In keeping with findings from previous studies of US adults, we found that both non-Hispanic whites and Hispanics failed to meet Healthy People 2010 targets for consumption of both fruits and vegetables. Although more Hispanics than non-Hispanic whites met the fruit objective (+5.6%), more non-Hispanic whites than Hispanics met the vegetable objective (−9.6%).
Our results indicated that more Spanish-speaking Hispanics met the fruit objective than did English-speaking Hispanics (+6.3%). The acculturation effect on fruit consumption, as measured by survey language preference, was strengthened following adjustment for sociodemographic variables in regression models. The findings for lower fruit consumption among English-speaking Hispanics are similar to those from the National Cancer Institute 7-item fruit and vegetable screener, which indicated that greater language acculturation among Mexican-American women was associated with significantly decreased fruit consumption (11
Conversely, fewer Spanish-speaking Hispanics met the vegetable objective than did English-speaking Hispanics (−6.0%). The effect of acculturation, as measured by survey language preference, on vegetable consumption remained significant in regression models after accounting for sociodemographic variables. Findings for fruit (34.7%) and vegetable (21.8%) consumption among English-speaking Hispanics were intermediate to intake among non-Hispanic whites and Spanish-speaking Hispanics, which is in keeping with the expectation that increased acculturation will yield dietary patterns more aligned with those of the population majority.
A more detailed comparison of screener items indicated some differential consumption by ethnicity and survey language preference, particularly regarding vegetable intake. Our findings may reflect true differences in vegetable consumption influenced by potential factors such as access and affordability, or they may reflect the ability of the translated screener to adequately query about vegetables categorized as "green salad" or "other vegetables." Small but significant differences in individual components of intake across Hispanic subgroups were also found in the 2005 California Health Interview Survey (CHIS), which used a short 7-item dietary screener similar to BRFSS (fruit juice, fruit, green salad, cooked-dried beans, fried potatoes, nonfried potatoes, and other vegetables). Specifically, intake of "other white potatoes" was different across subgroups of Hispanic women. Consumption of green salad also varied among women; South American women reported higher intake of green salad than did Central American women (20
). Results from CHIS suggest that dietary preferences differ among heterogeneous Hispanic subgroups, an aspect that state-specific analyses of BRFSS data may be able to further explore. Future research may investigate the reasons for the variation in fruit and vegetable intake, particularly since evidence supports decreased differential risk for cancer attributed to fruits compared with vegetables (21
Furthermore, because of recent overall changing dietary patterns in Mexico and Latin America, the premise that diet quality among Hispanics living in the United States decreases with increased acculturation may no longer hold true. In Mexico, findings from health and nutrition surveys have found temporal changes in dietary intake (22
), and results from multiple national nutrition surveys conducted among adult residents in countries in Latin America and the Caribbean indicate that consumption of fruits, vegetables, grains, cereals, and legumes has decreased and that consumption of saturated fat has increased (23
Assessment of fruit and vegetable intake using BRFSS has strengths as well as some limitations. Strengths of this study include using a representative population-based sample, including respondents from territories, and a large sample size. Although fruits and vegetables are just 1 part of a healthy diet, they are foods encouraged by the Dietary Guidelines for Americans 2010
) and are the only continuously monitored nutritional intake items in the state-based BRFSS. Compared with other dietary components that influence diet quality, as measured by the Healthy Eating Index 2005, whole fruit consumption has the highest correlation with overall dietary quality (r
= 0.45 for whole fruit, r
= 0.43 for total fruit, r
= 0.18 for vegetables, r
= 0.07 for total grains, and r
= −0.12 for milk) (25
We were restricted to survey language preference as the only measure by which to assess acculturation. Determining acculturation through assessment scales in population-based surveys such as BRFSS may be difficult due to limitations on the number of survey items to reduce respondent burden and maximize participation. However, a strong correlation was found (r
= 0.80) between a 1-item language preference question on a telephone survey and a validated acculturation assessment instrument among Latino adults, validating the use of survey language preference as a proxy for acculturation in this population (26
Although CDC provides a Spanish translation of the BRFSS survey, there may be differences in how individual states and territories translate the screener for use in their jurisdictions, depending on population characteristics of the majority of Spanish-speaking residents in that state. Adding a question about country of origin on the BRFSS may be a future consideration to address heterogeneity issues among respondents in this population. A preliminary report from the National Health Interview Survey for 2007 found that Hispanics or Latinos had a higher percentage (18.0%-19.3%) of households that were wireless-only (eg, residents used only cellular telephones) compared with non-Hispanic whites (11.3%-12.9%) (16
). To our knowledge, assessment of sociodemographic characteristics by race/ethnicity for cellular telephone users has not been published, and we are unsure how this may affect our estimates.
Finally, estimates calculated on the basis of abbreviated food frequency questionnaires, such as the BRFSS fruit and vegetable screener, are generally lower than those from studies that use other methods such as the National Health and Nutrition Examination Survey 24-hour recall, which limits comparison across surveys (28
). However, for the purpose of surveillance, the module has moderate validity when compared with other dietary assessment methods (29
). The exclusion of fried potatoes, such as French fries, in the BRFSS fruit and vegetable screener also contributes to lower estimates of overall intake.
Few assessments of surveillance data have determined differences in dietary intake by acculturation as measured by survey language preference. Although a causal relationship cannot definitively be established between acculturation and fruit and vegetable intake, our findings have implications regarding acculturation among the Hispanic population in the United States. Our analysis demonstrates how brief screeners can be used to determine possible disparities among minority groups and to monitor population goals to eliminate racial and ethnic health disparities.