What Is the Nutritional Quality of Pre-Packed Foods Marketed to Children in Food Stores? A Survey in Switzerland.

Food marketing targeting children influences their choices and dietary habits, and mainly promotes food high in fat, sugar, and salt as well as ultra-processed food. The aim of this study was to assess the nutritional quality of food and beverages marketed to children over the age of 3 and available on the Swiss market. Products with at least one marketing technique targeting children on the packaging were selected from five food store chains. Three criteria to assess nutritional quality were used: (1) nutritional composition (using the Nutri-Score), (2) degree of processing (NOVA classification), and (3) compliance with the World Health Organization (WHO) Nutrient Profile Model (NPM). A total of 735 products were found and analyzed. The most common marketing techniques used were childish names/fonts (46.9%), special characters (39.6%), and children's drawings (31.3%). Most products had a Nutri-Score of D or E (58.0%) and were ultra-processed (91.8%). Only 10.2% of products displayed the Nutri-Score. The least processed products generally had a better Nutri-Score (p < 0.001). Most products (92.8%) did not meet the criteria of the WHO NPM. Products that met the WHO NPM criteria, organic products, and products with a nutritional claim generally had a better Nutri-Score and were less processed (ps < 0.05). Pre-packaged foods and beverages marketed to children in the Swiss market were mostly of poor nutritional quality. Public health measures should be adopted to improve the nutritional quality of foods marketed to children in Switzerland and restrict the marketing of unhealthy foods to children.

The Nutri-Score nutrition label: Associations between the underlying nutritional profile of foods and lipoprotein particle subclass profiles in adults.

BACKGROUND AND AIMS: Lipoprotein particle concentrations and size are associated with increased risk for atherosclerosis and premature cardiovascular disease. Certain dietary behaviours may be cardioprotective and public health strategies are needed to guide consumers' dietary choices and help prevent diet-related disease. The Food Standards Agency nutrient profiling system (FSAm-NPS) constitutes the basis of the five-colour front-of-pack Nutri-Score labelling system. No study has examined FSAm-NPS index associations with a wide range of lipoprotein particle subclasses. METHODS: This was a cross-sectional study of 2006 middle-to older-aged men and women randomly selected from a large primary care centre. Individual participant FSAm-NPS dietary scores were derived from validated food frequency questionnaires. Lipoprotein particle subclass concentrations and size were determined using nuclear magnetic resonance spectroscopy. Multivariate-adjusted linear regression analyses were performed to examine FSAm-NPS relationships with lipoprotein particle subclasses. RESULTS: In fully adjusted models which accounted for multiple testing, higher FSAm-NPS scores, indicating poorer dietary quality, were positively associated with intermediate-density lipoprotein (ß = 0.096, p = 0.005) and small high-density lipoprotein (HDL) (ß = 0.492, p = 0.006) concentrations, a lipoprotein insulin resistance score (ß = 0.063, p = 0.02), reflecting greater lipoprotein-related insulin resistance, and inversely associated with HDL size (ß = -0.030, p = 0.045). CONCLUSIONS: A higher FSAm-NPS score is associated with a less favourable lipoprotein particle subclass profile in middle-to older-aged adults which may be a potential mechanism underlying reported health benefits of a healthy diet according to Nutri-Score rating.

Differential association of dietary scores with the risk of type 2 diabetes by metabotype.

PURPOSE: We aimed to examine the association between dietary patterns and type 2 diabetes mellitus (T2DM) while considering the potential effect modification by metabolic phenotypes (metabotypes). Additionally, we aimed to explore the association between dietary scores and prediabetes. METHODS: A total of 1460 participants (11.8% with T2DM) from the cross-sectional population-based KORA FF4 study were included. Participants, classified into three metabotype subgroups, had both their FSAm-NPS dietary index (underpinning the Nutri-Score) and ultra-processed foods (UPF) intake (using NOVA classification) calculated. Glucose tolerance status was assessed via oral glucose tolerance tests (OGTT) in non-diabetic participants and was classified according to the American Diabetes Association criteria. Logistic regression models were used for both the overall and metabotype-stratified analyses of dietary scores' association with T2DM, and multinomial probit models for their association with prediabetes. RESULTS: Participants who had a diet with a higher FSAm-NPS dietary index (i.e., a lower diet quality) or a greater percentage of UPF consumption showed a positive association with T2DM. Stratified analyses demonstrated a strengthened association between UPF consumption and T2DM specifically in the metabolically most unfavorable metabotype (Odds Ratio, OR 1.92; 95% Confidence Interval, CI 1.35, 2.73). A diet with a higher FSAm-NPS dietary index was also positively associated with prediabetes (OR 1.19; 95% CI 1.04, 1.35). CONCLUSION: Our study suggests different associations between poorer diet quality and T2DM across individuals exhibiting diverse metabotypes, pointing to the option for stratified dietary interventions in diabetes prevention.

