The Mexican Health Paradox

Acculturation as a Problematic Explanatory Construct

There is currently no uniform definition of acculturation (Trimble in Chun et al., 2003; Hunt et al., 2004; Lara et al., 2004; Rudmin, 2003), originally defined in the anthropological literature as the process by which “members of different autonomous cultural groups come into continuous firsthand contact, with subsequent changes in the original cultural patterns and custom of either or both groups (e.g., language, values, lifestyle, attitudes, identity)” (Redfield, Lipton, Herskovits [1936, p. 5], in Organista, 2007a). It is now generally understood as a complex, multidimensional process that cannot be adequately described as linear movement from “Mexican” to “American,” but as involving multiple and distinct outcomes depending on individual, minority group-level, and minority-majority group relations and factors.

Current Critiques of Acculturation

Despite the emergence of more nuanced theories of acculturation (e.g., Castro, Shaibi, & Boehm-Smith, 2009; Padilla & Perez, 2003) many large-scale studies of acculturation and Latino health do not reference a theoretical framework, and a majority does not even define acculturation (Hunt, et al., 2004; Thomson & Hoffman-Goetz, 2009). Instead, most use proxy variables such as language preference or time in the United States. While these epidemiological studies are vital in documenting population level health trends over time, a growing body of critiques challenge such acculturation research because it may inadvertently reify antiquated notions of acculturation as linear process (Escobar & Vega, 2000), and oversimplify the complexity of health declines over time in some Latino populations.

A major criticism of studies invoking acculturation and the cultural buffer hypothesis is that they assume that Mexican immigrants bring with them a monolithic culture that is “new” to the United States. Hunt et al. (2004) point out that such studies ascribe so-called “traditional” characteristics such as familism, traditional gender roles and “religiously based morality” to Hispanic culture, whereas “purportedly ‘modern’ characteristics, such as individualism and greater independence from family support, high stress and tolerance for self-destructive behaviors are ascribed to ‘mainstream’ culture” (p. 979). Such assertions are both derogatory toward immigrant culture and illogical given that these two cultures have existed side-by-side and with increasing interpenetration for well over a century (Rudmin, 2003). Hunt and colleagues (2004) also point out that we now live in a globalized world wherein mass communication and media have few borders, thus the idea that Mexican and U.S. culture are new to each other seems embedded in the antiquated melting-pot model of immigrant assimilation.

Limitations of Acculturation in Explaining the Mexican Health Paradox

The oversimplification of acculturation in relation to health outcomes is problematic on at least two levels. First, it points health and mental health interventions in the direction of a simple cause-and-effect relationship between acculturation and behavior change, rather than also targeting the broader sociopolitical or ecodevelopmental context and social determinants of health and mental health outcomes. Second, the descriptive nature of large-scale epidemiological studies raises the larger question regarding what processes are being tapped by single variable proxy measures. That is, might acculturation itself be a proxy for larger processes affecting health outcomes (Betancourt & Lopez, 1993)? Understanding the environmental context of health is necessary to appreciate the combination of factors that contribute to the paradox and its subsequent decline (Alegria, 2009).

While several theories have been proposed to explain the MHP, namely, the salmon bias (i.e., elderly or infirm immigrants returning to Mexico to die), the healthy immigrant effect (selective migration pursued by the most vigorous immigrants), and faulty census data (misclassification of Hispanics as non-Hispanic), none of these have been supported by data and research (Morales et al., 2002). While the cultural buffer hypothesis has been shown to have some validity in relation to protective health effects (see for example, Hovey, 2000), a deeper examination of the literature suggests complex processes affecting health declines, not accounted for by acculturation as a direct causal mechanism in-and-of itself. An expanded view of the context of acculturation can enhance the explanatory power of acculturation as a construct in relationship to health outcomes.

Refinement of Berry’s Model of Acculturation

The relevance of Berry’s theory of acculturation to understanding U.S. Latinos is that it primarily focuses on indigenous and other racial and ethnic minority groups in North America and their historical experience with European colonial powers. As such, Berry notes that while in theory two groups in continuous contact can influence each other, in reality a larger and more powerful group typically dominates a smaller and less powerful group, and thus wields more power to influence and even oppress the nondominant group. Such dynamics set the stage for an acculturation process that is resistant, stressful, and conflictual for minority groups such as non-White minority groups in the United States. While critics of acculturation research (e.g., Rudmin, 2003) note the almost exclusive focus on minority rather than majority groups, such a focus is intended to study the negative side of SES and health disparities in the United States.

