Assimilation and acculturation have long been recognized as important though complex correlates of change in health risk profiles of immigrants and the resulting ethnic populations(Reference Lara, Gamboa, Kahramanian, Morales and Hayes-Bautista1, Reference Williams and Collins2). The standard model that dominates research on acculturation and health suggests that new immigrant populations typically have a set of risk profiles that are distinctive from those of the population of the host society in which they have settled. These differences may reflect a combination of influences, including the maintenance of culturally distinctive behaviours characteristic of the country of origin; the distinctive influences of the immigration experience itself, including disruption of personal networks and exposure to discrimination; and the correlation of the decision to migrate across national boundaries with distinctive personal characteristics. Time spent in the host society – measured in years, and sometimes generations, among the descendents of immigrants – tends to erode these differences. Social epidemiologists frequently turn to the variables of time and, where relevant, linguistic change – the adoption of the language of the destination society – as correlates of changes in social and behavioural risk profiles away from those characteristic of the immigrant group itself, and towards those characteristic of segments of the broader population of the country of destination(Reference Lara, Gamboa, Kahramanian, Morales and Hayes-Bautista1, Reference Arcia, Skinner, Bailey and Correa3–Reference Hunt, Schneider and Corner5).
Recently, increased attention has been given in the social epidemiology literature to the influence of the social-spatial context of health. In particular, a growing literature investigates variation in local social environments with respect to variables such as quality of food supply, local modelling of healthy diets and personal habits, stressfulness of daily living and encouragement of physical activity(Reference Sampson, Raudenbush and Earls6–Reference Moore and Roux10). In the context of the social scientific study of immigrant incorporation, this emphasis is concordant with a well-documented relationship linking ethnic residential enclaves with the maintenance and intergenerational transmission of ethnic-specific cultures(Reference Gordon11–Reference Palloni and Arias13). Drawing on these broader social science research findings, epidemiologists have investigated the hypothesis that ethnic concentration of immigrant-derived populations in neighbourhoods is associated with the maintenance of group-specific social behavioural practices that influence health outcomes.
The Mexican-American population of the USA provides a particularly noteworthy case for the investigation of the effects of residential concentration on health. Compared with non-Hispanic whites, Mexican-Americans have lower mortality rates from all causes, and from leading causes including CVD and cancers at most common sites(Reference Palloni and Arias13–Reference Eschbach, Stimpson, Kuo and Goodwin18). Mexican-American mortality rates are consistently reported to be lower for immigrants than they are for Mexican-Americans born in the USA(Reference Palloni and Arias13, Reference Hummer, Rogers, Nam and LeClere15, Reference Eschbach, Kuo and Goodwin17–Reference Elo, Turra, Kestenbaum and Ferguson19). That mortality rates are lower for a Mexican-American population that is on average socio-economically disadvantaged has been described as an epidemiological paradox(Reference Markides and Coreil20, Reference Franzini, Ribble and Keddie21).
One of the leading hypotheses about the better than expected health and mortality outcomes for Mexican-Americans, as well as the apparently poorer outcomes for US-born Mexican-Americans compared with immigrants from Mexico, points to protective aspects of health-related behaviours among immigrants, including healthier diets, lower rates of smoking, substance and alcohol use, and higher rates of physical activity(Reference Palloni and Arias13, Reference Singh and Siahpush16, Reference Franzini, Ribble and Keddie21, Reference Markides and Eschbach22). Indeed, studies of acculturation have documented a relationship between acculturation and less healthy lifestyles in the Mexican origin population, although effects of acculturation are not entirely negative. Notably, both health-care access and use of screening improves with greater acculturation.
A small but growing number of studies have examined the hypothesis that a high concentration of Mexican populations in residential communities is associated with better health outcomes. To date, results have been mixed. Some studies report evidence of lower mortality, lower chronic disease morbidity, better mental health and higher self-rated health(Reference Aneshensel and Sucoff23–Reference Cagney, Browning and Wallace31). Others report weak, contradictory or null results(Reference Palloni and Arias13, Reference Lee and Ferraro32, Reference Frank, Cerda and Rendon33).