NutriGreen image dataset: a collection of annotated nutrition, organic, and vegan food products.

Introduction: In this research, we introduce the NutriGreen dataset, which is a collection of images representing branded food products aimed for training segmentation models for detecting various labels on food packaging. Each image in the dataset comes with three distinct labels: one indicating its nutritional quality using the Nutri-Score, another denoting whether it is vegan or vegetarian origin with the V-label, and a third displaying the EU organic certification (BIO) logo. Methods: To create the dataset, we have used semi-automatic annotation pipeline that combines domain expert annotation and automatic annotation using a deep learning model. Results: The dataset comprises a total of 10,472 images. Among these, the Nutri-Score label is distributed across five sub-labels: Nutri-Score grade A with 1,250 images, grade B with 1,107 images, grade C with 867 images, grade D with 1,001 images, and grade E with 967 images. Additionally, there are 870 images featuring the V-Label, 2,328 images showcasing the BIO label, and 3,201 images without before-mentioned labels. Furthermore, we have fine-tuned the YOLOv5 segmentation model to demonstrate the practicality of using these annotated datasets, achieving an impressive accuracy of 94.0%. Discussion: These promising results indicate that this dataset has significant potential for training innovative systems capable of detecting food labels. Moreover, it can serve as a valuable benchmark dataset for emerging computer vision systems.

A randomized controlled trial to test the effects of displaying the Nutri-Score in food advertising on consumer perceptions and intentions to purchase and consume.

BACKGROUND: Some research shows that advertising for high-fat, sugar, or salt (HFSS) products is contributing to a shift in consumer preferences toward products of poor nutritional quality, leading to unhealthy nutritional intakes that increase the risk of obesity and chronic diseases. A strategy of displaying simple and understandable nutritional information (like the front-of-pack nutrition label Nutri-Score) in food messages could be an aid to help guide consumers' choice towards healthier products. METHODS: A randomized controlled experiment was conducted on 27,085 participants randomly assigned to two experimental conditions or a control condition. In both experimental conditions (independent variable: advertising messages with vs. without the Nutri-Score), participants were exposed to advertisements for diversified food products with contrasting nutritional quality and belonging to nine different food categories. Participants were then asked questions about their perception, affective evaluation, and intentions to purchase and consume the products. In the control condition, they were not exposed to the advertisements. RESULTS: Overall, interaction effects between the two variables (1) the messages with vs. without the Nutri-Score and (2) the nutritional quality of products, were significant for all dependent variables, with effect sizes between large and medium. Overall, the better the products' nutritional quality, the more positive their perceptions, affective evaluations, and intentions to buy and consume them. When the Nutri-score was displayed in advertising messages (vs. when it was not), perceptions, affective evaluation, and behavioral intentions: (1) became more positive for products of good nutritional quality (Nutri-score A and B), (2) became more negative for products of poor nutritional quality (Nutri-score D and E), (3) changed little or not at all for products of intermediate nutritional quality (Nutri-Score C). CONCLUSIONS: This research is the first in the literature to demonstrate that displaying the Nutri-Score in advertising messages assists consumers in directing their choices towards healthier foods. Regulations mandating the display of the Nutri-Score in food advertising could be an effective public health measure.

Less Favorable Nutri-Score Consumption Ratings Are Prospectively Associated with Abdominal Obesity in Older Adults.