How stressful acculturation may be or not be depends on its historical course. Berry views the course of acculturation as a three-phase process characterized by (1) contact; (2) conflict; and (3) adaptation. Contact refers to the conditions under which two cultural groups meet (e.g., invasion, immigration, seeking refuge, etc.). Conflict refers to the predictable tension and fighting that results when one group attempts to dominate another, and adaptation refers to the eventual form of accommodation between groups that is intended to reduce conflict. While it is possible for little conflict to occur, the frequent dynamics of dominance and subordination make conflict a probable and often protracted phase of acculturation (e.g., non-White minority groups in the United States).

Which form of adaptation an ethnic minority group eventually experiences depends on the original conditions of contact, the degree of conflict between groups, and the dominant society’s tolerance for ethnic and racial diversity. Such factors influence the degree to which the nondominant group attempts to maintain or reinvent its culture of origin, and the degree to which positive relations between groups is possible. According to Berry, the four major forms of adaptation can be described by listing all possible combinations of yes/no answers to the following two questions: (1) Does the minority group attempt to retain its culture of origin, and (2) are positive relations possible between the minority and majority groups? (see Table 1)

OPEN IN VIEWER

Table 1.

Varieties of Adaptation to Acculturation for U.S. Minority Groups.

Attempt to retain culture?	Positive predominant relations possible with dominant culture?	Form of adaptation
No	Yes	Assimilation
Yes	Yes	Integration
Yes	No	Segregation
No	No	Marginalization

Source: Berry (2003).

For example, while indigenous in small numbers to what is now the southwestern United States, the historical course of acculturation for Mexican Americans has primarily involved voluntary immigration followed by domination, forced segregation and multiple forms of discrimination by the majority culture, as well as fairly high retention of Mexican culture. That Mexican culture is frequently replenished by regular immigration from Mexico is a factor not fully considered in Berry’s model. The slow historical pace of Mexican American adaptation from segregation to integration, a function of minority-majority relations, and frequent immigration flows from Mexico, is also not depicted in Table 1. Still, Berry’s fourfold model provides a basic understanding of how stressful acculturation can be for different U.S. minority groups. Indeed, these four forms of adaptation can be viewed as an acculturative stress gradient in which scenarios of assimilation and integration have been far less stressful forms of adaptation to American society than segregation and certainly marginalization. Such a preliminary conceptualization of acculturative stress can serve as a starting point for understanding a minority group’s level of social stratification as reflected in disparities in SES, health, and so forth.

Berry’s fourfold model has been criticized for its lack of predictive validity with regard to minority adjustment. For example, in a review of correlations between Berry’s four types of adaptation and various indices minority adjustment, taken from 18 of Berry’s Canada-based studies, Rudmin (2003) found that two-thirds of the correlations were nonsignificant, including 28 of 30 between integration and adjustment, a bicultural form of adaptation that Berry and many others view as optimal. Rudmin did find significant correlations between segregation and distress, for minorities with little economic and political power, as well as between marginality and stress. Thus while Rudmin’s review provides only partial support for Berry’s four-fold model, it seems premature to dismiss the model based on this preliminary review. That is, the 18 studies reviewed included very different acculturating minority groups (e.g., Korean, French, Portuguese, various Amerindian groups) and scenarios, and were typically small in sample size (i.e., which ranged from 31 to 150 with an average of 61 participants per study). Furthermore, while such quantitative critiques are useful, a historical and sociological analysis of majority-minority ethnic and race relations in Canada or the United States represents an alternative way of describing predominant forms of acculturation and adjustment.

