SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion

Objective

US Hispanics bear a disproportionate burden of excess body weight and associated morbidity, making obesity prevention and control in this population a public health priority. Development of culturally appropriate weight loss interventions requires understanding the influence of acculturation on attitudes and behavior related to obesity and weight loss.

Design and Methods

Using data from the National Health and Nutrition Examination Survey (2001–2004), we analyzed the associations of country of birth and language usage with weight perception, weight dissatisfaction, intention to lose weight, and intended weight loss of clinical significance among obese Hispanic adults (>20 years of age; unweighted n = 675) in multivariable logistic regression analyses.

Results

US birth was significantly associated with correct weight perception (odds ratio [OR], 4.45; 95% confidence interval [CI], 2.16–9.17) and weight dissatisfaction (OR = 2.95; 95% CI = 1.63–5.35). Compared with Spanish-speaking only, obese Hispanics, bilingual, obese Hispanics were more likely to attempt weight loss (OR = 2.78; 95% CI = 1.43–5.40) and achieve clinically significant weight loss (OR = 3.94; 95% CI = 1.51–10.3).

Conclusions

The findings suggest that lack of acculturation as measured by foreign birth and limited English proficiency may contribute to within-ethnicity health inequality among obese Hispanics, even after controlling for a variety of sociodemographic characteristics, including education and income, differential healthcare access and use, and health status. Further studies should examine the underlying mechanisms, thereby informing culturally competent intervention strategies targeted at obese US Hispanics.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion

The Latino/Hispanic population is one of the largest minority cultures in the United States. According to the 2010 US Census, an estimated 50.5 million individuals self-identified as Latino or Hispanic (heretoafter, referred to as Hispanic to be consistent with the term used in the data source for the study) comprise 16.3% of the population ([1]). It was projected that the Hispanic population would reach 102.6 million (24.4%) by 2050 ([2]). Hispanic adults have a higher age-adjusted prevalence of obesity than non-Hispanic whites (NHWs), that is, 39.1% versus 34.3% ([3]) and are affected disproportionately by the adverse health consequences of obesity, such as hyperlipidemia, metabolic syndrome, and type 2 diabetes ([4, 5]).

Culturally appropriate weight loss interventions are critically needed to effectively target the obese Hispanic population, thereby reducing the disproportionately high obesity-related disease burden in this population. This requires a better understanding of the attitudes and behavior among obese Hispanics that are related to body size and weight loss, such as weight perception, weight dissatisfaction, and weight loss intention and success. In particular, research is needed on how acculturation influences these overweight-related attitudes and behavior so that effective weight loss interventions tailored on the basis of acculturation status can be developed—a priority identified in the 2010 National Heart, Lung, and Blood Institute workshop on future research directions to prevent cardiovascular disease and its risk factors among Hispanics/Latinos ([6]).

Acculturation is a multidimensional process of cultural change an immigrant experiences when integrating into the majority culture. There is currently no gold standard measure of acculturation; several proxies exist that may assess different constructs of the multidimensional process ([7]). Prior studies have suggested that the relationship between acculturation and obesity in Hispanics may depend on the specific acculturation measure. For example, several studies have shown that the risk of obesity is greater among US-born Hispanics (vs. the foreign-born) and is positively associated with length of US residency among immigrants ([10]). On the other hand, other studies have found that less acculturation as indicated by low English usage in home and social settings is associated with increased odds of obesity and its comorbidities ([14]), but conflicting evidence also exists ([18, 19]). Further, very few studies have investigated the association of any acculturation measures with overweight-related attitudes or behavior ([19]).