Nutri-Score is a front-of-package (FOP) labeling designed to assist consumers in selecting healthier options at the point of purchase and ultimately enhance their health. This study aims to evaluate the association between the Nutri-Score system and incident abdominal obesity (AO) in community-dwelling older adults. A prospective cohort of 628 individuals aged ≥ 60 were recruited in Spain between 2008-2010 and were reexamined between 2015-2017. Dietary intake was evaluated utilizing a validated computerized dietary history. Food was categorized based on the Nutri-Score system into five levels from A (green, representing the best quality) to E (red, representing the poorest quality). A five-color Nutri-Score dietary index (5-CNS DI) in g/day/kg was calculated for each participant. AO was determined by a waist circumference (WC) of ≥102 cm for men and ≥88 cm for women. Logistic regression models were adjusted for the main potential confounders. During a mean six-year follow-up, 184 incident cases of AO occurred. The odds ratio (OR) and 95% confidence interval (CI) for AO, when comparing the highest and lowest quartiles of the 5-CNS DI, were 2.45 (1.17-5.14), with a p-value for trend of 0.035. In sensitivity analyses, the OR was 2.59 (1.22-5.52, p-trend: 0.032) after adjustment for WC at baseline, and 1.75 (0.74-4.18, p-trend: 0.316) after adjustment for ultra-processed food consumption. In conclusion, less favorable food-consumption ratings in the Nutri-Score are associated with incident AO in the elderly. These findings support the use of this FOP system to potentially improve metabolic health.

Checkmating the Nutri-Score: Will Food Patriotism Prevent the Harmonization of Front-of-Pack Labeling Schemes within the EU?

Front-of-pack labeling schemes are an effective but contested regulatory approach to nudge consumers towards healthy food choices. The Nutri-Score, being implemented by seven European countries, is one of the most elaborated and evidence-based examples. Therefore, the Nutri-Score has been deemed as the frontrunner within the EU Commission's attempt to harmonize front-of-pack labeling among EU member states under its Farm-to-Fork strategy (F2F) by the end of 2022. However, the endeavor is on the brink of failure due to massive resistance by Mediterranean member states and parts of the food industry capitalizing on patriotic narratives (e.g. Made in Italy). This comment investigates the Nutri-Score saga from a political and commercial determinants of health lens. It argues that an EU-wide roll-out of the label hinges on the specific interplay between political structures and stakeholder agency. As shown, the EU's weak decision-making power has been exploited by the No-Nutri-Score alliance.

A comparative analysis of nutritional quality, amino acid profile, and nutritional supplementations in plant-based products and their animal-based counterparts in the UK.

Plant-based (PB) food products have surged in popularity over the past decade. Available PB products in the UK market were extracted from NielsenIQ Brandbank and compared with animal-based (AB) counterparts in their nutrient contents and calculated Nutri-Scores. The amino acid contents of four beef products and their PB alternatives were analysed by LC-MS/MS. PB products consistently exhibited significantly higher fibre content across all food groups. Protein was significantly higher in AB products from all food groups except beef and ready meals. PB products were more likely to have higher Nutri-Scores compared to AB counterparts, albeit with greater score variability within each food group. Nutrient fortifications were primarily focused on dairy and ready meals; the most supplemented nutrient was vitamin B12 (found in 15% of all products). A higher proportion of EAAs in relation to total protein content was observed in all beef products.

The Nutri-Score Scale-A Tool for Assessing the Nutritional Quality of Processed Meat Products Available on the Polish Market.

Although meat and meat products are important sources of protein in the human diet, consumption appears to be a predisposing factor in the onset of several civilisation diseases, particularly red meat and its products. One way to reduce diet-related diseases is to guide consumers towards consciously purchasing healthier foods by including a nutrition declaration on product labels, such as by using a "front-of-pack" (FOP) labelling system. This study aimed to determine the Nutri-Score classes for processed meat products, distinguish products that are potentially better for consumers, and determine whether the refined algorithm significantly contributed to a change in product classification. An analysis of the labels of 1700 products available on the Polish market indicated that most processed meat products qualified as class D and E. Comparing the refined Nutri-Score calculation algorithm with the original algorithm resulted in a slight change in product allocation. Poultry products were ranked more favourably than red meat products. The most significant change in product allocation (by 35.2%) was achieved by reducing salt content by 30% and fat content by 10%. Among the processed meat products, some are more highly ranked and are hence considered better from a nutritional perspective than others in that group.

Nutri-Score of Meat, Fish, and Dairy Alternatives: A Comparison between the Old and New Algorithm.

Nutri-Score is a front-of-pack label that visualizes the nutritional quality of food products from most healthy (A, dark green) to least healthy (E, red). However, concerns have been raised about discrepancies between Nutri-Score labels and dietary recommendations. Therefore, the Nutri-Score algorithm has recently been adapted. To investigate the effect of the new algorithm, the Nutri-Score of plant-based meat, fish, and dairy alternatives (n = 916) was calculated with the old and new algorithms. In addition, the nutritional values of meat and milk alternatives with Nutri-Score labels A and B were compared under the old and new conditions and subsequently assessed for alignment with the criteria of Dutch dietary guidelines. The new algorithm resulted in a reduction in the number of products with labels A and B, ranging from 5% (cold cuts alternatives) to 55% (milk alternatives). The nutritional composition of products with labels A and B improved for meat alternatives (lower energy and saturated fatty acid contents; higher protein content) and milk alternatives (lower energy, salt, and sugar contents; higher protein and fiber contents). Overall, the new Nutri-Score algorithm is more in line with the Dutch dietary guidelines for plant-based meat and dairy alternatives, though challenges remain with respect to micronutrient (iron, calcium, vitamin B12), salt, and protein contents.