In the present study we investigate the relationship between ethnic residential concentration of Mexican-Americans and dietary intake. Specifically, we investigate the hypothesis that there is a strong relationship between ethnic concentration, e.g. residence in a barrio community, and types of foods consumed. This hypothesis has a high degree of plausibility, because a high level of ethnic concentration in a local community creates a context for the supply of ethnic-specific food products and for the modelling of dietary practices. The Mexican-American population lives in very diverse residential settings, ranging from homogeneous ethnic environments in near-border areas in the south-west, to neighbourhoods throughout the USA where they are highly integrated with non-Hispanics. Do dietary practices among Mexican-Americans in different neighbourhood settings differ in ways that suggest that integration with other groups leads to deterioration of dietary practices that help explain the increasing rates of chronic disease prevalence among more acculturated Hispanics?
Methods
Data source
The Third National Health and Nutrition Examination Survey (NHANES III), a large US survey conducted from 1988 to 1994, is a major source of information on the nutritional and health status of the US population aged 2 months or more(34). The strength of this survey is that it used the same stratified multistage probability design as previous National Health and Nutrition Examination Surveys(34). Weights indicating the probability of being sampled were assigned to each respondent, enabling results to represent the US population for each group. Mexican-Americans were over-sampled to produce statistically reliable health estimates for the largest ethnic minority subgroup in the USA. The data were collected via standardized questionnaires administered by health professionals at participants’ homes; standardized medical examinations by physicians, medical technicians and other health professionals at the National Health and Nutrition Examination Survey mobile examination centres (MEC); and laboratory tests on whole blood and sera. Interviews were conducted in English and Spanish after informed consents were obtained at the initial home interview. The interviewer gave each person selected for the survey a brochure which described the survey procedures using a question-and-answer format and included photographs of people being examined in the MEC rooms. The final page of the brochure was a paper that required the signature of each participant 18 years of age and older(34, Reference Woteki, Briefel, Hitchcock, Ezzatfa and Maurer35). Response rates were high, 78 % completed both the home interview and the medical examination. To get the percentage of Mexican-Americans at the census tract level, the NHANES III was merged with the 1990 US Census data. To avoid any potential identification of subjects, the merge of the NHANES III public database(36) with neighbourhood data (US Census Bureau, 1990)(37) was made by the National Center for Health Statistics (NCHS) Research Data Center (Hyattsville, MD, USA). We sent the statistical models needed for our analyses and the NCHS remote system sent us back the results. The study protocol was approved by the University of Texas Medical Branch Institutional Review Board.
Study sample
The sample for our analyses included 5306 Mexican-American men and women aged 17–90 years who completed both the home questionnaire and medical examination.
Measurements
The outcomes were food frequencies and serum levels of nutrients.
Food frequencies were assessed by a 1-month qualitative FFQ(Reference Thompson and Byers38, Reference Kant39). The NHANES III nutrient database for individual foods is derived from the US Department of Agriculture’s Survey Nutrient Database(Reference Kant39, 40). To ensure the accuracy of the nutrient contents of foods, substantial care was taken to include a wide variety of traditional Mexican foods (e.g. red chilli peppers)(Reference Dixon, Sundquist and Winkleby41, Reference Shah, Coyle, Kavanaugh, Adams-Huet and Lipsky42). The FFQ, administered during the household interview, was used to ask respondents about the average number of times foods were eaten during the 1-month period preceding the interview date. Frequencies of consumption of foods from the following food groups were ascertained: fruits, vegetables, grains and legumes.