The objective of this study was to examine the association of acculturation according to country of birth and language usage with four overweight-related outcomes: weight perception, weight dissatisfaction, stated intention of weight loss, and intentional weight loss of clinical significance (≥5%), among obese Hispanic adults in the United States. We hypothesized that more acculturated, obese Hispanics would be more likely than their less acculturated counterparts to perceive their weight status correctly, be dissatisfied with current weight, intend to lose weight, and achieve clinically significant weight loss.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion

Data source

The National Health and Nutrition Examination Survey (NHANES) is a national survey, involving household interviews and medical examinations conducted by the National Center for Health Statistics (NCHS). It was changed from a periodic annual survey to a continuous annual survey in 1999, and the continuous NHANES survey data have been released in two-year increments for public use. The NHANES uses a stratified, multistage probability cluster sampling design and weighting methodology that allows unbiased national estimates to be produced for the civilian, noninstitutionalized US population. NHANES sample weights adjust for unequal probabilities of selection, nonresponse, and planned oversampling (of low-income persons, adolescents 12-19 years, persons 60+ years of age, African Americans and Mexican Americans). The NCHS Ethics Review Board approved the protocols for the NHANES, including a waiver of the requirement for informed consent of participating respondents. Detailed documentation of the NHANES survey and public use data files can be found at http://www.cdc.gov/nchs/nhanes.htm.

Study population

A total of 3,293 and 2,860 Hispanics of all ages were included in NHANES 2001-2002 and 2003-2004, respectively. From this population, we excluded Hispanics who were younger than 20 years (n = 3,666) or pregnant (n = 159) or whose body mass index (BMI) was missing (n = 214) or less than 30 kg/m2 (n = 1,439) based on measured height and weight at the time of the NHANES medical examination. As a result, the analysis sample included 675 (n = 328 for 2001-2002 and n = 347 for 2003-2004) nonpregnant, obese Hispanics aged 20 years or older.

Measurements

Independent variables: Two measures of acculturation

Two constructs available in the NHANES 2001-2004 were country of birth and language usage (Phi correlation coefficient = 0.69), which have been commonly used in health research as proxies for migration and level of integration with mainstream American culture, respectively ([7, 16, 17]).

Country of birth

This was categorized as either US or foreign born. The latter encompassed all respondents who answered as born in Mexico or elsewhere (unspecified) than the United States.

Language usage

Acculturation by language usage was measured by the five-item Short Acculturation Scale (SAS) questionnaire that asks about language usage in various contexts, including overall use, as a child, in the home, when thinking, and with friends. Answers were based on a five-point scale, rated as (1) “only Spanish,” (2) “more Spanish than English,” (3) “both equally,” (4) “more English than Spanish,” and (5) “only English.” The unweighted sum of an individual's five-item responses was the person's language acculturation scale score, ranging between 5 and 25; the higher the score, the greater the level of language acculturation. Scores could not be calculated for 11 of the 675 subjects because of missing item responses.

Dependent variables: Four overweight-related outcome measures
Weight perception

Respondents were asked whether they perceived their weight to be “overweight,” “underweight,” or “about the right weight.” Given that we focused on individuals with a measured BMI ≥ 30 kg/m2, a correct weight perception was defined as a respondent self-identifying as “overweight,” whereas an incorrect weight perception was when the respondent answered “about the right weight” or “underweight.”

Weight dissatisfaction

Weight dissatisfaction was defined by a desire to change one's body weight. Respondents were asked if they would like to “weigh more,” “weigh less,” or “stay about the same.” If the respondent's answer was “weigh less,” they were regarded as expressing dissatisfaction with their current body weight.

Weight loss intention

Intentional weight loss was defined by two separate questions to which respondents answered that they had attempted weight loss within the past year whether or not they actually lost any weight or that they had lost weight within the past year and that the loss was intentional.

Intentional weight loss of clinical significance

For those who indicated intention to lose weight, a respondent was classified as achieving a weight loss of clinical significance if the measured weight at the time of the NHANES medical examination at least 5% less than the person's self-reported weight 12 months ago.

Covariates
Sociodemographic characteristics

We included age (20-39, 40-59, and ≥60 years), sex (male and female), and Hispanic origin (Mexican American and other Hispanic). The sample size was too small to further subcategorize other Hispanic. We categorized participants by annual family income (<$20,000 or ≥$20,000), highest achieved education level (<high school, high school, or general equivalency diploma [GED], or ≥ college), and employment status (currently employed or unemployed).

Healthcare access and use

We categorized participants by health insurance (yes/no) and routine place of care (yes/no). Participants also reported the number of encounters that they had had with a health-care provider in office, clinic, hospital emergency room, or home during the past 12 months.