Consistency of the Initial and Updated Version of the Nutri-Score with Food-Based Dietary Guidelines: A French Perspective.

BACKGROUND: To help consumers to make healthier food choices, 7 European countries have implemented the front-of-pack nutrition label Nutri-Score. The algorithm was updated in 2022-2023 by the Nutri-Score European scientific committee, based on the current scientific knowledge. OBJECTIVES: The aim of this study was to investigate the consistency of the newly internationally developed algorithm with the French food-based dietary guidelines (FBDG) and compare the respective performances of the initial and updated algorithm. METHODS: Three complementary French food composition databases were used to access extensive coverage of the food supply in France (n = 46,752): the Oqali, OpenFoodFacts, and CIQUAL databases. Based on the French FBDG, a list of 41 criteria was defined by which the consistency between French FBDG and the Nutri-Score was assessed (eg, consumption of fresh vegetables is promoted in FBDG, thus the Nutri-Score should rate favorably such products). RESULTS: Of all criteria, the initial algorithm met 63% (26/41) of them, whereas the revised algorithm met 85% (35/41) of them. Improvements achieved by the updated version of the Nutri-Score in alignment with the FBDG were particularly observed for high-fat products (ie, fatty fish, nuts, and seeds), sweet products (ie, ice creams and sweet spreads), salty products (ie, savory snacks and salted nuts), dairy beverages, and beverages with artificial sweeteners. CONCLUSIONS: The Nutri-Score's updated nutrient profiling system appears to rate foods more consistently regarding the French dietary guidelines and improved the currently existing system. This work supports the implementation of the updated nutrient profiling system underlying Nutri-Score.

Understanding the influence of stakeholders on the implementation of front-of-pack nutrition labelling in the absence of public debate: the case of Portugal.

In the context of the European Commission's proposal for the harmonization of front-of-pack nutrition labels (FoPLs) across the European Union scheduled for 2023, Portugal's position on the preferred format to be adopted remains unclear. Our study aims to (1) describe the context of the political decision regarding FoPL's implementation in Portugal, (2) identify and characterize the actors involved in the debate and (3) evaluate their influence in the decision-making process and their position on the various formats of FoPLs considered. A document review complemented by a qualitative study with eight experts revealed the actors involved in the implementation of FoPLs in Portugal. The stakeholder theory was used to qualify their level of influence on the decision. To complement the experts' analysis, the views of 1014 Portuguese consumers on stakeholders' influence were collected via an online questionnaire. The results show a strong mobilization of the Portuguese Ministry of Health on the issue through the production of several technical and scientific reports. The background of the political decision related to FoPLs in Portugal highlights the commitment of some influential stakeholders in favour of the traffic light format, including one of the major Portuguese retailers, Continente. Nevertheless, in the face of public support for the Nutri-Score from a growing number of actors, Portugal has not yet adopted a specific format, partly due to some resistance from the agricultural sector. Portugal, by lack of consensus and agenda setting, is now awaiting formal harmonization of FoPLs by the European Commission. This study provides useful insights for public policy actors when promoting interventional measures.

Complementarity between the updated version of the front-of-pack nutrition label Nutri-Score and the food-processing NOVA classification.

