Acculturation and obesity among migrant populations in high income countries – a systematic review

The underlying dynamics of the complex relationship between migration and migrant obesity may be further clarified with a theoretical perspective which takes into account the forces of acculturation and enculturation [27‐31]. As indicated above, the majority of people who migrate from low-income to high-income countries eventually adopt obesogenic behaviours, experience weight gain, and record higher body weights than their local counterparts. This change occurs as their immersion into the host culture increases over time [5, 6, 24] and is indicative of the broader and more complex process of acculturation – that is, the gradual exchange between immigrants’ original attitudes and behaviour and those of the host culture [32]. This phenomenon has consistently been associated with negative consequences for migrants, including added stressors, fewer coping resources, and poorer mental health [33].

In contrast some migrants may avoid these adverse effects of immigration, by assimilating less to the host culture and retaining more of their traditional heritage. This is termed enculturation[34], and often appears to be beneficial to migrants in terms of social and health effects. In particular, this phenomenon is commonly associated with the maintenance of several protective factors, including social/familial support, shared sense of ethnic identity, and continued practice of traditional values [35].

Thus, acculturation and enculturation are intertwined and representative of important psychosocial factors in the migration process - not least in terms of several health-risks. Indeed it would appear that acculturation is generally associated with adverse health-effects, and enculturation the opposite. Hence, in order to gain a clearer understanding of the impact of migration on migrant health, these two factors must be included in the analysis.

While past research has examined the relationship betweenx migrant health and acculturation, most studies have employed surrogate measures – such as for example length of residency in the host culture, place of birth or preferred language used by immigrants [36, 37] – rather than using standardised acculturation scales. Due to such lack of reliable and consistent methods for quantifying the degree of acculturation, comparative study and meta-analyses have been difficult, if not impossible, leaving a considerable gap in the literature on this topic. Furthermore, the fact that only few studies have taken enculturation into account, has meant that the research in this field has generally failed to separate the differential impact of acculturation and enculturation [38, 39]. This shortcoming represents another limitation of research in this area.

The two most commonly used standardised models of acculturation are uni-directional models (UDM) and bi-directional models (BDM). UDMs are based on the assumption that the maintenance of the original migrant heritage is not possible among those immigrants who become involved in the mainstream host society. Thus, acculturation is inevitably accompanied by a weakening of identification with heritage, and can be represented as a linear change over time.

Similarly, BDMs suggest that the relationship with both original heritage and the new host culture plays an important role in the process of acculturation [36]. This model allows for interaction effects of the original and the new cultures by assessing an individual’s cultural identity and participation in the larger society [40].

To systematically review the evidence on the relationship between acculturation - as measured with a standardized acculturation scale (i.e. UDM or BDM) - and overweight/obesity among adult migrants from low/middle countries to high income countries.

This review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). This system has been developed to provide clear, practical and systematic guidance in researching and writing systematic reviews.

A comprehensive search of the following databases was conducted during May 2011: Academic Search Complete (EBSCOhost); Academic Search Premier (EBSCOhost); CINAHL with Full Text (EBSCOhost); Global Health (EBSCOhost); MEDLINE with Full Text (EBSCOhost); and PsycINFO (EBSCOhost).

The search terms used were based on MeSH keywords for ‘obesity’, ‘enculturation’ and ‘acculturation’. Searches were conducted by inserting the following terms simultaneously into the EBSCOhost database. Within term groups (that is Obesity terms and Acculturation terms) search terms were divided with ‘or’, while the two term groups were connected with ‘and’. The search strategy including databases, keywords, and limits is defined in Table 1.

Articles obtained in the search were appraised in terms of relevance, quality and validity in three rounds. In the first round, articles were included or excluded based on their title. In the second round, articles were retained or excluded after reviewing their abstracts. In the third round, the full-text version of each of the remaining articles was obtained [41] and reviewed separately by two researchers. To determine the quality of the studies reviewed, quality assessment tools were utilised. The strength of the quantitative studies was ranked using the Quality Assessment Tool for Quantitative Studies developed by the Effective Public Health Practice Project from the Ontario Ministry of Health in Canada [42], and recommended by other authors [43, 44]. Studies were ranked as high, moderate or low in quality.

A total of 1135 papers were identified in the initial search. The vast majority of these were rejected on the basis of one or more of the following factors: the paper focussed only on acculturation, immigration, overweight or obesity rather than any relationship between the variables; the paper did not use a standardized (UDM or BDM) scale to measure acculturation; the paper did not include original, empirical research (commentaries, book reviews, ‘grey’ literature, etc.); or a combination of the above (see Figure 1). Overall, nine papers met the inclusion criteria (see Table 2).