Health status

Self-rated health status was classified as excellent, very good, or good, and fair or poor. Smoking status included non-smoker, current smoker, or past smoker. The presence of obesity-related co-morbidities referred to medical conditions that a respondent had ever been told by a doctor or other health professional that they had or clinical diagnosis such as diabetes, cardiovascular disease, hypertension, and hyperlipidemia. In addition to a self-reported doctor diagnosis, hypertension was also defined by either a measured systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg, or self-reported use of antihypertensive medication. Similarly, the definition of hyperlipidemia also included either a total cholesterol concentration ≥240 mg/dL or fasting low-density lipoprotein cholesterol concentration ≥160 mg/dL or self-reported use of lipid-lowering therapy.

Statistical analysis

Chi-square tests (PROC SURVEYFREQ) examined isolated associations of each acculturation measure with the four outcome measures and covariates. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) from the logistic regression procedure (PROC SURVEYLOGISTIC) were used to determine the significance of the independent association of an outcome measure with each acculturation measure while controlling for all the covariates. The two measures of acculturation (country of birth and language usage) were modeled separately.

All analyses were conducted in SAS, version 9.2 (SAS Institute, Cary, North Carolina) and took account of the complex sampling design and sample weights of the NHANES. Statistical significance was set at P < 0.05 (two-sided).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion

Associations of acculturation with covariates

The study sample included 328 US-born and 347 foreign-born Hispanic obese adults (305 born in Mexico and 42 born elsewhere) (Table 1). Acculturation as defined by language usage showed that 23% of the respondents use Spanish only, 51% use both Spanish and English of varying degree, and 26% use nearly all or exclusively English.

Table 1. Associations of acculturation with covariatesa
 Born in the United StatesLanguage acculturationa
    Low   High 
 NoYesP-value56-1011-1516-2021-25P-value
  1. Row percentages.

  2. a

    Acculturation by language is based on a five-item short acculturation scale, with scores ranging from 5 to 25; higher scores correspond to greater levels of language acculturation. See the Methods section for more details.

Unweighted sample size347328N/A196100117126125N/A
Weighted sample size (10 thousands)689702N/A302185227273344N/A
Sociodemographic characteristics        
Ethnicity (%)  0.25     0.003
Mexican American52.247.8 28.215.215.520.320.8 
Other Hispanic44.855.2 12.011.419.920.935.7 
Age (%)  0.13     0.001
20-3944.755.3 20.910.116.020.432.6 
40-5954.845.2 23.917.416.623.019.1 
60+57.842.2 28.523.024.112.312.1 
Sex (%)  0.02     <0.001
Male56.243.8 17.118.822.424.017.6 
Female44.455.6 27.010.112.917.832.2 
Education (%)  <0.001     <0.001
Less than high school74.125.9 40.822.516.310.110.4 
High school graduate36.064.0 14.89.720.226.029.3 
College30.469.6 5.45.915.929.942.9 
Family income (%)  0.41     <0.001
<$20,00053.446.6 35.711.217.315.919.9 
≥$20,00049.051.0 17.315.417.122.128.2 
Employment status (%)  0.31     <0.001
Unemployed52.747.3 33.311.917.114.023.7 
Currently employed47.952.1 17.114.917.024.027.0 
Healthcare access and use         
Health insurance (%)  <0.001     <0.001
No67.532.5 37.212.521.916.911.4 
Yes39.360.7 13.315.313.621.636.2 
Routine place of care (%)  0.007     0.002
No62.737.3 35.09.220.517.517.8 
Yes44.955.1 18.515.515.821.628.6 
Number of encounters with health care provider in office, ER, clinic or home (%)  <0.001     <0.001
065.134.9 31.612.523.217.715.1 
143.856.3 21.512.620.712.332.8 
2-352.647.4 24.919.712.327.415.7 
4-941.558.5 11.614.413.822.737.5 
10-1241.758.3 22.75.713.423.135.2 
≥1322.177.9 16.37.513.818.344.2 
Health status         
Self-rated health status (%)  0.39     0.001
Fair or poor54.046.0 30.017.115.110.926.9 
Excellent/very good/good48.052.0 17.413.618.524.825.8 
Smoking status (%)  0.15     <0.001
Current smoker42.857.2 18.27.025.320.728.6 
Past smoker50.549.5 20.818.15.121.234.9 
Never smoker52.347.7 25.715.517.720.220.9 
Obesity-related comorbidities (%)  0.51     <0.001
No48.052.0 22.89.217.018.632.5 
Yes51.049.0 22.618.617.122.419.3 