For example, Organista (2007a) has conducted a historical analysis of major U.S. Latino groups, emphasizing what he considers a generally inverse relation between degree of acculturative stress and adjustment at the population level. Table 2 illustrates the varying historical courses of acculturation for major U.S. Latino populations and varying consequent levels of acculturative stress and adjustment. By expanding the logic of Berry’s three course model of acculturation, Organista’s framework deepens our understanding of why some Latino groups fare worse than others, despite sharing common cultural values and practices. For example, Puerto Ricans generally fare worse on most indicators of adjustment (SES, health, etc.) while Cuban Americans generally fare the best, with Mexican Americans in the middle given their broader variance in acculturation.

OPEN IN VIEWER

Table 2.

Relation Between Acculturation, Degree of Acculturative Stress, and General Level of Adjustment for Major U.S. Latino Groups.

Latino group	Conditions of contact	Degree of conflict	Predominant form of adaptation	Degree of acculturative stress	General level of adjustment
Mexican Americans	Invasion + immigration	Medium	Segregation + integration	Medium	Medium
Cuban Americans	Legal refugees	Low	Segregation + integration + assimilation	Low	High
Puerto Ricans	Colonization	Medium	Segregation	Medium to high	Low to medium
Central Americans	Illegal refugees	Medium	Segregation	Medium	Low

Note: Plus signs signify progression from one predominant form of adjustment to another.

The historical course of acculturation for U.S. Latinos is consistent with their SES, psychosocial and health disparities. For example, Puerto Ricans have experienced a “double legacy of conquest and colonization” followed by racialized discrimination and hypersegregation alongside African Americans in the United States. (Organista, 2007a; p. 21). In contrast, the first wave of predominantly elite, professional, and White Cubans that fled the new Castro regime benefited from a generous open-door refugee policy replete with plentiful programs that maximized their adaptation to United States. Subsequent waves of similarly upper and middle professional class Cubans, most relatives of first wave, experienced the same reception, through the 1960s and 1970s, further consolidated a thriving ethnic enclave in South Florida. While a small wave of poorer and darker Cubans, called “Marielitos” because they sailed from the Cuban Port of Mariel, did receive a harsh reception during the 1980s (i.e., military detention centers, “tent city” encampments and deportations), the majority did eventually receive refugees status and were quickly absorbed into the Miami’s vibrant ethnic enclave (Garcia, 1996).

Thus no health paradox has been documented for Cubans given their initially high social capital, rapid integration and assimilation into U.S. society, and economically successful ethnic enclave. Nor is the paradox in evidence for Puerto Ricans whose historical experience of invasion and colonization in their country of origin, followed by dominance and subordination in the United States, has blurred the distinction of experiences in both locations (i.e., considerable acculturation stress following initial contact with American society followed by protracted conflict through adaptation in both Puerto Rico and the United States.). Furthermore, their status as U.S. citizens has led to much revolving door migration between Puerto Rico and the East Coast, further blurring any clear immigration experience. Hence, the rather profound differences in the acculturation histories of U.S. Latinos comprise the basis of different trajectories of adjustment rooted in varying levels of acculturative stress.

Mental Health and Acculturation

The MHP is strongly supported by major mental health research (Alderete et al., 2000; Alegria et al., 2007, 2008; Vega et al., 1998), especially in mental disorders with rates more sensitive to stress (i.e., depression and anxiety disorders, alcohol and substance use disorders) than disorders with more stable prevalence (e.g., schizophrenia).

The Mexican American Prevalence and Services Survey (MAPSS) conducted by Vega et al. (1998) provides one of the most compelling demonstration of the MHP. These researchers surveyed lifetime prevalence of psychiatric disorders in a stratified random sample of 3,012 adults of Mexican background in Fresno County, California. This population based sample was divided into short-term (less than 13 years) and long-term (13 years or longer in U.S.) immigrants, and U.S. born Mexican Americans. Results revealed a dramatic pattern of positive relations between increasing time in the United States and lifetime prevalence when comparing short-term, long-term immigrants, and U.S. born Mexican Americans for any mental disorder (18.4%, 32.3%, & 48.7%, respectively), any abuse/substance use disorder (9.7%, 14.3%, & 29.3%, respectively), any anxiety disorder (7.6%, 17.1%, & 24.1%, respectively), and any affective disorder (5.9%, 10.8%, & 18.5%, respectively).