Serum levels of nutrients have been shown to correlate well with dietary intake of respective nutrients(43). The micronutrients examined included those considered to be of potential public health significance and thought to decrease the risk of cancer or CVD(44–Reference He, Nowson and MacGregor48). Serum levels of the following nutrients were obtained: lycopene, Se, vitamin E, vitamin D, vitamin A, vitamin C, vitamin B12, folate, α-carotene and β-cryptoxanthin. Serum levels of nutrients were determined by nutritional biochemistry. MEC collected blood samples and used the following assay or instrumentation methods for laboratory assessments: ‘Quantaphase Folate’ RIA Kit (Bio-Rad Laboratories, Hercules, CA, USA) for folate; HPLC (Waters Corporation, Milford, MA, USA) for vitamins A, C and E and carotenoids (lycopene, α-carotene, β-cryptoxanthin); INCSTAR 25-OH-D RIA Kit (INCSTAR, Stillwater, MN, USA) for vitamin D (25-hydroxyvitamin D3); 125I-folic/57Co-B-12 for vitamin B12; and graphite furnace atomic absorption using Perkin–Elmer model 3030 and 5100 instruments (Perkin–Elmer Co., Norwalk, CT, USA) for Se(34).
A measure of contextual acculturation, the percentage of Mexican-Americans at the census tract level (a higher percentage indicates more isolation or less integration with other ethnic groups)(Reference Eschbach, Ostir, Patel, Markides and Goodwin25) was used as a continuous variable. Other variables were age (years, used as a continuous variable) and gender (male and female).
Statistical analyses
All statistical analyses were carried out using the statistical software packages SAS for Windows version 9·1 (SAS Institute, Inc., Cary, NC, USA) and SUDAAN version 7·11 (Research Triangle Institute, Research Triangle Park, NC, USA). All analyses incorporated sampling weights that adjusted for unequal probabilities of selection. Because of the complex survey design used in NHANES III, traditional methods of statistical analysis based on the assumption of a simple random sample may not be reliable. Sample weights are needed to produce correct estimates of population quantities. Other aspects of the sample design (e.g. PSU (primary sampling units) pairings) should be taken into account to obtain correct standard errors and significance levels for hypothesis testing(Reference Skinner, Holt and Smith49, Reference Brogan50). With continuous outcome variables, frequency of foods (e.g. cereals, tomatoes) or serum levels of nutrients (e.g. Se, lycopene), we used age- and gender-adjusted linear regression analyses (REGRESS procedure) to examine the independent association of the percentage of Mexican-Americans at the census tract level with food frequencies and serum levels of nutrients.
Results
The study population comprised 2682 Mexican-American men (50·6 %) and 2624 women (49·4 %); 35 % of subjects were aged 17–29 years, 29·2 % were aged 30–44 years, 13·5 % were aged 45–59 years, 17·5 % were aged 60–74 years and 4·8 % were 75 years of age and older. Age group distributions did not differ by gender. Eighty-eight per cent of subjects came from three of the four US–Mexico border states: California, Texas and Arizona. These states correspond to the south-west area of the USA where the majority of Mexican-Americans reside.
Table 1 shows the multivariate linear regression analyses for the relationship between consumption of specific foods (more detailed description is provided in the table) and percentage of Mexican-Americans at the census tract level. It shows that increased percentage of Mexican-Americans in the neighbourhood was associated with less consumption of melons (unstandardized beta coefficient (b) = −1·21, se 0.52, P = 0·0266), any other fruits (e.g. apples, bananas; b = −4·57, se 1.34, P = 0·0017), carrots (b = −1·61, se 0·70, P = 0·0273), spinach/greens (b = –1·17, se 0·27, P = 0·0001) and broccoli (b = –1·84, se 0·5, P < 0·0001). On the other hand, increased percentage of Mexican-Americans in the neighbourhood was associated with more consumption of the group of fruits that included peaches, nectarines, apricots, guava, mango and papaya (b = 0·32, se 0.68, P = 0·63). Although this association did not reach statistical significance, it suggests that these fruits – especially the traditional mango and papaya – may be important diet components of high-density Mexican-American neighbourhoods. By contrast, increased percentage of Mexican-Americans in the neighbourhood was associated with more consumption of corn products (b = 11·12, se 2·98, P = 0·0006), flour tortillas (b = 7·17, se 2·63, P = 0·0097), tomatoes (b = 2·76, se 0·94, P = 0·0060), hot red chilli peppers (b = 4·05, se 1·48, P = 0·0097) and legumes such as beans, lentils or chickpeas/garbanzos (b = 11·56, se 1·81, P < 0·0001).