Greater acculturation by both measures was associated with female sex (P = 0.02 for US birth; P < 0.001 for more English usage), more education (P < 0.001 for both), having health insurance (P < 0.001 for both), having a routine place of care (P = 0.007; P = 0.002), and more frequent encounters with a health care provider (P < 0.001 for both). In addition, more English usage was associated with non-Mexican Hispanic origin (P = 0.003), younger age (P = 0.001), higher income (P < 0.001), being currently employed (P < 0.001), better self-rated health status (P = 0.001), past or current cigarette use (P < 0.001), and the absence of obesity-related comorbidities (P < 0.001).

Associations of overweight-related outcomes with acculturation and covariates

Greater acculturation by both measures was associated with higher percentages of respondents with correct weight perception (P < 0.001 for US birth; P = 0.002 for more English usage) and weight dissatisfaction (P < 0.001 for both measures) (Table 2). Additionally, more English usage was associated with higher percentages of respondents who intended to lose weight (P = 0.01) and achieved clinically significant weight loss (P = 0.03).

Table 2. Associations of weight-related outcomes with acculturation and covariatesa
 Weight perceptionWeight dissatisfaction
 OverweightAbout the right weight/underweightP-valueDissatisfiedSatisfiedP-value
Unweighted sample size59279N/A60173N/A
Weighted sample size (10 thousands)1,233156N/A1,279112N/A
Born in the United States (%)  <.0.001  <.0.001
No84.215.8 89.110.9 
Yes93.36.7 94.85.2 
Language acculturation (%)  0.002  <0.001
584.215.8 88.711.3 
6-1092.97.1 94.25.8 
11-1582.417.6 81.418.6 
16-2088.811.2 93.96.1 
21-2597.22.8 98.02.0 
Sociodemographic characteristics      
Ethnicity (%)  0.10  <0.001
Mexican American87.712.3 89.410.6 
Other Hispanic90.59.5 96.53.5 
Age (%)  0.98  0.05
20-3988.511.5 92.57.5 
40-5989.210.8 92.37.7 
60+88.411.6 87.312.7 
Sex (%)  <0.001  <0.001
Male81.019.0 86.513.5 
Female94.65.4 96.13.9 
Education (%)  0.10  0.04
Less than high school86.313.7 88.511.5 
High school graduate93.96.1 95.05.0 
College89.610.4 94.06.0 
Family income (%)  0.01  <0.001
<$20,00084.415.6 86.014.0 
≥$20,00089.610.4 94.15.9 
Employment status (%)  0.56  0.50
Unemployed89.710.3 90.99.1 
Currently employed88.311.7 92.57.5 
Healthcare access and use      
Health insurance (%)  0.07  0.14
No84.315.7 89.610.4 
Yes91.09.0 93.16.9 
Routine place of care (%)  0.02  0.10
No80.619.4 88.211.8 
Yes91.68.4 93.36.7 
Number of encounters with health care provider in office, ER, clinic or home (%)  0.12  <0.001
085.214.8 88.012.0 
183.716.3 86.313.7 
2-391.68.4 96.04.0 
4-992.87.2 96.33.7 
10-1296.04.0 98.11.9 
≥1385.015.0 86.413.6 
Health status      
Self-rated health status (%)  0.87  0.26
Fair or poor89.011.0 90.69.4 
Excellent/ very good/good89.310.7 92.77.3 
Smoking status (%)  0.53  0.73
Current smoker85.914.1 91.09.0 
Past smoker91.58.5 93.56.5 
Never smoker89.110.9 91.88.2 
Obesity-related comorbidities (%)  0.59  0.045
No88.012.0 90.39.7 
Yes89.510.5 93.66.4 
 Intention to lose weightIntentional weight loss of clinical significance
 YesNoP-valueYesNoP-value
  1. Row percentage.

