Association of Socioeconomic and Geographic Factors With Diet Quality in US Adults

Abstract

Importance: Poor diet quality is a key factor associated with obesity and chronic disease. Understanding associations of socioeconomic and geographic factors with diet quality can inform public health and policy efforts for advancing health equity.

Objective: To identify socioeconomic and geographic factors associated with diet quality in a large US cohort study.

Design, setting, and participants: This cross-sectional study included adult men and women who enrolled in the Cancer Prevention Study-3 at American Cancer Society community events in 35 US states, the District of Columbia, and Puerto Rico between 2006 and 2013. Participants completed a validated food frequency questionnaire between 2015 and 2017. Data were analyzed from February to November 2021.

Exposures: The main exposures included self-reported race and ethnicity, education, and household income. Geocoded addresses were used to classify urbanization level using Rural-Urban Commuting Area codes; US Department of Agriculture's Food Access Research Atlas database classified residence in food desert.

Main outcomes and measures: Poor diet quality was defined as lowest quartile of dietary concordance with the 2020 American Cancer Society recommendations for cancer prevention score, based on sex-specific intake categories of vegetables and legumes, whole fruits, whole grains, red and processed meat, highly processed foods and refined grains, and sugar-sweetened beverages.

Results: Among 155 331 adults, 123 115 were women (79.3%), and the mean (SD) age was 52 (9.7) years), and there were 1408 American Indian or Alaskan Native individuals (0.9%); 2721 Asian, Native Hawaiian, or Pacific Islander individuals (1.8%); 3829 Black individuals (2.5%); 7967 Hispanic individuals (5.1%); and 138 166 White individuals (88.9%). All key exposures assessed were statistically significantly and independently associated with poor diet quality. Compared with White participants, Black participants had a 16% (95% CI, 8%-25%) higher risk of poor diet quality, while Hispanic/Latino had 16% (95% CI, 12%-21%) lower risk and Asian, Native Hawaiian, and Pacific Islander participants had 33% (95% CI, 26%-40%) lower risk of poor diet quality. After controlling for other characteristics, rural residence was associated with a 61% (95% CI, 48%-75%) higher risk of poor diet quality, and living in a food desert was associated with a 17% (95% CI, 12%-22%) higher risk. Associations of income with diet quality and education with diet quality varied by race and ethnicity (income: P for interaction = .01; education: P for interaction < .001). All diet score components were associated with disparities observed.

Conclusions and relevance: This cross-sectional study found that multiple individual-level socioeconomic and geographic variables were independently associated with poor diet quality among a large, racially and ethnically and geographically diverse US cohort. These findings could help to identify groups at highest risk of outcomes associated with poor diet to inform future approaches for advancing health equity.

Figure 1.. Association of Income Level and Risk of Poor Diet Quality, by Race and Ethnicity

Odds ratio (OR) and 95% CI of lowest quartile of American Cancer Society diet quality score overall by income and stratified by race and ethnicity. Models included age, sex, energy intake, race and ethnicity, income, education, Rural-Urban Commuting Area code, and residence in a food desert, and a race × income interaction term. The reference group was White participants with income $50 000 to less than $75 000. P for interaction = .01.

Figure 2.. Association of Educational Attainment With Risk of Poor Diet Quality, by Race and Ethnicity

Odds ratio (OR) and 95% CI of lowest quartile of American Cancer Society diet score overall, by attained education, stratified by race and ethnicity. Models included age, sex, energy intake, race and ethnicity, income, education, Rural-Urban Commuting Area code, and residence in a food desert, and a race × education interaction term. The Reference group was White participants with a college degree. P for interaction < .001.