Diabetes and stress : an anthropological review for study of modernizing populations in the US-Mexico border region

Introduction: Diabetes is a growing worldwide problem, characterized by considerable ethnic variation and being particularly common in modernizing populations. Modernization is accompanied by a variety of stressful sociocultural changes that are believed to increase the risk of diabetes. Unfortunately, there is little accurate knowledge about impact of stress on the risk of diabetes in the US–Mexico border area. Methods: Literature searches were performed in PubMed and Google Scholar to identify anthropological studies on stress and diabetes. Snowball and opportunistic sampling were used to expand the identified literature. In total, 30 anthropological studies were identified concerning the role of stress and modernization on diabetes among Indigenous peoples. This article reviews the available information regarding stress and diabetes in different populations from various anthropological perspectives. Results: Four different concepts of stress were indentified: physiological, psychological, psychosocial and nutritional stress. Unlike physiological and nutritional theories of diabetes, psychological and psychosocial theories of stress and disease lack etiological specificity. No study addressed all four concepts of stress and few studies addressed more than two concepts. Most studies concerned nutritional stress and the developmental origins of diabetes. Most studies were conducted on the Pima Indians of Arizona and Mexico. All four stress concepts have some evidence as determinants of diabetes. Conclusion: These theoretical concepts and ethnographic results can provide the basis for developing comprehensive research protocols and public health intervention targeted at diabetes. A comprehensive view of stress can potentially explain the high


Diabetes as a worldwide epidemic
Type II diabetes mellitus (T2DM) has become a worldwide epidemic.From an estimated 2.8% global prevalence (171 million persons) in 2000, its prevalence rose to 347 million cases in 2008 1 and is expected to rise up 5.3% (438 million cases) by 2030 2,3 .Once thought to be related to affluence and development, countries in the process of economic development now face the largest burden from diabetes 2,4 .Diabetes poses an enormous public health burden, accounting for 11.6% of global health spending and 6.8% of global all-cause mortality 5 .Undiagnosed diabetes, more serious in developing countries with limited surveillance resources and more extensive rural areas with reduced healthcare access, worsens the burden 6 .Undiagnosed diabetes was 3.5 times more prevalent in low socio-economic status groups than in high 7 .Incomerelated health disparities are particularly important in developing countries or Indigenous populations.Diabetes prevalence is exhibits considerable ethnic disparities 4,8 .Diagnosed diabetes among US Hispanics and non-Hispanic Blacks was twice the rate of non-Hispanic Whites 6,9 .Similarly, total diabetes was higher among Hispanic than non-Hispanic White populations in 46 states, with undiagnosed diabetes higher among Hispanics in all 50 states 10 .
Diabetes is a complex multi-factorial disorder 19 , with genetic 19 and environmental 11,12,20,21 determinants.Anthropologists have often tended to favor one or the other side, biomedical or cultural, as the predominant explanation.Medical anthropologists are concerned with the subjective lived experience of people with diabetes and the ways bodily illnesses reflect sociocultural power relationships 4,22,23 .However, biological anthropologists have often favored unidentified 'thrifty genotype' genetic susceptibilities among Indigenous peoples to explain their high prevalence of diabetes 11,12,24,25 .
Statistical evidence from genome-wide association studies has implicated numerous regions in elevated diabetes risk 26 .However, independent replication rates are very low and causal variants remain unknown 26,27 .The rapid (<2-3 generations) emergence of diabetes does seem incongruous with a common genetic etiology for 300 million Indigenous people in thousands of cultures across the globe 28,29 .
The drastic rise in diabetes prevalence among Indigenous peoples in the 20th century seems more likely to be explained by known social factors than unknown genetic risks 19,22,28,30,31 .An understanding of culturally constructed models of diabetes causality is necessary for designing effective public health interventions 32,33 .And an exclusive focus on subjective experience risks overlooking known biomedical determinants of diabetes 22 .A middle way is suggested that incorporates both biological and cultural stress and stressors in diabetes etiology 22,34 .
This article is a review of the anthropological literature on the prevalence and incidence of diabetes in modernizing populations from an anthropological perspective.This review focuses on the role of psychosocial stress, broadly conceived as modifiable behavioral or social determinants, in environments characterized by modernization and economic development.