Unweighted sample size370304N/A86588N/A
Weighted sample size (10 thousands)804587N/A1661,225N/A
Born in the United States (%)  0.63  0.10
No54.245.8 9.390.7 
Yes61.438.6 14.585.5 
Language acculturation (%)  0.01  0.03
542.257.8 6.393.7 
6-1058.741.3 11.388.7 
11-1561.738.3 11.688.4 
16-2064.036.0 19.380.7 
21-2565.234.8 12.687.4 
Sociodemographic characteristics      
Ethnicity (%)  0.51  0.36
Mexican American55.344.7 12.687.4 
Other Hispanic62.237.8 10.789.3 
Age (%)  0.68  0.14
20-3954.245.8 9.890.2 
40-5964.435.6 14.785.3 
60+54.245.8 13.786.3 
Sex (%)  0.02  0.84
Male51.348.7 12.287.8 
Female62.737.3 11.788.3 
Education (%)  <0.001  0.01
Less than high school47.352.7 6.193.9 
High school graduate64.835.2 11.288.8 
College66.233.8 16.683.4 
Family income (%)  0.49  0.52
<$20,00055.344.7 10.989.1 
≥$20,00060.139.9 12.787.3 
Employment status (%)  0.59  0.05
Unemployed57.142.9 8.891.2 
Currently employed58.241.8 13.586.5 
Healthcare access and use      
Health insurance (%)  0.004  0.99
No49.350.7 12.487.6 
Yes65.434.6 12.387.7 
Routine place of care (%)  0.07  0.05
No39.460.6 8.291.8 
Yes64.335.7 13.286.8 
Number of encounters with health care provider in office, ER, clinic or home (%)  0.008  0.16
048.351.7 8.791.3 
147.152.9 6.993.1 
2-369.730.3 14.185.9 
4-959.540.5 16.283.8 
10-1273.326.8 7.492.6 
≥1356.943.1 18.481.6 
Health status      
Self-rated health status (%)  0.80  0.64
Fair or poor58.241.8 11.588.5 
Excellent/ very good/good59.940.1 12.887.2 
Smoking status (%)  0.43  0.72
Current smoker60.539.5 12.687.4 
Past smoker61.438.6 13.886.2 
Never smoker55.244.8 10.989.1 
Obesity-related comorbidities (%)  0.10  0.74
No53.446.6 11.388.7 
Yes62.038.0 12.587.5 

Obese Hispanics other than Mexican Americans were more likely to be dissatisfied with their weight (P < 0.001). Obese Hispanic women were more likely to express correct weight perception (P <0.001), weight dissatisfaction (P <0.001), and intention to lose weight (P = 0.02) than obese Hispanic men, but they were no more likely to achieve clinically significant weight loss. Obese Hispanics with at least high school education were more likely to be dissatisfied with their current weight (P = 0.04), intend to lose weight (P < 0.001), and achieve clinically significant weight loss (P = 0.01). Higher income was associated with greater percentages of respondents with correct weight perception (P = 0.01) and weight dissatisfaction (P < 0.001). A high percentage of insured obese Hispanics said that they had intended to lose weight within the past year (P = 0.004). Obese Hispanics with a routine place of care were more likely to have correct weight perception (P = 0.02). Obese Hispanics reporting 2-12 encounters with a health-care provider in the past year, but not those reporting 13 or more encounters, were more likely to express weight dissatisfaction (P < 0.001) and intention to lose weight (P = 0.008) than obese Hispanics with one or no visit. Hispanics with obesity-related comorbidities were more likely to be dissatisfied with their weight (P = 0.045).

