Background Online grocery shopping interventions using optimal defaults have improved the nutritional quality of grocery purchases, but no studies have explored impacts on dietary intake. This analysis examines the potential of an optimal defaults online grocery shopping intervention to promote healthier intake by assessing (1) the relationship between the nutritional quality of grocery shopping purchases and the nutritional quality of intake, and (2) intervention effects on the nutritional quality of participants' dietary intake. Methods A randomized controlled trial tested the effects of an optimal defaults intervention promoting grocery purchases consistent with a diabetic diet. As part of the 5-week, 3-arm study, participants were randomly assigned to grocery shop either (1) in person as usual (control group), (2) online (online group), or (3) online with a prefilled, default grocery shopping cart (defaults group) during a 3-week intervention. All groups received recipes corresponding to the Dietary Approaches to Stop Hypertension diet. Participants submitted grocery receipts and completed interviews weekly. Fifty-five adults provided a complete food frequency questionnaire post intervention. Linear regression tested the relationship between the nutritional quality of grocery shopping purchases and the nutritional quality of intake. Analysis of variance examined the effect of study group on the nutritional quality of dietary intake. Results The nutritional quality of household grocery purchases was positively associated with the nutritional quality of participants' intake ( P < .001). The effect of study group on the nutritional quality of intake was not significant ( P = .07), although trend-level results were in the expected direction with the defaults group having the highest nutritional quality of the 3 groups. Conclusions Optimal defaults interventions' effects on grocery purchases may extend to dietary intake. Future research using more precise and objective measures of dietary intake could elucidate potential effects.
Ferrante et al. (Tue,) studied this question.