OBJECTIVE: To compare the initial and the updated versions of the front-of-pack label Nutri-Score (related to the nutritional content) with the NOVA classification (related to the degree of food processing) at the food level. DESIGN: Using the OpenFoodFacts database - 129,950 food products - we assessed the complementarity between the Nutri-Score (initial and updated) with the NOVA classification through a correspondence analysis. Contingency tables between the two classification systems were used. SETTINGS: The food offer in France. PARTICIPANTS: Not applicable. RESULTS: With both versions (i.e. initial and updated) of the Nutri-Score, the majority of ultra-processed products received medium to poor Nutri-Score ratings (between 77·9 % and 87·5 % of ultra-processed products depending on the version of the algorithm). Overall, the update of the Nutri-Score algorithm led to a reduction in the number of products rated A and B and an increase in the number of products rated D or E for all NOVA categories, with unprocessed foods being the least impacted (-3·8 percentage points (-5·2 %) rated A or B and +1·3 percentage points (+12·9 %) rated D or E) and ultra-processed foods the most impacted (-9·8 percentage points (-43·4 %) rated A or B and +7·8 percentage points (+14·1 %) rated D or E). Among ultra-processed foods rated favourably with the initial Nutri-Score, artificially sweetened beverages, sweetened plant-based drinks and bread products were the most penalised categories by the revision of Nutri-Score while low-sugar flavoured waters, fruit and legume preparations were the least affected. CONCLUSION: These results indicate that the update of the Nutri-Score reinforces its coherence with the NOVA classification, even though both systems measure two distinct health dimensions at the food level.

The effects of health-related food taxes on the environmental impact of consumer food purchases: secondary analysis of data from a randomised controlled trial in a virtual supermarket.

OBJECTIVE: To measure the effects of health-related food taxes on the environmental impact of consumer food purchases in a virtual supermarket. DESIGN: This is a secondary analysis of data from a randomised controlled trial in which participants were randomly assigned to a control condition with regular food prices (n 152), an experimental condition with a sugar-sweetened beverage (SSB) tax (n 131) or an experimental condition with a nutrient profiling tax based on Nutri-Score (n 112). Participants were instructed to undertake their typical weekly grocery shopping for their households. Primary outcome measures were three environmental impact indicators: greenhouse gas (GHG) emissions, land use and blue water use per household per week. Data were analysed using linear regression analyses. SETTING: Three-dimensional virtual supermarket. PARTICIPANTS: Dutch adults (≥ 18 years) who were responsible for grocery shopping in their household (n 395). RESULTS: GHG emissions (-7·6 kg CO2-eq; 95 % CI -12·7, -2·5) and land use (-3·9 m2/year; 95 % CI -7·7, -0·2) were lower for the food purchases of participants in the nutrient profiling tax condition than for those in the control condition. Blue water use was not affected by the nutrient profiling tax. Moreover, the SSB tax had no significant effect on any of the environmental impact indicators. CONCLUSIONS: A nutrient profiling tax based on Nutri-Score reduced the environmental impact of consumer food purchases. An SSB tax did not affect the environmental impact in this study.

Chemical Data and Relationships for a Scoring Algorithm of Extra Virgin Olive Oil's Nutritional Value.

Extra virgin olive oil (EVOO) is a valuable product and is highly appreciated by consumers for its great nutritional value. However, to date, there has been a lack of uniform systems capable of ranking the nutritional value of EVOO based on its chemical composition in terms of macro- and micronutrients (including phenolic compounds and tocopherols). The aim of this study was to propose a scoring algorithm to rank the nutritional value of EVOO samples, considering their chemical composition in macro- and micronutrients and their sensitivity to oxidation phenomena. Data from more than 1000 EVOO samples were used to assess the variability of the data, considering the selected negative parameters (free acidity, peroxide value, spectrophotometric indices) and positive components (composition in tocopherols via HPLC-DAD, phenolic compounds via HPLC-DAD, and fatty acids via GC-MS) so as to ensure the universal validity of the scoring algorithm. The dataset included samples from the main producing countries worldwide, in addition to Australia, across several production years; data were selected to represent different production realities. A mathematical model was set up for each chemical component, resulting in six variable values. By combining these values with a dimensionless constant value, the algorithm for computing the nutritional value score (NVS) was defined. It allows the nutritional value of an oil to be ranked on a scale of 0 to 100 based on its chemical composition. The algorithm was then successfully tested using chemical data from about 300 EVOO samples obtained from laboratories from different Italian regions. The proposed NVS is a simple and objective tool for scoring the nutritional value of an EVOO, easy to understand for both producers and consumers.

Dietary index based on the Food Standards Agency nutrient profiling system and risk of Crohn's disease and ulcerative colitis.