Table 1 Multivariate analysis resultsFootnote * for frequency of foods as a function of the percentage of Mexican-Americans at the census tract level: outcome data in Mexican-American men and women (n 5306) were obtained from the Third National Health and Nutrition Examination Survey (1988–94) and linked to the 1990 Census
b, unstandardized beta coefficient; se, standard error of the beta coefficient.
* Adjusted for age and gender.
Table 2 shows the multivariate linear regression analyses for the relationship between serum levels of nutrients and percentage of Mexican-Americans at the census tract level. It shows that increased percentage of Mexican-Americans at the census tract level was associated with lower levels of lycopene (b = −3·77, se 0·62, P < 0·0001), Se (b = −4·99, se 1·55, P = 0·0033), vitamin C (b = −0·10, se 0·03, P = 0·0025) and folate (b = −0·88, se 0·33, P = 0·0117). Increased percentage of Mexican-Americans at the census tract level was also associated with higher levels of β-cryptoxanthin (b = 1·53, se 1·18, P = 0·20) and vitamin B12 (b = 191·2, se 179·1, P = 0·32) but did not reach statistical significance.
Table 2 Multivariate analysis resultsFootnote * for serum levels of nutrients as a function of the percentage of Mexican-Americans at the census tract level: outcome data on Mexican-American men and women (n 5306) were obtained from the Third National Health and Nutrition Examination Survey (1988–94) and linked to the 1990 Census
b, unstandardized beta coefficient; se, standard error of the beta coefficient.
* Adjusted for age and gender.
Discussion
In the present study, we focused on spatial aspects of assimilation in relation to dietary quality using data from a geo-coded NHANES III data set, where the characteristics of tract populations from the 1990 census were attached to individual records to investigate the relationship between the social characteristics of tract populations and nutrient profiles measured through self-reported dietary recalls and measurement of serum nutrients. The emphasis on spatial characteristics of residential communities as a possible correlate of changes in nutrition is consistent with a very old social science research finding that residential concentration of immigrant and ethnic populations serves to maintain ethnic-specific cultural patterns(Reference Gordon11, Reference Massey12). It is also consistent with a recent research stream in epidemiology that investigates the socio-economic characteristics of neighbourhoods of residence in relation to health-related behaviours and health outcomes(Reference Kawachi and Berkman51). It also identifies a variable with a broad range among Mexican-Americans, who are distributed across a broad spectrum of residential environments, ranging from a high degree of ethnic segregation in the border region of Texas, to full integration with non-Hispanics in urban and suburban communities throughout the USA(52).
Dietary patterns including vegetables and fruits have been associated with lower risk of all-cause mortality using data from the National Health Interview Surveys(Reference Kant, Graubard and Schatzkin53) and the Breast Cancer Detection Demonstration Project(Reference Kant, Schatzkin, Graubard and Schrairer54); and with lower risk for CVD using data from the Physicians’ Health Study(Reference Liu, Lee, Ajani, Cole, Buring and Manson55), the Nurses’ Health Study(Reference Forman, Rimm, Stampfer and Curhan56, Reference Hung, Joshipura, Jiang, Hu, Hunter, Smith-Warner, Colditz, Rosner, Spiegelman and Willett57), the Health Professionals’ Follow-up Study(Reference Hung, Joshipura, Jiang, Hu, Hunter, Smith-Warner, Colditz, Rosner, Spiegelman and Willett57, Reference Hu, Rimm, Stampfer, Ascherio, Spiegelman and Willett58) and the Framingham Nutrition Studies(Reference Millen, Quatromoni, Nam, O’Horo, Polak and D’Agostino59). In addition, case–control and cohort studies showed that vegetables and fruits have been associated with reduction in the risk of some cancers including mouth and pharynx, oesophagus, stomach, colon-rectum, larynx, lung, breast (vegetables only), ovary (vegetables only), bladder (fruits only) and kidney(Reference Riboli and Norat60–Reference Pavia, Pileggi, Nobile and Angelillo62).