However, to better understand the MHP, MAPSS findings were also compared to the National Comorbidity Survey (NCS), and a Mexico City prevalence survey. The NCS was the first national probability survey of mental disorders to contain a representative sample of Latinos, but only English speaking (i.e., more acculturated). The Mexico City survey allowed for a comparison of the mental health profiles of Mexican immigrants in the United States with their counterparts in Mexico in order check for any differences due to immigrant selection bias (i.e., immigrants healthier than nonmigrant counterparts). Comparisons revealed comparable lifetime prevalence rates between U.S. born Mexican Americans, and rates for NCS general U.S. population and English speaking Latinos (i.e., about 50% lifetime prevalence for any disorder assessed). In contrast, the lifetime prevalence rate for any disorder for immigrants in the MAPSS was half the rate of U.S. born Mexican Americans, and comparable to the prevalence rate in the Mexico City survey (25.4, 48.7 & 23.4, respectively). Vega et al. concluded that their results support the MHP but not because of the selective migration hypothesis given the comparable prevalence rates for Mexican immigrants in the MAPSS and their Mexico City counterparts. While this study did rely on time in the United States as the proxy index of acculturation, it examined a countywide representative sample, compared three distinct time periods, and compared data with both national U.S. and Mexico City data in order to better attribute the declining mental health of Mexican Americans to the stressful effects of acculturation experienced at the population level.

While unusually rigorous, the MAPSS still leaves us to study, more in-depth, which aspects of acculturative stress are related to poor mental health outcomes. Such research will likely tap contradictory nuances at the individual level. For example, in their study of sociopsychological predictors of acculturative stress, Miranda & Matheny (2000) found that first generation immigrants face more acculturative stress than their second-generation counterparts who are more likely to grow up speaking English and thus can navigate society more easily. This finding is interesting because mental health outcomes are worse in second generation and English-speaking Mexican Americans as noted above. Thus additional aspects of acculturation, such as the role of ethnic identity development and segmented assimilation, suggest that each generation is likely to face different types and degrees of stress in need of closer investigation (Caplan, 2007).

In light of Cook et al.’s (2009) findings on the role of perceived discrimination in predicting poor mental health outcomes, recent interest in colorism and the role of phenotyping may shed additional light on the perceived discrimination as a moderating factor in the relationship between acculturation and health. The construct of phenotyping and the role of skin color on mental health among Latinos add a dimension largely ignored in the Latino paradox/acculturation literature. For example, Montalvo & Codina (2001) review a variety of studies highlighting the relationship between level of acculturation, phenotype, and life chances. In general, the authors found that lighter skin Mexican immigrants were more likely to rise up the socioeconomic ladder and live outside ethnic barrio enclaves, while darker skinned Mexican Americans were more likely to remain poor and have lower self-esteem. Interestingly, their review also found evidence of an inverse relationship between acculturation and likelihood of self-identifying as “brown.” This trend suggests that time in the United States is likely to be associated with discrimination and increased feelings of “otherness” due to skin color and indigenous features.

As noted by Padilla & Perez (2003), experiences of stigma and discrimination influence acculturation-related choices at the individual level (i.e., how ethnocentric to be, or how much to attempt or integrate or assimilate outside of one’s group). For example, in a qualitative study of the integration experience of first and second generation Mexican immigrant women, Viruell-Fuentes (2007) found that second generation women were more likely to describe experiences of discrimination and stigmatization (also called “othering”) than their first-generation counterparts. These second generation women, who from childhood had more interaction with mainstream society (outside the ethnic enclaves of their first-generation counterparts) described feeling like outsiders, and struggling to know where they “fit in.” As the author concludes, the second generation’s experience of social exclusion, consistent with a segmented assimilation framework, suggests that “‘othering’ processes and discrimination [are] potential pathways through which ‘becoming American’ might erode the health of immigrants and their descendants” (p. 1,525).

The above studies underscore how social contextual and interpersonal factors may be implicated and interrelated, and raise questions regarding how generational status, ethnic identity, phenotyping, perceived discrimination, and family cultural conflict interrelate and may differentially affect mental health and other health outcomes for Mexican Americans adjusting to American society.