Covariate-adjusted associations of overweight-related outcomes with acculturation

As shown in Table 3, the associations of US birth with correct weight perception (OR = 4.45; 95% CI = 2.16-9.17) and weight dissatisfaction (OR = 2.95; 95% CI = 1.63-5.35) persisted after controlling for all the covariates. Language acculturation remained significantly associated with intention to lose weight and clinically significant weight loss in covariate-adjusted logistic regression models, although its associations with weight perception and weight dissatisfaction were no longer significant. Compared with exclusively Spanish-speaking obese Hispanics (language usage score = 5), those speaking some or all English showed greater intention to lose weight and greater likelihood of achieving clinically significant weight loss. However, the covariate-adjusted ORs reached statistical significance only for bilingual obese Hispanics reporting about equal use of Spanish and English (language usage scores of 11-15; OR = 2.78; 95% CI = 1.43-5.40) regarding weight loss intention, and for those bilingual individuals reporting more use of English than Spanish (language usage scores of 16-20; OR = 3.94; 95% CI = 1.51-10.3) regarding clinically significant weight loss.

Table 3. Covariate-adjusteda associations of weight-related outcomes with acculturation
 Weight perception (overweight)Weight dissatisfaction (dissatisfied)Intention to lose weight (yes)Intentional weight loss of clinical significance (yes)
 OR (95% CI)P-valueOR (95% CI)P-valueOR (95% CI)P-valueOR (95% CI)P-value
  1. OR = odds ratio, CI = confidence interval.

  2. Adjusted for the following covariates: sociodemographic characteristics (ethnicity, age, sex, education, family income, and employment status), healthcare access and use (health insurance, routine place of care, and number of clinical encounters), and health status (self-rated health status, smoking status, and obesity-related comorbidities).

Born in the United States        
NoReference Reference Reference Reference 
Yes4.45 (2.16, 9.17)<0.0012.95 (1.63, 5.35)<0.0011.09 (0.71, 1.68)0.701.84 (0.82, 4.09)0.14
Language acculturation        
5Reference Reference Reference Reference 
6-102.69 (0.89, 8.14)0.082.11 (0.68, 6.53)0.202.33 (0.99, 5.45)0.0511.65 (0.32, 8.49)0.55
11-151.00 (0.50, 1.99)1.000.52 (0.24, 1.11)0.092.78 (1.43, 5.40)0.0032.39 (1.00, 5.72)0.05
16-201.32 (0.49, 3.57)0.581.97 (0.48, 8.06)0.351.98 (0.81, 4.87)0.133.94 (1.51, 10.3)0.005
21-254.45 (0.81, 24.5)0.095.50 (0.70, 43.1)0.111.83 (0.63, 5.28)0.261.78 (0.68, 4.67)0.24

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion

Hispanic individuals in the United States bear a disproportionate burden of obesity and obesity-related risk factors (e.g., metabolic syndrome) and chronic diseases (e.g., type 2 diabetes) ([3]). Furthermore, the obesity prevalence varies substantially within the Hispanic population, with acculturation being a likely contributor to the heterogeneity ([13, 14, 20]). To better understand the influence of acculturation on within-ethnicity variation in obesity, this study investigated the associations of country of birth and language usage with overweight-related attitudes (weight perception and dissatisfaction) and behavior (weight loss intention and success) among obese US Hispanic adults. The study sample was representative of 7 million US-born and 6.9 foreign-born Hispanic obese adults in the general population. The findings showed that US-born obese Hispanics were more likely to perceive themselves as overweight and desire to weigh less than their foreign-born counterparts, and that obese Hispanics with higher levels of acculturation, measured by language, were more likely to attempt weight loss and actually achieve at least 5% weight loss. These relationships were significant even after controlling for sociodemographic characteristics, including education and income, differential healthcare access and use, and health status.