However, higher concentrations of Mexican-Americans in a neighbourhood are correlated with poverty and disadvantage; therefore, the pattern of low consumption of some fruits (e.g. cherries, berries) and some vegetables (e.g. broccoli) in our study may reflect unaffordable costs for foods or lower availability in a neighbourhood food environment(Reference Eschbach, Mahnken and Goodwin26). Indeed, in another study, Bodor et al.(Reference Bodor, Rose, Farley, Swalm and Scott63) reported that greater fresh vegetable availability within 100 metres of residences was a positive predictor of vegetable intake.
Studies of health and mortality patterns of Mexican-Americans living in the USA have previously reported greater longevity(Reference Palloni and Arias13, Reference Eschbach, Kuo and Goodwin17–Reference Elo, Turra, Kestenbaum and Ferguson19) and lower biological risk profiles(Reference Crimmins, Kim, Alley, Karlamangla and Seeman64) for Mexican-American immigrants residing in the USA compared with non-Hispanic whites and US-born Mexican-Americans. US-born Mexican-Americans appear to have mortality rates and biological risk profiles similar to or not much worse than those of non-Hispanic whites, which some commentators appear to ascribe as unexpected because of the much lower average socio-economic status of US-born Mexican-Americans compared with non-Hispanic whites.
The explanation of the lower mortality and better than expected biological risk profiles of Mexican-Americans remains a matter of investigation and debate. Recent work has substantially removed data quality concerns as the principal explanation of these patterns(Reference Elo, Turra, Kestenbaum and Ferguson19, Reference Turra and Goldman65, Reference Turra and Elo66). A second hypothesis suggests that the greater propensity to immigrate of persons with better health may play a leading role, although direct evidence for this hypothesis remains weak. A third set of explanations points to healthier socio-cultural risk profiles as a contributing element. More nutritious diets for immigrants, lower rates of smoking and substance use, and stronger social support, are frequently hypothesized to contribute to the Mexican-American mortality advantage. These hypotheses are concordant with evidence showing decreases in the quality of diets and health-related behaviours with increasing time and generation in the USA(Reference Lara, Gamboa, Kahramanian, Morales and Hayes-Bautista1, Reference Kant39, Reference Guendelman and Abrams67, Reference Montez and Eschbach68).
On the other hand, high consumption of legumes (especially beans) and hot red chilli peppers may reflect cultural preferences and more affordable foods for this Mexican-American population(Reference Dixon, Sundquist and Winkleby41, Reference Shah, Coyle, Kavanaugh, Adams-Huet and Lipsky42). It has been reported that dietary patterns are different and generally less healthy for US-born compared with immigrant Mexican-Americans. Dixon et al.(Reference Dixon, Sundquist and Winkleby41) reported that US-born Mexican-Americans consumed significantly more fat and less fibre and vitamins, and were less likely to meet dietary guidelines than were immigrant Mexican-Americans. Also, Guendelman and Abrams(Reference Guendelman and Abrams67) reported that first-generation Mexican-American women had higher average intakes of protein, vitamins A and C, folic acid and Ca than second-generation Mexican women, whose nutrient intake resembles that of white non-Hispanic women. Other studies have shown that acculturation to the US culture among Mexican-Americans was associated with increased dietary fat and sugar along with higher waist circumference and abdominal obesity(Reference Dixon, Sundquist and Winkleby41, Reference Mazur, Marquis and Jensen69–Reference Perez-Escamilla and Putnik71). A potential explanation on how acculturation affects diet among Mexican-Americans is related to a higher food store availability and consumption of fast food in inner-city neighbourhoods(Reference Unger, Reynolds, Shakib, Spruijt-Metz, Sun and Johnson72, Reference Galvez, Morland, Raines, Kobil, Siskind, Godbold and Brenner73). The consumption of these other foods may influence the selection of a healthier diet profile in our study population of Mexican-Americans.