Definitions and measures of stress
The human stress response stems from the body's reaction to physical, biotic, and/or sociocultural stimuli resulting in the need to appraise adaptive activity 34 .Contemporary stress research derives from Hans Selye, who recognized that stress has both adaptive and destructive effects and proposed to measure stress by hormones that maintain or disturb physiological homeostasis 35 .Stressors included both internal (nervous tension) and external (infectious) factors, and could cause both specific (affecting particular bodily systems) and non-specific (whole-body) responses 35 .Stressors could be bad (distress), which can lead to exhaustion and ill health, or good (eustress), which provides a healthy level of tension.
The vague definition of stress as wear and tear on the body, made it difficult to measure the effects of stress on health 35- 38 .But the theoretic distinction between stress as a psychophysical state, and stressors as noxious environmental stimuli, allowed study of external social stressors, subjective mental or emotional states, and their relationship [36][37][38] .Since homeostasis was maintained by the neuroendocrine system, stress could be measured objectively by hormones of the hypothalamic-pituitary-adrenal (HPA) axis, and disturbed homeostasis could increase disease susceptibility 36,38 .Chronic stress resulted in fatigue, led to exhaustion, reduced resistance and increased ill health due to disturbed homeostasis and eventually caused disease or death 39 .This notion of allostasis contrasts to homeostasis and refers to an alternative but dysfunctional physiological set point induced by various stressors including aging and measured by various biomarkers 40,41 .
The following operational definitions are used in this article: • Physiological stress refers to the general adaptation response involving the HPA axis and is measured by levels of 'stress hormones' (adrenalin, norepinephrine, cortisol) that underlie the human body's adaptive response to external stimuli.
• Psychological stress refers to subjectively perceived mental or emotional states, as measured by various survey instruments, open-ended interviews or other self-reports.
• Psychosocial stress refers to objective measures of external sociocultural stressors associated with modernization or acculturation, including socioeconomic status, lifestyle incongruence (LI), urbanization, chronic (migration, trauma) or acute (familial or work-place conflicts) stressors, changes to traditional lifestyles (diet, physical activity) and environmental stressors (toxins, extreme weather events).These stressors are similar to the social determinants of health studied by public health practitioners 42 .
• Nutritional stress refers to under-nutrition, overnutrition or malnutrition, particularly in the perinatal period, as it pertains to Barker's theory of the developmental origins of chronic disease 20,25 .

Methods
A literature review was performed in PubMed using the  22,43 .A subsequent refinement added the search terms 'Hispanic' or 'Mexican-American' or 'Mexican' in the text and returned 3 articles, none of which were pertinent to this topic.These meager results indicated that a formal literature search using PubMed was inadequate to the task.
Due to these limitations, three additional procedures were incorporated to identify pertinent literature.First, similar searches were conducted on GoogleScholar: 'diabetes' and 'stress' in the title returned 2010 references, while 'diabetes' and 'anthropology' in the title returned 4 articles, one of which was used here 44 .Second, abstracts of the 83 PubMed articles were reviewed and the 2010 GoogleScholar references were browsed to determine their relevance to the research question.Third, it was quickly determined that a better method to identify pertinent historical and recent studies was to use a combination of snowball sampling from the bibliographies of found references 22,44,45 , and opportunistic sampling from PubMed when retrieving known abstracts, to identify further substantive articles of interest.

Study limitations
The main study limitation identified was the inadequacy of search engine results using PubMed and GoogleScholar.
While this approach can be a good starting point for systematic reviews of biomedical topics, it was inadequate to identify a sufficient number of references on the desired topic.Another limitation was the multi-dimensional and often incompatible definitions of stress used across a wide range of studies, which resulted in large numbers of irrelevant hits in both Pubmed and GoogleScholar.Finally, the number of studies identified that address T2DM as related to stress in ethnic populations or anthropological perspectives was limited.

