Examining Sociodemographic Disparities in Household Purchases of Fruit Drinks with Policy-Relevant Nutrition Claims

Abstract

\n \n \n Fruit drinks are a top source of added sugars among US children and carry misleading nutrition claims. Disparities exist in food label use and fruit drink (FD) consumption. We aimed to describe household (HH) demographic characteristics associated with purchase of 1) any FDs and 2) FDs with specific nutrition claims among HH with children 0–5 y.\n \n \n \n We merged HH monthly-level FD purchasing data from 2017 Nielsen Homescan (n 60,712) with product-level nutrition claims data. For our first aim, we examined differences in predicted probabilities of purchasing any FDs by HH race/ethnicity, income, and education. In aim 2, to account for observed differences in likelihood of purchasing any FDs by demographic characteristics, we constructed inverse probability (IP) weights based on each HH s likelihood of purchasing any FD. We then used IP weighted multivariable logistic\n regression models to examine differences in predicted probabilities of purchasing FDs with specific claims by demographic characteristics, holding the likelihood of purchasing any FDs constant. We used the Holm method to adjust for multiple comparisons.\n \n \n \n One in three HH with young children purchased any FDs. Non-Hispanic (NH) Black (52%), Hispanic (36%), low-income (39%), and low-educated HH (41%) were more likely to purchase any FDs than NH white (31%), high income (26%) and high educated HH (30%) (all p\xa0<\xa00.001). In IP weighted analyses, NH Black HH were more likely to purchase FDs with “Natural” and fruit claims (6.9% and 3.7%) than NH white HH (4.5% and 2.8%) (both p\xa0<\xa00.01). Low and middle income (15.0% and 13.8%) and low and middle educated HH (15.5% and 14.5%) were more likely to purchase FDs with “100% Vitamin C” claims than high income and high educated HH (10.8% and 12.9%) (all p\xa0<\xa00.025). Low educated HH were less likely to purchase FDs with a “Natural” claim (3.9%) than high educated HH (5.2%) (p\xa0=\xa00.002).\n \n \n \n There are sociodemographic disparities in FD purchases and FD purchases with claims among HH with young children in the US, with low income, low education, and NH Black HH being more at risk. Experimental studies are needed to determine if nutrition claims may be contributing to observed disparities in FD consumption.\n \n \n \n Healthy Eating Research National Institutes of Health.\n