So the question we asked was whether there was evidence of dietary advantages in more rather than less ethnically homogeneous Mexican-American communities that could help explain lower rates of incidence and mortality for some chronic diseases for Mexican-Americans as a population group. Therefore, consumption of higher amounts of legumes (beans, lentils or chickpeas) may protect the health of a population with high concentration of Mexican-Americans in the neighbourhood. Indeed, consumption of higher amounts of legumes may have a protective effect against cancer. Kolonel et al.(Reference Kolonel, Hankin and Whittemore74) reported that intake of legumes (whether total legumes, soya foods specifically, or other legumes) was inversely related to prostate cancer risk. Also, Correa(Reference Correa75) examined data from forty-one countries and found a significant inverse correlation between bean consumption and mortality due to prostate, breast and colon cancer. In other human or animal studies, high consumption of dry beans has been associated with lower rates of myocardial infarction among Costa Ricans or fewer colon adenocarcinomas among rats(Reference Kabagampe, Baylin, Ruiz-Narvaez, Silesw and Campos76, Reference Hughes, Ganthavorn and Wilson-Sanders77). These findings may be part of the explanation why those Mexican-Americans living in neighbourhoods with a high concentration of Mexican-Americans exhibit lower cancer incidence or lower overall mortality(Reference Eschbach, Ostir, Patel, Markides and Goodwin25, Reference Eschbach, Mahnken and Goodwin26). In Mexico, common beans are the second source of protein, carbohydrates, vitamins and minerals after corn(Reference Diaz-Batalla, Widholm, Fahey, Castano-Tostado and Paredes-Lopez78, Reference Espinosa-Alonso, Lygin, Widholm, Valverde and Paredes-Lopez79). Beans contain complex carbohydrates and are rich in Mg, Cu and α-linoleic acid; these components may improve insulin sensitivity and lipid profiles(Reference Kabagampe, Baylin, Ruiz-Narvaez, Silesw and Campos76). Beans are also an excellent source of non-nutritive constituents such as fibre, protease inhibitors, phytic acid, isoflavonoids, lignans and polyphenols such as tannins. These compounds have antioxidant, antimutagenic and anticarcinogenic activities and are also free radical scavengers(Reference Diaz-Batalla, Widholm, Fahey, Castano-Tostado and Paredes-Lopez78–Reference Gonzales de Mejia, Castano-Tostado and Loarca-Pina82). In addition, capsaicin, the major pungent ingredient in red peppers, decreases the growth (e.g. inducing the apoptosis) of human and in vitro prostate cancer cells(Reference Mori, Lehmann, O’Kelly, Kumagai, Desmond, Pervan, McBride, Kizaki and Koeffler83), human leukaemic cells(Reference Ito, Nakazato, Yamato, Miyakawa, Yamada, Hozumi, Segawa, Ikeda and Kizaki84), gastric(Reference Kim, Kim, Pyo, Kim, Kim, Yu and Han85) and hepatic carcinoma cells in vitro (Reference Jung, Kang and Moon86). Finally, consumption of tomatoes has been found to have protective cardiovascular effects, with potential protection for prostate, oesophagus, stomach, lung and breast cancer(Reference O’Kennedy, Crosbie, van Lieshout, Broom, Webb and Duttaroy87–Reference Giovannucci, Rimm, Liu, Stampfer and Willett89).
One limitation of our study is the cross-sectional design of the NHANES III, which prevented us from drawing causal inferences. Dietary assessment tools also have inherent limitations. A serum level of nutrients and 1-month qualitative FFQ are not representative of individual nutrient intakes because of day-to-day variation in food consumption. However, serum levels of nutrients are an objective measure, and we included a proxy for contextual acculturation – i.e. neighbourhood density (percentage of Mexican-Americans at the census tract level) – that may capture other contextual factors related to the environment where Mexican-Americans live(Reference Eschbach, Ostir, Patel, Markides and Goodwin25–Reference Patel, Eschbach, Rudkin, Peek and Markides28).