Theories about stress and chronic disease
Psychosocial stress and chronic disease: The very diversity of studies on stress and diabetes (Table 1) suggested a need to better characterize and categorize the nature of stress and help ensure better comparability among studies 46 .Two major theoretical perspectives on chronic diseases and stress were identified: psychosocial stress (sociocultural and psychological) and biomedical stress (physiological or hormonal and nutritional).Theories on stress and chronic disease distinguish between subjectively perceived stress (a private mental or emotional state), external factors that induce stress (sociocultural stressors),physiological changes involving the HPA axis that underlie behavioral responses to stressors, and nutritional stress 2,21,[36][37][38] .Psychosocial stress can be psychological (subjective psycho-emotional states) or sociocultural (social support, financial issues, chronic familial or workplace conflict, acute traumatic events).Psychological stress, as subjective psycho-emotional states (anxiety, depression) can be inferred from verbal reports, written questionnaires or open-ended interviews.The impact of stress can vary depending on personality, interpretation, social support and coping style 36,37,[47][48][49] .The sociocultural context and cultural meaning of events can influence how psychosocial stressors are interpreted 47,49,50 .Psychological stress can affect health status by increasing behavioral risk factors, such as tobacco or alcohol consumption 51,52 .
Physiological stress can be measured by levels of stress hormones (adrenalin, norepinephrine and cortisol) 51 .Persistent elevation of stress hormones in people exposed to social situations normatively described as 'stressful' (including spousal loss, workload and migration) suggested a connection between subjective psychoemotional states and their social causes 51 .Stress hormones can increase diabetes risk by stimulating release of glucose or insulin and lead to obesity and hyperinsulinemia 11,51 .Stressors themselves can be chronic (everyday stress) or acute (migration, traumatic life events), and either sociocultural (including environmental) or nutritional in nature 2,46,47 .Excessive chronic stress can eventually lead to emotional, physical, and behavioral breakdown 39 .Social stressors included crowding, occupational mobility, social isolation, family breakdown, migration, urbanization and economic development [36][37][38] .Acute stressors are traumatic events that demand psychological, behavioral or social readjustment, such as migration, divorce or job loss 48,49 .Chronic stressors derived from discord related to routine social roles, including family, job, marriage or money 48,49 .
Specific psychosocial stressors were related to health outcomes in modernization 49,50 .It has been argued that the stress of modernization differs from both acculturation (Western education, speaking English) and biomedical risk factors (diet, obesity, physical activity) [48][49][50]53 . Speifically, lifestyle incongruence is the disjunction between people's desired socioeconomic status and their actual attainment and can lead to chronic disease 48,49 .Lifestyle incongruence is primarily a theory of consumption during modernization 9,49,50 .
These theoretical developments provided a research focus for studies of stress and diabetes and indicated a need to measure a variety of stress processes or social stressors 37,38 .In Cassell's theory, social stressors produce non-specific effects on health [36][37][38] .This lack of etiological specificity meant that there are no 'stress diseases' but only altered incidences of routine diseases due to stress 36- 38 .Consequently, all disease outcomes (rather than particular diseases of interest) must be studied to characterize the effects of exposure to social stressors 37 .
Nutritional stress and chronic disease: Nutritional stress is another stress theory related to modernization.The developmental origins of chronic disease postulates that perinatal malnutrition leads to chronic disease in adulthood.Fetuses adapt in utero to their nutritional milieu, as determined by maternal condition and diet during pregnancy and previous nutritional experience 54,55 .The fetus responds to under-nutrition by altering changing hormone production, tissue sensitivity to hormones, and growth rates.The result is reduced fetal growth rates, low birth weight (LBW), stunting, reduced rates of postnatal growth, and altered organ/body weight ratios 55 .The timing of malnutrition can impact different organs differently, depending on critical growth periods affected 55 .Therefore, unlike psychosocial stress, nutritional stress can induce etiological specificity.
The developmental origins model proposed that diabetes can be a long-term consequence of the hormonal changes induced by maternal malnutrition 20,21 .This model proposed that early-life malnutrition led to decreased fetal insulin, insulin-like growth factors and glucose which control fetal growth, thus leading to reduced fetal growth rates and permanently altered insulin and glucose metabolism 20,55 .The thrifty phenotype hypothesis posits fetal and early life malnutrition impose nutritional thrift upon the individual, leading to impaired pancreas development and increased susceptibility to T2DM 20,21 .Low birth weight also induces pre-diabetic traits of high blood glucose and insulin, insulin resistance, and impaired beta cell function 30,[56][57][58][59] .Subsequent childhood over-nutrition resulted in accelerated childhood growth insulin resistance, glucose intolerance, obesity and T2DM diabetes in later life 20,21,60 .Offspring glucose and insulin levels were associated with low maternal BMI independently of offspring birth weight 55,60 .The developmental origins model postulates that prenatal nutritional stress permanently altered physiological and metabolic pathways.Subsequent nutritional stress, including obesity and nutritionally poor diets, can lead to T2DM in mid-life 61 .
Epidemiological studies have suggested that adult noncommunicable chronic diseases have a host of determinants, from genetic and adult lifestyle factors to environmental factors acting in early life 62 .In summary, psychosocial stress in conjunction with nutritional deficits can potentially explain the diabetes epidemic in modernizing countries and among Indigenous peoples worldwide.