All of the studies were published in the 1990s and 2000s, and all were conducted in the US. Eight studies utilised cross-sectional designs, while one was longitudinal (see Table 2). Of the studies reviewed, participant mean age was 41 years (SD = 15), and the male: female ratio across all studies was 32:68. The focus of the articles included was limited to immigrant and refugee resident populations, and thus, temporary residents were not represented in any of the studied populations. Study sample ethnicity included, Mexican-Americans [45, 46, 50, 51], Hmong (Thailand and Laos) US residents [49]; Korean-Americans [52], Puerto-Rican Americans [48, 51], and Soviet-American women [53]. Six studies used BDM-scales to measure acculturation [45, 46, 49, 50, 52, 53] and three studies used UDM-scales [47, 48, 51]. Two studies used self-report measures for the BMI measure [46, 52].

Table 2 presents the nine papers’ qualitative evaluations using Quality Assessment Tool for Quantitative Studies [46, 50, 53]. Three of the nine studies were classified as a ‘strong’ quality ranking while the remaining six were of ‘moderate’ quality [45, 47‐49, 51, 52]. Studies moderate in quality were so determined due to small sample sizes, data collection weaknesses, participant attrition and participant representation.

The acculturation scales used in the reviewed studies are presented in Table 2. The scales were valid and reliable, and all comprised items gauging degree of acculturation/enculturation by typically focusing on language use, use of media in host country, ethnic social relations, social networks, lifestyle, values and/or attitudes. The studies employed both UDM and BDM scales, making it difficult to compare specific results across these studies using a common metric. However, the standardization, validity and reliability of the scales used, ultimately facilitated comparability and synthesis of the results obtained with each type of scale.

Of the nine studies reviewed, seven reported overall positive associations between the degree of acculturation and body weight variables [45‐50, 52]. The conclusions drawn from these studies generally complemented one another. Three studies [50, 51, 53] found negative associations between higher degree of acculturation and obesity in women. Lee et al. [52] found no significant relationship between acculturation and obesity in their female sub-sample, while more acculturated males recorded higher BMI.

In their study on US-born and Mexico-born Mexican-American adults living in the US, Barcenas et al. [46] indicated significant differences in the relationships between acculturation level and BMI, gender, and birthplace. In Mexican-born participants longer residency in the US was associated with higher BMI. Women residing in the US for more than 15 years recorded a mean BMI of 2.38 kg/m² (β (adjusted) =2.12, 95% CI, 1.53-2.72, p < .0001) higher than those who had lived in the US for less than five years. Similarly, men living in the US for more than 15 years recorded an average BMI of 1.10 kg/m² (β (adjusted) = 1.47, 95% CI, 0.59-2.34, p < .001) higher than residents of less than five years. Furthermore, the risk of obesity in men and women with a low degree of acculturation increased with every additional year of residence in the US by 2% (OR = 1.02, 95% CI, 1.00-1.03) and 1% (OR = 1.01, 95% CI, 1.01-1.02) respectively. For highly acculturated men and women the increase in risk was 4% (OR = 1.04, 95%CI, 1.01-1.07) and 3% (OR = 1.03, 95% CI, 1.00-1.05) respectively. In US-born Mexican-Americans, high acculturation and birthplace accounted for 6% and 25% of risk of obesity in men and women respectively.

These general findings of positive correlations between obesity and acculturation in Mexican-Americans, were further supported in the study by Ahluwalia et al., [45] where comparable general associations between the degree of acculturation and BMI were established. Further, Khan et al. [51] found that 2nd and 3rd generation Mexican-Americans had higher BMIs than 1st generation migrants.

A similar study was conducted by Hazuda et al. [50] who established a more ambiguous and complex relationship between acculturation and obesity in Mexican-Americans. In this study, overweight was assessed using BMI, sub scapular-to-triceps skin-fold thickness (SUBTR) ratio, and waist-to-hip ratio (WHR). Both acculturation and socio-economic status (SES) were associated with anthropometry measurements. Specifically, low SES was associated with a high WHR (p < .001), and high scores on the ‘family attitude’ subscale of acculturation was associated with a high SUBTR-ratio (p < 0.01). Among females, low SES was associated with high BMI, SUBTR-ratio and WHR (p < .01 – p < .001), but, interestingly, high scores on the ‘functional integration’ acculturation subscale were associated with lower BMI and SUBTR-ratios (p < .01 – p < .0001). This apparently protective effect for Mexican-American women was also evident in Khan et al. [51], where higher acculturation was associated with lower BMI for Mexican-American women (β = −.56, p < .01) but not for males nor either gender in the other ethnicities examined (Puerto Rican and Cuban).

