Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BHNRC) » Beltsville Human Nutrition Research Center » Food Surveys Research Group » Research » Publications at this Location » Publication #331668

Research Project: What We Eat in America - Dietary Survey: Data Collection, Interpretation, Dissemination, and Methodology

Location: Food Surveys Research Group

Title: Differences in fruit and vegetable intake of U.S. adults by sociodemographic characteristics evaluated by two methods

Author
item Hoy, M Katherine
item Goldman, Joseph
item Moshfegh, Alanna

Submitted to: Journal of Food Composition and Analysis
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/30/2017
Publication Date: 7/3/2017
Citation: Hoy, M.K., Goldman, J.D., Moshfegh, A.J. 2017. Differences in fruit and vegetable intake of U.S. adults by sociodemographic characteristics evaluated by two methods. Journal of Food Composition and Analysis. 64(Part 1 2017):97-103. https://dx.doi.org/10.1016/j.jfca.2017.06.012.
DOI: https://doi.org/10.1016/j.jfca.2017.06.012

Interpretive Summary: Socioeconomic factors have been associated with fruit and vegetable (FV) intake. Estimates of FV consumption vary depending on what is counted towards intake, which will be determined by the research purpose. This study estimates FV consumption of adults 20 years and older in the U.S. by sociodemographic characteristics including gender, age, race/ethnicity, income, and education level using two different methods. An epidemiological method included all FV after disaggregating foods and beverages. A behavioral method considered foods that provided 0.2 cup equivalents (CE) or more per 100 grams, which included discrete items such as an apple or green beans, and mixed dishes that provided approximately ½ cup FV per 1 cup serving. One day of dietary intake data from What We Eat in America NHANES 2009-2012 was used. Using the behavioral method, estimates of fruit and vegetable intake, respectively, were around 10% and 30% lower than the epidemiological method. With some exceptions, mean FV intakes and percentage reporting any intake were higher among females, older individuals, Non-Hispanic whites and, in some cases Hispanics, individuals at higher income levels, and those with less than high school education. Within each sociodemographic characteristic, the significance of differences in estimates of FV intake did not change between the two approaches. These results can inform evaluation of the impact of dietary guidance messages and needs for nutrition education.

Technical Abstract: Socioeconomic factors have been associated with fruit and vegetable (FV) intake. Nutrition behaviors such as choosing and consuming FV may be among many factors related to racial/ethnic health disparities. Intake can be measured using an epidemiological approach that includes FV from all foods after disaggregation, or a behavioral approach that includes choices consistent with dietary guidance that excludes incidental amounts present in small quantities (e.g. tomato in catsup) and sources high in added sugars, fat, and sodium. The purpose of this study is to describe and compare FV intake by sociodemographic characteristics using both methods. One day dietary intake data of adults 20+ years (N=10,563) from What We Eat in America, NHANES 2009-2012 were used. FV in foods was estimated using the Food Patterns Equivalents Database (FPED) 2009-2012 and is expressed as cup equivalents (CE). Mean intakes and percent reporting were compared by gender, age, ethnicity, income, and education using both methods. Comparisons were made by t-tests with regression adjustment for covariates. Results were considered significant at P<0.001. Using the epidemiological approach, estimates of intake ranged from 0.8 to 1.2 CE for fruit and 1.1 to 1.7 CE for vegetables. When estimated by the behavioral approach, fruit intake ranged from 0.8 to 1.1 CE, and vegetable intake excluding potatoes was 0.8 to 1.3 CE. Significance of most differences within each sociodemographic did not change between the two methods. With some exceptions, mean FV intakes and percentage reporting any intake were higher among females, older individuals, Non-Hispanic whites and, in some cases Hispanics, individuals at higher income levels, and those with less than high school education. These results can inform evaluation of the impact of dietary guidance messages and needs for nutrition education.