Anthropological literature on modernization, stress and diabetes
Theories of psychosocial and nutritional stress and their impact on chronic disease are related to anthropological studies of modernization.Modernization is the process of economic and sociocultural change, by which traditional cultures industrialize and develop a capitalist economy characterized by division of labor, reduced importance of kinship, and changes in traditional lifestyles 48,63,64 .Modernization is related to increased chronic disease morbidity and mortality and is a major focus of contemporary medical anthropology 48,[65][66][67] .The unique contribution of anthropological studies is that the global diabetes epidemic strongly impacts modernizing populations 5,67 and occurs during the nutritional transition 68,69 involving reduced physical activity, increased obesity and diets rich in processed carbohydrates 1,20,70 .
The real challenge in anthropological studies of diabetes is how to operationalize 'modernization'.Drastic lifestyle changes involving nutrition, activity patterns, residence, and psychosocial stress related to post-WWII modernization are related to diabetes among Indigenous peoples 11,71 .
Urbanization is a composite variable for several aspects of modernization, including decreased physical activity, increased healthcare access, and differential access to Western foods 72 .Urbanization also correlates with access to wage-labor jobs and socioeconomic factors 47,50 .With some exceptions 18,43,73 , data to compare diabetes incidence or prevalence between Westernized and traditional Indigenous peoples are sparse 12 .Anthropologists have operationalized the lifestyle changes associated with modernization by measuring activity patterns, acculturation scales, affluence and consumption, dietary changes and nutrition, urbanization and other psychosocial factors 11,18,71,73,74 (results in Table 2).

Psychological and physiological stress
Anthropological studies on subjective stress, though infrequent 51 , have been conducted [74][75][76][77] .Levels of epinephrine were higher in Westernized urban Samoans than in Samoans having a more traditional lifestyle as rural agriculturalists, but norepinephrine levels did not differ 76,77 .Rural villagers had higher levels of life satisfaction and emotional stability compared with urban Samoans, thus correlating stress hormone levels to subjective stress 77 .A study of Mexican-American migrant farm-workers in Wisconsin demonstrated associations between dopamine β hydroxylase (DBH, a norepinephrine precursor), acute stressors (familial separation, migration), and diabetic status 78 ..These locations also corresponded to differences in language, education, occupation and blood pressure (BP) 81 .Activity levels were categorized by occupational information recorded on health certificates 81 .More westernized males and those with more sedentary life-styles had elevated CVD-related mortality 81 .

