
Demographic variables collected using self-report included: age (continuous), sex (dichotomous male/female), education based on years in school (categorised into primary, secondary, some college or college graduate or higher), marital status (categorised into single/separated/divorced/widowed as not married or married/other union) and monthly family income. Monthly family income was categorised following discussions with community leaders to provide context for different levels of income. Individuals reporting they did not know their income were categorised as ‘poor’ as they were considered subsistence farmers/fishers. Those making < $250 balboas per month [1 balboa = US$1] were categorised as ‘low income’. Those making $250 balboa per month or more were categorised as ‘moderate income’. These levels of income were specific to the indigenous community as mainland Panamanian income levels are generally higher.
Statistical analysis
To evaluate food insecurity in the indigenous Kuna population of Panama and understand its relationship on mental health, a series of analyses were conducted. First, frequencies were calculated to describe the demographic characteristics of the population surveyed (age, gender, education, income and marital status). Secondly, the proportion and 95 % CI of food insecurity was determined for the categories of each demographic characteristic and compared using Chi-squared tests. Third, a logistic regression model was run with food insecurity as the outcome and independent variables including each of the demographic characteristics to understand independent correlates of food insecurity. Finally, three unadjusted and three adjusted multiple linear regression models were run to understand the relationship between food insecurity and mental health outcomes. In the first unadjusted and adjusted models, depression score served as the dependent variable, with food insecurity as the independent variable and demographic characteristics as covariates in the adjusted model. In the second unadjusted and adjusted models, serious psychological distress score served as the dependent variable, with food insecurity as the independent variable and demographic characteristics as covariates in the adjusted model. In the third unadjusted and adjusted models, perceived stress score served as the dependent variable, with food insecurity as the independent variable and demographic characteristics as covariates in the adjusted model. Analyses were run using Stata v.14, and significance was determined based on two-tailed alpha of P < 0·05.
Results
The survey was completed by 211 individuals, of which 209 answered all questions in the food insecurity scale and were included in this analysis. Table 1 presents the sample demographics for the population. Of the sample, 40·1 % were age 18–39, 73 % were female, 70·3 % were married and 53·7 % had more than a primary school education. Approximately 51·9 % were poor, while 37 % reported low income and 11·1 % reported moderate income.
Table 1 Sample demographics (n 209)
Food insecurity was reported by 83·3 % (95 % CI 77·5, 88·0) of the sample population. Table 2 presents food insecurity status by age, gender, education and income. The only demographic variable where food insecurity prevalence varies across levels was age. Individuals in younger age groups reported higher levels of food insecurity, with 90·2 % (95 % CI 81·7, 95·7) of those between 18 and 39 reporting food insecurity, compared to 85·1 % (95 % CI 74·3, 92·6) of those between 40 and 59 and 71·9 % (95 % CI 58·5, 83·0) of those between ages 60 and90 years.
Table 2 Food insecurity status by age, gender, education, and income (95 % CI and OR values)
Bold font indicates statistical significance (P < 0·05).
Table 3 presents results of the regression model for demographic correlates of food insecurity in the Kuna Indian sample. All demographic variables (gender, age, education, income and marital status) were included in the same model to investigate independent associations after taking other demographic factors into account. Ages 60–90 compared to age 18–39 remained independently associated with food insecurity (OR = 0·18, 95 % CI 0·04, 0·77).
Table 3 Independent correlates of food insecurity
Bold font indicates statistical significance (P<.05).
Adjusted models included covariates for gender, age, education, income, and marital status.
Table 4 presents the unadjusted and adjusted regression models for each of the three mental health outcomes. In unadjusted models, food insecurity was significantly positively associated with depressive symptoms (β = 0·44, 95 % CI 0·10, 0·78) and serious psychological distress (β = 0·54, 95 % CI 0·23, 0·85), but not with stress (β = –0·06, 95 % CI –0·24, 0·12). After adjusting for demographics, food insecurity was still significantly positively associated with depressive symptoms (β = 0·43, 95 % CI 0·03, 0·84) and serious psychological distress (β = 0·68, 95% CI 0·30, 1·07).
Table 4 Independent association between food insecurity and mental health
Bold font indicates statistical significance (P<.05).
Outcome for each model was continuous mental health score, primary independent was food insecurity.
Adjusted models included covariates for gender, age, education, income, and marital status.
Discussion
In an adult indigenous population in Panama, food insecurity was very high at 83·3 % of individuals surveyed. Younger individuals were significantly more likely to be food insecure, with 90 % of those between the ages of 18–39 reporting food insecurity. When investigating the relationship between food insecurity and mental health outcomes, those who were food insecure reported more depressive symptoms and serious psychological distress but did not report higher levels of perceived stress. These results provide the first investigation of food insecurity and mental health in a remote indigenous population in Panama and provide important considerations when designing an intervention to promote mental health and well-being in this population.
