Table 1.

RCT studies investigating effects of the Nutri-Score on food purchases.

Year of Publication	Country	Authors Affiliated w Nutri-Score	Study Description	Results	Notes	Effect on FSA-NPS	Reference
2019	CO	no	Randomised field trial in a university cafeteria, with randomly provided information on the Nutri-Score. n= 484 participants.	Using the Nutri-Score led to more protein, more calories and more expenditures (on healthy items only) in purchases. Purchases of ‘unhealthy’ products did not decrease.	Customers were 10% more likely to buy a healthier item than controls. Information on the Nutri-Score system increased the store’s sales.	Not investigated	Mora-Garcia et al., 2019 [16]
2019	SG	no	RCT investigating the effect of the Nutri-Score compared to the UK’s multiple traffic light system (MTL) and no label in online grocery store. n = 154 participants in a 3 × 3 crossover (within-person) design.	The Nutri-Score and MTL performed significantly better vs. no-label controls. NS performed statistically better than MTL and control-based on average Nutri-Score. MTL (but not the Nutri-Score) statistically reduced calories or sugar from beverages.	Thorough study with crossover design.	Not investigated	Finkelstein et al., 2019 [17]
2021	BE	no	A difference-in-difference analysis of a natural experiment in 43 supermarkets of a major retailer in Belgium versus 14 control stores, studying the impact of shelf tags with the Nutri-Score on consumer purchases.	The proportion of Nutri-Score B and C product sales was more favourable in intervention than control stores and less favourable for Nutri-Score D product sales. A positive impact was found for 17/58 food categories (vegetable, fruit and dairy products, and confectionery), a negative impact for 16/58 categories (bread and bakery products).	The impact on consumer purchases was mixed as difference–in-differences found were favourable for Nutri-Score B and C products and unfavourable for Nutri-Score D products. Shelf labelling on its own is unlikely to significantly influence consumer behaviour.	Not investigated	Vandevijvere and Berger 2021 [10]
2020	FR	no	RCT investigating four FOPLs (SENS, Nutri-Score, Nutri Repère, Nutri-Couleurs) to improve the nutritional quality of food purchases in real-life grocery shopping settings.	The Nutri-Score increased purchases of foods in the top-third of their category by 14% nutrition-wise, but had no impact on purchases of foods with medium, low or unlabelled nutritional quality.	The Nutri-Score improved the nutritional quality of labelled foods purchased by only 2.5% in the FSA-NPS score. Effect sizes were 17 times smaller on average than those found in comparable laboratory studies.	Yes, but based on four product groups only	Dubois et al., 2020 [18]
2019	FR	yes	Three RTCs in students (n = 1866), low-income individuals (n = 336), and cardiovascular patients (n = 1180) investigating the effect of the Nutri-Score on overall nutritional quality of purchases in an online supermarket compared to the RIs and no label.	Shopping cart contents were lower in calories and saturated fatty acids and higher in fruits and vegetables in the Nutri-Score arm than in the other arms.	No significant difference between the Nutri-Score and no-label groups or between RIs and no-label groups.	Yes, but no significant effect of Nutri-score versus no label	Egnell et al., 2019 [19]
2021	FR	yes	Three RTCs in students (n = 1866), low-income individuals (n = 336), and cardiovascular patients (n = 1180) investigating the effect of the Nutri-Score on purchasing intentions in an online supermarket compared to RIs and no label.	Shopping carts of participants simulating purchases with the Nutri-Score affixed to pre-packaged foods contained a higher proportion of unpacked products—especially raw fruits and meats, i.e., with no FoPL—comparedto participants purchasing with no label or withRIs.	This is a sequel paper to the one above (#17). It is a post-hoc analysis, viz. analyses that were not originally planned.“The Nutri-Score appears to decrease purchases in processed products resulting in higher proportions of unprocessed and unpacked foods, in line with public health recommendations.”	Not investigated	Egnell et al., 2021 [20]
2021	FR	yes	RCT investigating the effect of the Nutri-Scorecompared to RIs and no label. Participants (n = 336) went on a simulated grocery shopping at an experimental online supermarket.	The Nutri-Score performed significantly better versus RIs (overall nutritional quality of the shopping cart, and lower caloric and saturated fatty acids content), but not versus no label.	This is one of the three arms in the study mentioned above published as a separate paper.	Yes, but no significant effect of Nutri-score versus no label	Egnell et al., 2021 [21]
2021	NL	no	Investigate the effect of the Nutri-Score on (n = 192) consumer attitudes, taste perception, and purchase intention in an online environment by comparing the Nutri-Score’s efficacy on three different snacks labelled with Nutri-Score A, B, and C.	No effects of the Nutri-Score were observed on attitudes, taste perception, or purchase intention.	This study is similar to those conducted by the Nutri-Score’s developers.	Not investigated	Folkvord and Pabian 2021 [22]

BE = Belgium; FR = France; CO = Colombia; NL = the Netherlands; SG = Singapore; FSA-NPS = Food Standard Agency Nutrient Profile Score; RIs = reference intakes.