Psychosocial stress
The effects of modernization on diabetes risk was not directly assessed, but there was a significantly higher ageadjusted incidence of diabetes-related mortality among Samoans (32.2 deaths/100 000 population) relative to the US (13.4 deaths/100 000) 81 .
Finally, 2 unspecified American Indian populations (Northern Plains, Southwest) were studied using a 25 item early-life traumatic events and a 30 item chronic stress instrument 43 .Results showed that diabetes in the Northern Plains population was associated with early-life trauma, while the Southwest population diabetes was associated with discrimination and community addiction problems 43 .
Lifestyle incongruence: Lifestyle incongruence refers to a discrepancy between material consumption (lifestyle) and economic means 48 .The effects of LI on BP in Western Samoa were studied using a 21 point scale for lifestyle/material possessions and an 8 point occupational scale 82 .For men, LI was associated with elevated systolic BP.For women, diastolic BP was higher when occupational rank exceeded lifestyle, the opposite of LI 82 .In American Samoa, LI also varied by sex 83 .The effects of modernization versus status incongruence on hypertension were directly compared in a Mexican community 84 .The modernization model predicted that stress accumulated with the degree of modernization and increased chronic disease risk 84 .Modern traits included wage-labor, Western education, urban residence, reduced kinship support network, and adoption of modern attitudes, behaviors and values 84 .They also measured subjectively perceived stress (anxiety, depression, life-satisfaction, recent life events, social support, tension and worry) 85 .The alternative explanation was lifestyle incongruence 84,85 .Results showed that systolic and diastolic BP increased with increasing modernization but not with LI 84 .Among the Mississippi Choctaw tribe, LI was associated with elevated non-fasting plasma glucose 85 , while psychological stress was associated with elevated diastolic BP 85 .These mixed results suggest that more research on the effects of LI and acculturation on diabetes is needed.Cultural factors may also modify the impact of lifestyle incongruence 53 .

Dietary change:
The effects of traditional diet and dietary change, often conceptualized biomedically as adoption of a high-fat, high-carbohydrate, low-fiber diet, have been studied as a risk factor for diabetes among Indigenous peoples 11,12,71 .An interesting anecdote about the impact of refined carbohydrates on diabetes came from India in 1907.The earliest reported cases of diabetes occurred in areas near the first industrial rice mills, which removed most of the bran fiber from the grain 71 .By contrast, areas that continued to use traditional home-pounding of rice, which left half the bran fiber, had lower incidence of diabetes 71 .
Many traditional Australian Aboriginal, Pima Indians and Pacific islander foods protect against hyperglycemia by being digested more slowly than Western foods and result in lower levels of blood glucose and insulin [86][87][88] .Westernized diabetic Aborigines who temporarily (7 weeks) reverted to a traditional hunter-gatherer diet and lifestyle had improved blood glucose, insulin and triglycerides and reduced insulin resistance 89 .Pima Indians who ate a Western diet were 2.9 times more likely to develop diabetes, while those on a mixed diet were intermediate at 1.6 times more likely, compared with those eating a traditional diet 88 .Over-eating by non-diabetic Pima Indians increased plasma insulin 90 .Dietary changes among the Eastern Oklahoma Cherokee, including replacement of traditional beans, corn and squash with refined flours and a change from boiling or broiling to deep-frying, are thought responsible for the increase in diabetes prevalence, from 0% before 1941, to 13% in 1974 91 .Traditional methods for processing wheat, including parboiling and baking whole grains, resulted in lower glycemic response compared to bread baked with industrially milled flours 92 .Unhealthy dietary changes often result from greater availability and affordability of energy rich, nutrition poor foods 93 .The increase in nutritional risk factors is an important factor in the increasing prevalence of diabetes and other chronic diseases in developing countries 72 .
The effects of acculturation and dietary factors on CVD risk factors were studied in Yaqui and Tepehuanos Indian communities in Mexico 52, 94 .Diabetes was assessed as a CVD risk, along with obesity and triglycerides 52 .A 12 point acculturation scale was used and included language use, media exposure and social relations 95 .The Yaqui Indians were more acculturated than the Tepehuanos, who were subsistence farmers with a vegetarian diet until Western foods were introduced in 1998 94 .The Yaqui were more obese (48% vs 7% prevalence), had more hypertriglyceridemia (15% vs 43% prevalence), less LDL-cholesterol (34% vs 42% prevalence), and a remarkable 22fold higher prevalence of diabetes (18.3% vs 0.8%) than the more traditional Tepehuanos 52 .Impaired fasting glucose was higher among the Yaqui (17%) than the Tepehuanos (5%) 52 .Neither total daily energy consumption nor total fiber intake differed between groups, but they did differ in the energy derived from saturated fats (1412 Kj/day Yaqui, 978 Kj/day Tepehuanos) 52 .The Yaquis' higher diabetes prevalence was attributed to their higher saturated fat consumption 52 .
A 64 item food questionnaire was administered in 12 Yaqui communities at 2 time-points, 1994 (traditional vegetarian diet) and 2004 (Westernized diet) to assess total caloric intake and food proportions derived from protein, carbohydrates and saturated and unsaturated fats 96 .Traditional foods (green vegetables, bread, beans, potatoes and mixed-root tortillas, occasional meat) and Western foods (refined flours, canned foods, pasta, soft drinks, and 'junk' food) were compared 94 .In 10 years of follow up, dietary saturated fat doubled while unsaturated fats declined by 75%, fiber intake declined from 53 g/day to 49 g/day, and total caloric intake increased from 1476 KCal/day to 2100 KCal/day 94 .These dietary changes were associated with increased overweight/obesity (11% to 22%), hypertriglyceridemia (3% to 17%), impaired fasting glucose (6% to 15%), and decreased HDL (71% to 10%) 94 .
Hypertension prevalence doubled from 2% to 4%, and metabolic syndrome increased from 0% to 10%, although diabetes prevalence did not change at 1% 52, 94 .Acculturation among the Tepehuanos Indians was characterized by dietary changes involving increased consumption of calories and saturated fats, reduced fiber and was associated with an increased prevalence of risk factors 94 .

