Ultra‐processed and Ready‐to‐eat Food and Beverage Purchases Differ by Race, Education, and Income in a Longitudinal US Study

Abstract

It is not clear whether processed foods contribute to sociodemographic disparities in the diets of Americans. Our objective was to determine whether race/ethnicity, education, and income were associated with ultra‐processed or ready‐to‐eat food and beverage purchases among US households. We analyzed data from the 2000‐2012 Homescan panel (n=157,142). We classified 1.2 million barcoded products by level of food processing and convenience. Longitudinal random effects linear and multinomial logistic regression models were used to examine time‐varying associations.The calorie contribution of ultra‐processed products to food purchases was 蠅58% for all subpopulations. Inverse associations of ultra‐processed food purchases with education and income emerged across time as households of higher, but not lower, socioeconomic status (SES) shifted away from these foods. Compared to white households, black households had higher ultra‐processed beverage purchases yet lower ultra‐processed food purchases. Blacks were more likely than whites to be in the highest quartile for purchases of moderately processed foods (predicted probability: 45.8% vs 21.0%) and foods requiring cooking (41.3% vs 21.7%) and purchased more cooking oils and table sugar. Black race and lower education were associated with lower ready‐to‐eat convenience food purchases. Findings suggest that ultra‐processed foods among low SES households and ultra‐processed beverages among blacks may contribute to US diet disparities. Sociodemographic differences in patterns reflective of greater home‐cooking warrant further study.Funding source: Robert Wood Johnson Foundation (67506, 68793, 70017, 71837), NIH (R01DK098072), CPC 5 R24 HD050924