The rates of food insecurity were high in the Kuna population studied, and analysis by demographic characteristics suggests a particular need to tailor interventions for younger adult age groups. While studies in indigenous groups have shown overall high levels of food insecurity, the influence in this population on younger adults is unique. For example, 62·5 % of Inuit communities in Canada reported food insecurity, 19–25 % of the Alaska Native population has been noted as experiencing food insecurity and 67·8 % of children from an indigenous community in Ecuador lived in a food insecure household(Reference Walch, Bersamin and Loring8,Reference Huet, Rosol and Egeland14,Reference Walrod, Seccareccia and Sarmiento33) . A systematic review of food insecurity in rural and peasant populations found very little work has been conducted in Central and South America, highlighting the importance of our findings focussed on an indigenous population in Panamá(Reference Restrepo-Arango, Gutierrez-Builes and Rios-Osorio6). Given the high prevalence of food insecurity in this community, and the evidence from other regions on the impact on long-term health, this study highlights an urgent need to begin addressing food insecurity in the Kuna community. Additional research is needed to identify the unique situations faced by younger indigenous individuals contributing to increased prevalence in those ages 18–39, to allow targeted interventions for this age group.
Additionally, food insecurity was associated with both depression and serious psychological distress in the Kuna Indian community, indicating the need to focus on social determinants of health when addressing mental health concerns. This relationship has been shown in other populations, but this is one of the first to our knowledge focussed specifically on the relationship between food insecurity and mental health in an indigenous population, and the first investigating the indigenous Kuna community. A study of aboriginal people living off reserves noted a relationship between food insecurity and high stress(Reference Willows, Veugelers and Raine34), and a study of First Nations people in Canada found an association between food insecurity and self-report of a mental health diagnosis by a health professional(Reference Deaton, Scholz and Lipka35). A study of coping strategies in Bangladesh found that while food insecurity was high across all groups, the means of coping varied by ethnic identity, with indigenous groups using more protective coping strategies(Reference Ali and Vallianatos36). One particularly strong factor that is often present in indigenous communities is social support mechanisms and social cohesion in the community(Reference Power7,Reference Walch, Bersamin and Loring8) . In a study of social structures and food insecurity in a rural Kenyan island community, Nagata et al. found higher food insecurity associated with lower levels of instrumental social support(Reference Nagata, Fiorella and Salmen37). Strong social networks may lower levels of stress and anxiety generally associated with food insecurity(Reference Sorsdahl, Slopen and Siefert21), and therefore, may modify the direction of the relationship in indigenous populations. Another factor that may be specific to indigenous populations is access to indigenous farming and food sources. A study in Botswana found households with more access to traditional and indigenous foods had lower levels of food insecurity(Reference Kasimba, Motswagole and Covic38). Indigenous communities often view health in a different way than western models of health, incorporating physical, mental and spiritual factors and encompassing the health of themselves and those around them(Reference Stephens, Nettleton and Porter39). This perspective lends itself to solutions that address concerns like food insecurity and mental health in the same intervention. Recognising that food insecurity is not only an aspect of food but is also tied to mental health and incorporating the cultural framework of indigenous populations into mental health treatment delivery may help improve food insecurity and mental health simultaneously.
While this study provides unique information on an underserved indigenous group in Panamá, there are some limitations. First, as each indigenous population has specific cultural and environmental factors unique to their community, the generalisability of these results may be limited. However, given the paucity of research on this topic in indigenous groups, these results recommend future research in populations previously not included in food insecurity and mental health research, and urgent action to address high levels of food insecurity. Secondly, the limited sample size and cross-sectional nature of the data collection preclude these results from indicating causality and should be considered as an association only. And finally, mental health symptoms were collected through self-report and not validated through a clinical interview. While a mental health diagnosis cannot be assumed, the self-report of symptoms is appropriate for preliminary research and was agreed upon as the best method by leaders of the community in which this work was conducted.
In conclusion, over 80 % of an indigenous community on the islands off Panamá reported food insecurity, with the highest levels reported between the ages of 18–39. Food insecurity was associated with higher levels of depressive symptoms and serious psychological distress, suggesting a need to incorporate social determinants of health into mental health treatment models. This study supports a targeted effort to address food insecurity in this population and the need to incorporate mental health and access or availability of food into a combined intervention.
Acknowledgements
Acknowledgements: We would like to acknowledge the support and participation of the Ustupu and Ogobsucum communities and the partnership with the Panama Ministry of Health that made this study possible. Financial support: This study was partially supported by Indigenous Health International (IHI), an approved 501(c)(3) non-profit organisation in the United States. Conflicts of interest: The authors declare that they have no competing interests. Authorship: LEE obtained funding for the study, coordinated data collection, conducted statistical analyses. RJW interpreted statistical analyses. All authors participated in data collection, drafted the manuscript, and revised the article critically for important intellectual content. All authors approved the final manuscript. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving research study participants were approved by the Western Institutional Review Board. The Western Institutional Review Board, is an accredited organisation that has been providing human subjects and regulatory compliance for more than 40 years across 70 countries. Verbal consent was obtained from all subjects. Verbal consent was witnessed and formally recorded. Consent for publication: Consent for publication was obtained from leaders of both study communities.