Physical activity change:
Decreased energy expenditure is primary feature of lifestyle change during modernization that can lead to obesity and diabetes 71 .Historically recent (>1930s) changes in activity patterns among Oklahoma Cherokee involved decreased walking and increased reliance on vehicular transportation and were associated in time with increased diabetes-related mortality (from 0 cases in 1941 to 26.2/100 000 in 1977), which was 63% higher than Caucasian rates 91 .Most studies on the role of reduced physical activity as a diabetes risk factor concerned the Pima Indians.Questionnaires were used to estimate the amount of time spent in various activities 96 .Traditional Mexican Pimas had 23 h/wk of occupational physical activity, compared with 5 h/wk for more Westernized Arizona Pimas 96 .These differences corresponded to higher prevalence of diabetes among Arizona Pima (38.2%) than Mexican Pima (6.2%) 96 .A later study confirmed these results, with Arizona Pimas having physical activity of 3.1 h/wk (female) to 12.1 h/wk (male), compared with Mexican Pima females (22 h/wk) and males (32.9 h/wk) 97 .

Prenatal nutritional stress
The only anthropological studies concerning the developmental origins of diabetes were conducted on the Pima Indians.Several large studies examined the relationship between known birth weights and serum glucose or insulin levels 98,99 .Diabetes prevalence and 2 hour fasting glucose concentration had a U-shaped relationship with birth weight 98,99 .The LBW Pima were more insulin resistant, after controlling for weight and height 98 .The highest adult prevalence of diabetes (52.4%) was found in the lowest onethird (<2500 g) and the highest one-third (>4500 g) of birth weights.High prevalence among high birth weight (HBW) babies was attributed to gestational diabetes 99 .The LBW babies were 3.8 times more likely to have diabetes as adults, compared with middle birth weight babies 99 .Only approximately 6% of Pima babies were LBW and there was no evidence of maternal nutritional stress during pregnancy, making Barker's developmental model an incomplete explanation of the high (≥40%) adult diabetes prevalence 99 .The high prevalence of diabetes among LBW babies was attributed to selective survival of diabetes-prone LBW infants through an unknown genetic mechanism 99 .In another study, HBW Pima babies exhibited a dose-response relationship between extent of bottle feeding with cow's milk and adult diabetes prevalence 100 .Diabetes at age 30-39 years had 20% prevalence among Pima babies exclusively breast fed for at least 2 months, 25% prevalence among the cohort with some breast feeding, and 30% prevalence among those exclusively bottle fed with cow's milk 100 .Perinatal nutritional stress associated with later-life diabetes seemed related to nutritional overload of cow milk bottle-feeding, or possibly a missing nutrient specific to breast milk which affected infant insulin sensitivity 100 .