Infant Birth Outcomes and Acculturation

Perhaps the most robust support for the MHP comes from research on infant birth outcomes in California. Guendelman and colleagues (Guendelman, 1995; Guendelman & English, 1995; Guendelman et al., 1990; Guendelman, Thornton, Gould, & Hosang, 2006), among others (see, Acevedo-Garcia, Soobader, & Berkman, 2007; Bender & Castro, 2000; Campos et al., 2008; Harley & Eskenazi, 2006; Johnson & Marchi, 2009; Magana & Clark, 1995; Padilla, Boardman, Hummer, & Espitia, 2002; Scribner & Dwyer, 1989; Sherraden & Barrera, 1997; Weigers & Sherraden, 2001) have repeatedly found favorable birth outcomes (i.e., healthy birth weight) in first generation Mexican immigrants that disappear by the second generation. In fact, Guendelman and English (1995) found an increase in low-birth weight and preterm deliveries among Mexican women after just five years of residence in the United States. The authors attribute this increase in unfavorable pregnancy outcomes over time to an increased risk profile among Mexican immigrant that includes higher rates of risky health behaviors such as drinking, smoking and drug use during pregnancy. Longer term residents were also more likely to work in physically strenuous, low-paying jobs throughout pregnancy such as cleaning houses or factory work, potentially increasing their exposure to toxins during activity.

Expanding on the above findings, a review by Rumbaut (1997a) found the following main differences between first and second generation Mexican origin women, with U.S.-born women tending to have in common the following characteristics with U.S.-born Anglo women: worse dietary patterns (higher fat consumption, reduced fiber & fresh produce consumption), increased likelihood for smoking and drinking during pregnancy, and less reported familial support.

Several studies have shed light on components of acculturation affecting birth outcomes. Sherraden & Barrera (1997) applied segmented assimilation theory to the birth outcomes of 41 second generation Mexican American women in Chicago and found that high levels family/social support was more strongly associated with positive birth outcomes and upward assimilation than were maladaptive health behaviors such as smoking and drinking and poor diet. Interestingly, high support families were more likely to visit Mexico and send their daughters there to stay with relatives for vacations. In comparison, women with low family support and poorer birth outcomes knew little to nothing about their parents’ Mexican heritage (i.e., where they were from in Mexico, etc.). These findings suggest that a strong sense of ethnic identity/heritage within a cohesive family unit has an important role in birth outcomes and merit further study.

A study by Johnson & Marchi (2009) found that neighborhood was related to maternal stress and infant birth weight, with English-speaking Mexican women living in ethnic enclaves likely to report higher levels of stress than those living in diverse or white neighborhoods. Although counterintuitive to what the authors had hypothesized based on segmented assimilation theory, the role of discrimination and racism within Latino groups adds yet another dimension to the experience of acculturation. Likely interrelated with issues of phenotyping and ethnic-identity development, Latinos who do not speak Spanish but live in ethnic enclaves may not feel fully accepted by their more Mexican-oriented counterparts. Surprisingly, the authors found a four-fold increase in low-infant-birth weight among English-speaking Mexican women living in ethnic-enclaves versus English-speakers in other neighborhoods (racially/ethnically diverse). This finding is consistent with Berry’s conception of marginality, and helps to fill in some of the details of this form of adaptation that may be related to poor health outcomes.

Type 2 Diabetes and Acculturation

Obesity and Type 2 diabetes Mellitus (DM2) are particularly interesting health problems to explore in light of the current state of acculturation theory. Among Mexican Americans, the risk of diagnosed diabetes is 87% higher than for non-Hispanic Whites, and is 66% higher among Latinos in general (National Diabetes Fact Sheet, 2011). Latinos generally suffer more complications and DM2-related amputations and deaths than non-Hispanic Whites, and rates of diabetes are higher among Mexican Americans and Puerto Ricans as compared to non-Hispanic Whites and Cuban Americans (Organista, 2007b). However, the role of acculturation in the DM2 epidemic is not well understood, partly due to the tendency to group all Latinos together, or to use proxy measures of acculturation across studies (Antecol & Bedard, 2006; Abraído-Lanza, Chao, & Flórez, 2005).

