Exploring disparities in the proportion of ultra-processed foods and beverages purchased in grocery stores by US households in 2020

Abstract

Objective: American diets are increasingly based on ultra-processed foods (UPF). Current research, particularly on socio-economic differentials, is lacking. This study aimed to provide an updated examination of US household purchases of UPF and how this differs by race-ethnicity, household income and household education.

Design: The NielsenIQ Consumer Panel 2020 was utilised for analysis. Each food and beverage product purchased by US households was assigned a level of processing under the Nova level of processing classification system. The volume of UPF purchased overall and by food group was determined for each Nova processing group and examined by race-ethnicity, education and income. Results were stratified by race-ethnicity within each income group. A P value < 0·0001 was considered significant.

Setting: This study analysed data from the Nielsen IQ Consumer Panel 2020 which recorded household food purchases in the USA.

Participants: The Nielsen IQ Homescan Consumer Panel is a nationally representative longitudinal survey of around 35 000 and 60 000 US households.

Results: Of 33 054 687 products purchased by 59 939 US households in 2020, 48 % of foods and 38 % of beverages were considered UPF. Categories with the highest proportion of purchases deriving from UPF included carbonated soft drinks (90 %), mixed dishes and soups (81 %) and sweets and snacks (71 %). Slightly higher but statistically significant proportions of UPF purchases occurred in the lowest income and education groups and among non-Hispanic whites.

Conclusions: It is concerning that household purchases of UPF in the USA are high. Policies that reduce consumption of UPF may help reduce diet-related health inequalities.

Keywords: Food environment; Food purchases; NOVA classification; Race–ethnic differences; Ultra-processed foods.

Figure 1.

Proportion of food and beverages purchased by US households by level of processing. Footnotes: University of North Carolina calculation is based in part on data reported by NielsenIQ through its Homescan Services for all food categories including beverages for 2020 across the US market. NielsenIQ, 2020⁽³⁴⁾. Authors’ calculations are based in part on data reported by NielsenIQ through its Homescan Services for all food categories including beverages for 2020 across the US market. NielsenIQ, 2020. The conclusions drawn from the data are those of UNC and do not reflect the views of NielsenIQ. NielsenIQ is not responsible for and had no role in, and was not involved in, analysing and preparing the results reported herein.

Figure 2.

Proportion of (a) foods and (b) beverages purchased by US households by both levels of processing and food category. Footnotes: University of North Carolina calculation is based in part on data reported by NielsenIQ through its Homescan Services for all food categories including beverages for 2020 across the U.S. market. NielsenIQ, 2020⁽³⁴⁾. Authors’ calculations are based in part on data reported by NielsenIQ through its Homescan Services for all food categories including beverages for 2020 across the US market. NielsenIQ, 2020. The conclusions drawn from the data are those of UNC and do not reflect the views of NielsenIQ. NielsenIQ is not responsible for and had no role in, and was not involved in, analysing and preparing the results reported herein.

Figure 3.

Proportion of beverages (a) and foods (b) purchased by US households by level of processing, by demographic subgroup. Footnotes: University of North Carolina calculation is based in part on data reported by NielsenIQ through its Homescan Services for all food categories including beverages for 2020 across the US market. NielsenIQ, 2020⁽³⁴⁾. Authors’ calculations are based in part on data reported by NielsenIQ through its Homescan Services for all food categories including beverages for 2020 across the US market. NielsenIQ, 2020. The conclusions drawn from the data are those of UNC and do not reflect the views of NielsenIQ. NielsenIQ is not responsible for and had no role in, and was not involved in, analysing and preparing the results reported herein.