Discussion and Conclusions
A survey of the literature on stress and chronic disease identified four concepts of stress: physiological, psychological, psychosocial, and nutritional.To varying degrees, all are determinants of diabetes and should be studied in any comprehensive study.The anthropological survey demonstrated how studies of various anthropological populations undergoing modernization have operationalized stressors that influence diabetes risk.The remaining question of interest was how to measure modernization stressors in the US-Mexico border area, in order to study their influence on diabetes risk.
In fact, many of these stress concepts have empirical support in Hispanic populations.Hispanic youth with metabolic syndrome have higher serum cortisol than overweight healthy controls, as expected from physiological stress theory 101 .Acculturation measures in Hispanic populations are associated with reduced prevalence of breastfeeding 103 , which is associated with increased risk of diabetes in adulthood 102 , as suggested by the developmental origins model of diabetes 20,21 .Measures of modernization are associated with reduced dietary quality (decreased intake of fiber and fruit, and increased consumption of saturated fats and sugar) among diabetic Hispanics 104,105 .Hispanic household food insecurity is associated with increased consumption of poor-quality foods, including fats and sugar 106 .But no studies were found in the border area that addressed all four concepts of stress in diabetes etiology.

Several
suggestions can be made for best practice.Throughout, the importance is underscored of a comprehensive view of stress and use of multiple indicators 34,75 .If blood samples can be collected, diabetes can be directly inferred by fasting glucose levels using established WHO criteria.Tissue sampling would also allow measurement of stress hormones (salivary cortisol, serum epinephrine).If blood sampling is not feasible, diabetes prevalence could be estimated by self-reporting (having been told by a physician that they are diabetic).Furthermore, the few available studies indicate that 'stress hormones' do not always measure stress 76,77,79 .In that regard, it may be inadvisable to collect tissues at all.Biomedical risk factors should be measured, including diet (by 24 hour recall or food frequency questionnaires [FFQs]), physical activity levels (to assess work and leisure activity patterns) 104 .Surveys should address the presence of barriers and motivating factors for increased physical activity 107 .Obesity (BMI) can be measured using basic anthropometry (height and weight) in either clinical or field situations.Lifestyle incongruence can be inferred using Dressler's instrument 50 .Childhood nutritional stress can be estimated by parental recall 106 , which is reliable up to 42 years later 102 .It can also be retrospectively estimated in adults by measurement of adult leg length (the difference between height and seated height) [108][109][110] .Acculturation can be measured by unidimensional scales, like the Short Acculturation Scale (SAS), which utilizes the single dimension of language use 95 Such uni-dimensional acculturation scales have been criticized because culture is more than language, culture change is not identical to stress, and changes in diet and physical activity do not fully explain the increased risk of chronic disease encountered during modernization 49,50,85,111 .Multidimensional scales, like the Acculturation Rating Scale for Mexican Americans-Revised (ARMSA II), also measure the attitudes, beliefs, values and behaviors that constitute the actual processes of cultural change, while still incorporating language and social factors like immigration status and length of residence [112][113][114] .

Conclusion
In the authors' view, lifestyle changes, multidimensional acculturation scales, FFQs, subjective stress instruments 115 , and anthropometry are needed to fully assess the impact of stress as a determinant of diabetes.This knowledge can be used to design more effective public health interventions and accurately target diabetes prevention efforts toward individuals at greatest risk of developing diabetes in the US-Mexico border area or similar modernizing populations.