Conflicting findings about behavior change over time underscore the complex determinants of diabetic health outcomes among Mexican Americans. For example, Mexican American dietary change with increased length of residency in the United States has been the source of much research interest, but studies have shown inconsistent results in regards to explaining rising obesity trends among Mexican Americans because dietary changes have been documented as both positive and negative (Abraido-Lanza et al., 2006; Akresh, 2007; Caballero, 2005; Crespo, Smith, Carter-Pokras, & Andersen, 2001; Perez-Escamilla & Putnik, 2007; Schulz et al., 2005; Yeh, Viladrich, Bruning, & Roye, 2009). A literature review by Antecol & Bedard (2006) showed that studies looking at the relationship between acculturation and factors contributing to DM2 (including obesity, diet, alcohol use, etc.) are filled with contradictory results depending on which measure of acculturation was used.

In a dated yet insightful study, Hazuda, Haffner, Stern, & Eifler (1988) used data from the San Antonio Heart Study to measure the role of acculturation and socioeconomic status on obesity and diabetes among Mexican Americans. The study represents one of a few that utilized multiple acculturation indices to help account for the nonlinear process of acculturation and focuses specifically on Mexican Americans (as opposed to all Latinos). The authors found that acculturation was negatively correlated with prevalence of obesity and DM2 in both men and women. That is, less acculturated men and women of low SES showed higher prevalence of DM2 and obesity than their more acculturated and higher SES counterparts.

Interestingly, Hazuda and colleagues (1988) also found that rates of diabetes correlated with Indian heritage (Mexican Americans with darker skin and Indian heritage were also more segregated in lower SES barrios), suggesting a genetic predisposition. Results call into question the relationship between SES, ethnoracial appearance and stress potentially exacerbating genetic predisposition to DM2. While the authors did not explore this relationship, the work of Montalvo & Codina (2001) suggests interactions among these factors that merit further exploration. The results from Hazuda and colleagues’ study seem to contradict the MHP, and raise questions about the complexities of contributing factors to diabetes. Caballero (2005) discusses the role of the following factors among Latino populations: genetic predisposition, social and cultural factors of immigration, level of acculturation, body image, depression, education-level, fears, general level of familial integration and support, and health literacy. Biopsychosocial factors experienced by Latino immigrants may also affect adherence to treatment regimens, and the risk of complications and mortality due to DM2.

Cultural factors may also contribute to lower obesity rates among more acculturated, higher SES Latinos. A qualitative focus-group study by Diaz, Mainous, & Pope (2006) found conflicting views regarding ideal body weight and health as dictated by U.S. culture versus traditional Latino culture. One male interviewee in the Diaz study summarized his experience of cultural conflict: “In the Latin culture, from when you are a child, you hear that being fat is ‘saludable’ [healthy]. That is wrong, but in the culture, if you are not fat, then you are sick. You have to convince the family and the individuals themselves that being fat does not mean you are ‘saludable’” (p. 329). While the research clearly shows that DM2 and obesity is a problem of epidemic proportions among Mexican Americans and Puerto Ricans, our understanding of the epidemic and the specific role of acculturation remains inconclusive due to current limitations of the construct.

Taken together, the above studies underscore the need to enhance the role of social determinants in future of acculturation research on DM2. To this end, Castro and colleagues (2009) have recommended the use of an ecodevelopmental framework for understanding factors contributing to the development of DM2 among Latinos. Specifically, they suggest that traditional systems models which account for contextual factors at multiple levels (e.g., individual-level, family-level, community-level, etc.) should be expanded to include a temporal developmental perspective, wherein key life milestones (e.g., transition to adolescence, adulthood or old age) would better map the ways in which contextual factors interact over the life course to affect the development of DM2. For example, the role of familial dietary norms (often associated with acculturative change) may differentially affect Mexican Americans at different life stages (e.g., during childhood and adolescence versus young adulthood). The authors further suggest that an ecodevelopmental framework that accounts for gender, age, socioeconomic status, acculturation, and other social and cultural factors over time, and at multiple levels (micro, meso, and macro), can help us better understand the development of DM2, and will ultimately allow us to more effectively intervene to prevent DM2 disparities among Latinos and other populations.