New research reveals that ultra-processed foods account for nearly half of US household grocery purchases, with lower-income and less-educated families bearing the brunt, underscoring the need for urgent national policy reform.
Study: Exploring disparities in the proportion of ultra-processed foods and beverages purchased in grocery stores by US households in 2020. Image Credit: Rimma Bondarenko / Shutterstock
In a recent article published in the journal Public Health Nutrition, researchers analyzed purchases of ultra-processed foods (UPFs) by American households, examining differences based on education levels, household income, and race and ethnicity.
They found that almost half of all food purchases and more than one-third of beverages were UPFs; snacks, soft drinks, and mixed dishes accounted for the highest proportions of UPFs. Purchases of UPFs were more common in households with lower incomes and members with lower levels of education, as well as in non-Hispanic white households. The study included a nationally representative sample of 59,939 US households, analyzing over 33 million purchased products during 2020.
Statistical analyses used a significance threshold of p < 0.0001 to determine meaningful differences between groups.
Background
UPFs are highly engineered products made with additives and ingredients derived from food and are becoming increasingly dominant in American diets. These foods have been linked to several chronic diseases such as heart disease, cancer, obesity, depression, sleep issues, and even premature death.
Unlike traditional food research focused on nutrients, recent studies use the NOVA classification system to assess food processing levels, aiming to highlight foods designed to encourage overconsumption.
UPFs are identified not only by their high processing but also by the use of certain additives rarely found in home cooking. Current estimates show that approximately half of American beverage and food purchases are ultra-processed. Nutrition-related health problems, including obesity, also show significant disparities across ethnic, racial, and socioeconomic groups.
Past research undertaken from 2007 to 2012 revealed that more educated individuals consume fewer UPFs, while white and Black adolescents consume the most. Income and age were inversely related to UPF intake. While purchase data provide a clearer picture than dietary surveys, earlier findings indicated that Black households bought more processed drinks but fewer processed foods.
Proportion of food and beverages purchased by US households by level of processing
About the study
In a quickly changing food market, researchers sought to update research on UPF purchasing patterns, exploring how race, income, education, and family structure intersect with food purchasing.
The study utilized data from a consumer panel survey of 59,939 American households, drawn from a nationally representative sample of approximately 35,000 to 60,000 households per year, conducted in 2020. Participating households scanned packaged beverage and food purchases and provided demographic data, including race, ethnicity, education, income, and household composition.
Ethnicity and race were grouped as non-Hispanic, Hispanic, non-Hispanic white, and non-Hispanic Black. Income was divided into three levels based on the Federal Poverty Level (FPL), namely <185%, 185–350%, and >350%. Education was classified into at most high school, some college, or college graduate and above. Households were also classified based on whether they included children.
Using commercial databases, each product was matched with nutrient labels and ingredient data. Items without nutrition labels or barcodes (like fresh produce and meats) were excluded. Products were grouped into nine food and eleven beverage categories.
Based on the NOVA classification system, items were categorized as minimally processed, containing processed ingredients, processed foods, and ultra-processed foods (UPFs). UPFs were identified by checking for substances not typically used in home kitchens, including flavor enhancers and sweeteners. If items originally classified as UPFs were found to contain none of the NOVA markers or additives, the product was reclassified as processed foods rather than UPFs.
The authors used statistical tests to compare the purchase volumes of UPFs across demographic sub-groups, applying Bonferroni adjustments to account for multiple comparisons.
Proportion of (A) foods and (B) beverages purchased by US households by both level of processing and food category
Findings
In 2020, among 59,939 American households and over 33 million purchased products that were analyzed, 38% of beverages and 48% of packaged foods were classified as UPFs, accounting for 43% of all products purchased.
The groups with the largest share of UPFs were carbonated beverages, which contained 90% UPFs, followed by soups and mixed dishes at 81%, and snacks and sweets at 71%. In contrast, fruits, vegetables, fats, nuts, and oils had virtually no UPFs. Importantly, although these minimally processed categories had almost no UPFs, they made up only a small fraction (12%) of total purchases, compared to the top three UPF categories (carbonated soft drinks, sweets and snacks, and mixed dishes/soups), which together comprised 36% of all purchases.
Non-Hispanic white households had the highest percentage of UPF purchases (49%), followed by non-Hispanic Black (47%), Hispanic (46%), and other ethnic groups (43%). Lower-income and less educated households purchased a higher proportion of UPFs than higher-income and more educated households.
These patterns held across racial-ethnic groups even when stratified by income. Differences were observed across food categories, with non-Hispanic white households purchasing more carbonated UPF drinks, and non-Hispanic Black households purchasing more UPF dairy beverages and fruit juices. The "other" race/ethnicity group consistently had the lowest proportion of UPF purchases, except in the dairy beverage category. Grain products and fruit/vegetable juices showed the greatest variability by group, with at least a 10% difference between the lowest and highest proportions of UPFs purchased.
It is important to note that this study measured only in-home purchases. Previous research has shown that non-Hispanic Black households are more likely than other race-ethnic groups to purchase sugar-sweetened beverages outside the home, which was not captured in this dataset. Therefore, the observed lower in-home beverage UPF purchases among non-Hispanic Black households may not reflect their total beverage consumption.
Conclusions
This study confirms the dominance of UPFs in American household food purchases, consistent with previous research. Notably, non-Hispanic whites now purchase more ultra-processed beverages than other groups, differing from earlier findings. One possible explanation for this shift is purchasing behavior changes during the COVID-19 pandemic, which may have disproportionately affected minority groups and altered patterns of food insecurity and at-home purchasing. Additionally, the limitation of only capturing in-home purchases may confound observed racial and ethnic differences in UPF beverage purchasing patterns.
Despite using a nationally representative sample and full-year data collection, limitations include reliance on purchase (rather than consumption) data, potential underestimation due to food waste, exclusion of restaurant purchases, possible misclassification using the NOVA classification system, and pandemic-related purchasing shifts. Despite these, scan-based data likely ensures reasonable accuracy.
The findings emphasize the urgent need for American policy initiatives addressing UPF consumption to enhance diet quality and mitigate chronic disease risks, particularly given the absence of federal regulations compared to other countries.