Weight misperception among foreign-born obese Hispanics may signify a lack of awareness of healthy weight levels and health risks related to excess weight ([21, 22]). Also, cultural norms and beliefs play an important role in shaping body image and desirable weight, thereby influencing individual perspectives on obesity and weight loss ([23]). Based on the evidence of a strong, positive association of substantial weight gain with duration of US residence and across generations among Hispanics, other researchers have urged for the development of obesity prevention and control programs tailored for Hispanic immigrants, particularly those living in new receiver communities ([13, 20]). The association of foreign birth with more common weight misperception and less desire to weigh less in obese Hispanics, as documented in this study, suggests that changing the attitudes toward obesity and weight loss should be a necessary component of weight management programs for this population. This may be accompanied by an emphasis on conserving traditional lifestyle patterns, which has been shown to protect against the negative health consequences of acculturation to an obesogenic environment among immigrants living in the United States ([13, 20, 26]). More research on the reasons for the distorted perception and its potential impact is needed to guide the development of effective intervention strategies.

Our finding that obese Hispanics with low English language acculturation have lower likelihood of weight loss intention and success extends prior evidence showing a positive association between Spanish fluency and obesity among Hispanics, particularly Mexican Americans ([14, 15]). Other studies have also found that Hispanics with low language acculturation are at increased risk of poor lipid and diabetes control ([16, 17]). The lack of culturally appropriate weight management and diabetes prevention programs for Hispanics, especially those with limited English proficiency, is a critical concern because culturally undifferentiated interventions that originate in mainstream populations may be inefficient or even unintentionally increase health disparities ([27, 28]).

Cultural tailoring goes far beyond literal translation of English materials into the language of minority group preference. Instead, interventions should target the “deep structure” of cultural groups and address cultural meanings that are likely to motivate behavior change. Hispanic cultures differ in obvious ways from mainstream American culture, for example, superficial differences such as the importance of particular holidays, the role of food and music, and sports preferences. Lying beneath these surface elements, which are frequently expressed as cliches and stereotypes, are a culture's “deep structure.” The latter relates to culturally rooted values and beliefs, regarding such salient issues as ancestry, social networks, family structure, health, heredity of disease, individual and family risk of disease, choices for risk modification, acceptable means of risk communication and intervention, and barriers to resource access ([29, 30]). Within a culture, heterogeneity exists based on the site of family origin (e.g., region in Mexico), process of acculturation, age at immigration, historical timing of immigration, gender, socioeconomic and legal status, and occupation. Hence, proper cultural adaptation also needs to address these within-population and between-individual differences—through individualized tailoring ([31]). Past research has suggested that Hispanic-focused weight management interventions should incorporate traditional foods, support cultural traditions, and include a family focus ([25]).

This study has several limitations. First, the NHANES is a serial, cross-sectional survey, which precludes the possibility of making causal inferences. Second, NHANES oversamples for Mexican Americans, whereas data are insufficient to categorize non-Mexican Hispanics into ethnic subgroups. However, research has shown that health outcomes can vary substantially by subgroup within the Hispanic population ([32]). Third, although country of birth and language usage are two commonly used proxies of acculturation, the data set does not capture other dimensions of acculturation, for example, duration of US residence, degree of identity with home culture and traditions, and neighborhood immigrant composition. Fourth, longitudinal weight measurements were not available in NHANES. Self-reported weights are subject to reporting errors and biases, which may vary by age, sex, ethnicity, and overweight status ([33]). Prior studies based on NHANES III data have indicated that underreporting of weight is more common among Mexican American women than among NHW women ([34]). In this study, we used respondents' measured weights at the time of the interview and self-reported weights 12 months ago for calculating weight loss. A comparison of measured weights with self-reported weights at the time of the interview suggested that the magnitude of underreporting was greater among respondents with higher language acculturation than their counterparts, whereas there was no significant difference by country of birth. If this holds true for self-reported weights 12 months ago, then our observed association of language acculturation and intentional weight loss of clinical significance may be a conservative estimate of the actual relationship.

In conclusion, our results suggest that independent of sociodemographic characteristics, healthcare access and use, and health status, acculturation is significantly associated with overweight-related attitudes and behavior among obese Hispanics. Further investigation is needed to understand the role of acculturation in health inequality within the Hispanic population and to identify potential mechanisms. Finally, because Latinos comprise 16% of the US population and account for one-half of the nation's growth (1), effective, culturally appropriate weight management interventions in this population present a prime opportunity to address health disparities and control the substantial downstream personal and societal costs of obesity and diabetes.