Other studies examined the impact of acculturation on health factors such as smoking, physical activity, fat intake, and BMI, and have included other ethnicities with similar results. For example, Bertera et al. [47] found that acculturation was positively associated with obesity in Salvadoran immigrants in the US, accounting (in combination with other variables such as gender and SES) for 45% of the variance in immigrant obesity (p < .01-.001). Similarly, Fitzgerald et al. [48] discovered that less acculturated Puerto Ricans in the US were 54% less likely to be obese than their more acculturated counterparts. In yet other research on Hmong-American immigrants, higher degree of acculturation also correlated with higher BMI [49]. This relationship was mainly attributed to the changes in diet and physical activity following the adaptation to an environment with higher average income [49].

Comparable effects have also been found in studies among Korean-Americans [52] whereby more acculturated men (but not women) had a higher BMI than their enculturated counterparts, with a mean difference in BMI of 1.7kg/m² (p < .05) [52]. This effect persisted when controlling for age, income, education, working status, and smoking as well as physical activity, fat intake and parents’ body size. In addition, and perhaps paradoxically, acculturated men and women were more likely to engage in light physical activity than enculturated men and women (men OR = 4.34, p < .005; women OR = 7.38, p < .005) [52].

The final paper included in this review [53] reported on a longitudinal study which examined the relationship between acculturation and cardiovascular disease risk factors in Soviet Union-born middle-aged women in the US. Similar to some of the other research reviewed, this study established a significant inverse correlation between acculturation score and obesity in women. Specifically, women with higher acculturation scores at baseline predicted lower changes in BMI over the following ten years (r = −.14, p = .05). Similarly, higher enculturation scores predicted higher changes in mean BMI and waist circumference (r = .25, p = .01; r = .34, p = .01, respectively) in the following decade [53].

Five of the nine studies included measurement of SES [47, 48, 50‐52]. Three found no significant correlations between SES and obesity [47, 51, 52], while Fitzgerald et al. [48] established an inverse correlation between these two variables, and Hazuda et al. [50] found both positive and negative correlations for men and women, respectively.

A strength of this review relates to the chief goal of the paper which was to identify and review the research which has used valid and comparable acculturation scales to measure the relationship between acculturation and obesity in migrants to high-income countries. While only nine studies met the inclusion criteria for this paper, the fact that all of the reviewed research employed acculturation scales in their methodology effectively facilitated comparison and synthesis of the main themes and variations emanating from this research. This has not been done before and has thus far represented a gap in the literature on this topic. Nonetheless, although the studies included in our review used validated acculturation scales they varied in terms of number of items and domains covered. For example, some studies [45‐48, 51] included one dimension that focused on language acculturation rather than a broader array of behaviours related to the migration and settlement process. Others included questions regarding language use and preference, social connections, and overall eating patterns [49], while others still focused on functional integration (i.e. adoption of the values, attitudes, and behaviour of the host society) [50], structural acculturation [52], and language, identity, and behaviour [53]. In addition, when examining the relationship between acculturation and obesity, some studies used self-reported BMI measures whereas other used measured BMI, making it difficult to draw conclusions across studies. Thus, the variation in acculturation domains and measurement identified in this review emphasises the urgent need for a standardized international acculturation scale.

Another strength of this review centres on the broader research methodology of the included studies. Eight of them employed a cross-sectional design. While the correlational nature of this methodology prohibits a statistical assessment of cause and effect, it can be more readily applied to large representative samples as well as take into account a number of potentially confounding factors. Further, given the specific nature of the variables measured in the reviewed studies, it is still possible to hypothesise with relative certainty on the direction of the associations found despite the nature of cross-sectional research designs. That is, it is much more likely that acculturation has an impact on body weight than vice versa, and while several other factors may mediate or moderate the relationship, this direction of the association remains the most plausible.

A major limitation to the review, relates to the fact that all of the studies focussed on immigrants in the US – a country which is among the ten most obese nations in the world [64]. In light of this, it is conceivable that obesity may not be related to acculturation in general, but rather to the specific culture to which immigrants acculturate. That is, immigration and acculturation to a country where obesity is not a prominent health issue may not increase migrant BMI. More research examining immigration to countries other than the US is therefore needed for more generalisable conclusions.

Another limitation relates to acculturation scales used. The introduction to this study identified the bi-directional model of acculturation as being more complex and sophisticated than the uni-dimensional model. The current review included studies which used both types of scales, making it difficult to compare specific results across these studies using a common metric. There is also some debate focusing on socioeconomic status and its possible role on the relationship between acculturation and obesity. However, not all the nine studies took this into account and therefore lack of adequate control for SES perhaps partly explain the distinct findings among women.

Finally, the review included studies which used both measured and self-reported BMI (Barcenas et al., 2007; Lee et al., 2000). This comprises a limitation as the accuracy of self-reported BMIs may be questionable.

¹The authors define high, middle and low income countries based on United Nation criteria.