The NHANES III questionnaire does not distinguish between traditional and non-traditional fruits (e.g. papaya or mango v. apricots) or other foods (e.g. corn tortillas v. corn muffins) among Mexican-Americans; this may lead to biased estimations or underestimations of some traditional foods in this population. In addition, NHANES III includes a mixed group of unprocessed-corn products such as bread or muffins but, with the exception of corn tortillas, does not include the consumption of processed-corn and specifically masa products that are essential foods in the diet of countries of Hispanic origin in the Americas. Masa is used to make tortillas (or tortillas chips), tamales, pozole, arepas and empanadas(Reference Bello-Perez, Osorio-Diaz, Agama-Acevedo, Solorza-Feria, Toro-Vazquez and Paredes-Lopez90, Reference Martinez-Flores, Figueroa, Martinez-Bustos, Gonzalez-Hernandez, Rodriguez Garcia, Banos Lopez and Garnica-Romo91). Masa is obtained after thermal-alkaline treatment, or a nixtamalization process, of the corn kernels. It involves lime-cooking (calcium hydroxide solution), followed by steeping for 12–16 h, washing and stone-grinding the corn grains to produce masa. Cooking the corn with lime significantly increases its Ca (>700 %), P and Fe content(Reference Martinez-Flores, Figueroa, Martinez-Bustos, Gonzalez-Hernandez, Rodriguez Garcia, Banos Lopez and Garnica-Romo91). Ca from masa acquires great relevance because it represents almost the only source of Ca in some Latin American countries. Masa products provide an important source of energy, proteins, dietary fibre, antioxidants and nutrients such as phytochemicals and carotenoids (e.g. lutein, zeaxanthin, β-cryptoxanthin)(Reference De la Parra, Serna Saldivar and Liu92). However, lime-cooking affects the amount of resistant starch and the quality of protein. For example, the partial removal of the pericarp or bran leads to finished products that are considered as semi-whole grain foods(Reference De la Parra, Serna Saldivar and Liu92). Also, digestibility of the protein is decreased slightly, possibly because hydrophobic interactions, protein denaturation and cross-linking of proteins occur during maize processing that change the solubility of these components, which could affect amino acid release during enzymatic digestion(93).
The Hispanic population is the largest minority group in the USA, and Mexican-Americans constitute the majority of this group. Isolated Mexican-American communities tend to maintain many of their traditional foods; however, public health campaigns are necessary not only to promote these traditional foods but also to make accessible other essential foods in their diet. On the other hand, more research is needed to assess potential health-protective effects of traditional Mexican-American foods such as avocados or specific kind of beans (e.g. pinto, black).
In conclusion, an increased percentage of Mexican-Americans at the census tract level was associated with less consumption of selective or non-traditional foods (e.g. some fruits such as melons, apples, berries; or vegetables such as broccoli) and low levels of serum Se and vitamin C, but it was associated with more consumption of traditional foods such as corn products, legumes (beans, lentils and chickpeas), tomatoes and hot red chilli peppers. Thus, consumption of these traditional foods may make a difference to the health risk profiles in this population. Further studies are needed to determine if other nutrients or foods (e.g. masa products) that were not include in the data may influence dietary profiles in high-density Mexican-American neighbourhoods. Also, research is needed to explore whether unhealthier practices such as the consumption of fast foods or sedentary lifestyles are common among isolated Mexican-American neighbourhoods.
Acknowledgements
Sources of funding: The study was supported by research grants W81XWH-06-1-0290 from the Department of Defense and P50 CA10563-02 from the National Cancer Institute. The sponsors had no role in the design, methods, data collection, analysis, or manuscript preparation. The interpretation and reporting of these data are the sole responsibility of the authors. Financial disclosure: None of the authors has any conflict of interest related to this work. Author contributions: C.A.R.-O. and K.E. conducted the literature review. C.A.R.-O., H.J. and Y.-F.K. participated in the acquisition of data and provided statistical expertise. C.A.R.-O., K.E. and J.S.G. were responsible for supervision of the study and obtained funding. All authors participated in study conceptualization and design, interpretation of data, and editing the manuscript.