BACKGROUND: Nutri-score is now widely available in food packages in Europe. AIM: To study the overall nutritional quality of the diet in relation to risks of Crohn's disease (CD) and ulcerative colitis (UC), in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort METHODS: We collected dietary data at baseline from validated food frequency questionnaires. We used a dietary index based on the UK Food Standards Agency modified nutrient profiling system (FSAm-NPS-DI) underlying the Nutri-Score label, to measure the nutritional quality of the diet. We estimated the association between FSAm-NPS-DI score, and CD and UC risks using Cox models stratified by centre, sex and age; and adjusted for smoking status, BMI, physical activity, energy intake, educational level and alcohol intake. RESULTS: We included 394,255 participants (68.1% women; mean age at recruitment 52.1 years). After a mean follow-up of 13.6 years, there were 184 incident cases of CD and 459 incident cases of UC. Risk of CD was higher in those with a lower nutritional quality, that is higher FSAm-NPS-DI Score (fourth vs. first quartile: aHR: 2.04, 95% CI: 1.24-3.36; p-trend: <0.01). Among items of the FSAm-NPS-DI Score, low intakes of dietary fibre and fruits/vegetables/legumes/nuts were associated with higher risk of CD. Nutritional quality was not associated with risk of UC (fourth vs. first quartile of the FSAm-NPS-DI Score: aHR: 0.91, 95% CI: 0.69-1.21; p-trend: 0.76). CONCLUSIONS: A diet with low nutritional quality as measured by the FSAm-NPS-DI Score is associated with a higher risk of CD but not UC.

Dietary indices underpinning front-of-pack nutrition labels and health outcomes: a systematic review and meta-analysis of prospective cohort studies.

BACKGROUND: Nutrient profiling systems are increasingly used to characterize the healthfulness of foods for front-of-package (FOP) labeling, which have been proposed as an effective public health strategy to help people make healthier food choices. OBJECTIVE: This study aimed to review available evidence from cohort studies that evaluated the association of dietary indices underpinning FOP nutrition labels with all-cause mortality and incidence of cardiovascular diseases (CVDs) or cancer. METHODS: PubMed, Web of Science, and Scopus were systematically searched up to October 2023. We included articles if they were prospective cohort studies, if the exposure was any dietary index underpinning FOP nutrition labels [e.g., the modified Food Standard Agency-Nutrient Profiling System (FSAm-NPS) and the Health Star Rating System], and if outcomes were all-cause mortality or incidence of or mortality due to CVD and cancer. Random-effects models were used to calculate the pooled hazard ratios (HRs) and 95% CIs. RESULTS: We identified 11 records (7 unique prospective studies), which were included in the systematic review. The meta-analysis comprised 8 studies analyzing the FSAm-NPS dietary index (DI) as exposure. The pooled HRs associated with a 2-unit increase in the FSAm-NPS DI of all-cause mortality, CVD, and cancer risk were 1.06 (95% confidence interval [CI]: 0.99, 1.13; I2: 80%), 1.08 (95% CI: 1.00, 1.18; I2: 70%), and 1.09 (95% CI: 1.00, 1.19; I2: 77%), respectively. The Chilean Warning Label score and the Health Star Rating systems were examined by 1 study each and were significantly associated with the outcomes. CONCLUSIONS: DIs underpinning most common FOP nutrition labels and reflecting nutrient-poor diets show a tendency toward an increased incidence of CVD and cancer, but the observed effects are quite modest in magnitude. Further studies at the population level are needed to support the widely shared hypothesis that FOP labels, possibly in conjunction with other interventions, may contribute to reduce noncommunicable disease risk. This meta-analysis was registered at PROSPERO as CRD42021292625.

The impact of nutritional labeling on adult snack choices: A controlled field experiment in a non-commercial professional setting.

Each year, 2.8 million people die because of comorbidities associated with being overweight. Snacking substantially contributes to people's calorie intake. One way to nudge consumers towards healthier alternatives is the implementation of nutritional labeling. This study reports on a controlled field experiment that evaluates the effect of two nutritional labels on free snack choices (n = 739). Participants at a conference could choose between nuts, cookies, and candy bars as a snack, presented at the bar at six different locations across 2 bar counters. The labels were set up in front of each snack in three conditions: no labeling (control), a calorie label, or a traffic light label (i.e., the Nutri-Score). The location of the snacks on the counter (Either side of the counter; Center, Right, Left) and the time-of-day (Morning (=reference) vs Afternoon) were statistically controlled for. The results show that calorie labels could not successfully nudge consumers toward healthier snack choices (nuts instead of candy bars or cookies). In contrast, the Nutri-Score label significantly increased the probability of choosing nuts over candy bars. The Nutri-Score also increased the chance of choosing nuts over cookies, but the difference was not significant. No prior studies to our knowledge have directly compared calorie labels to the Nutri-Score. This study suggests that the Nutri-Score label can be a more successful intervention than calorie labels to nudge consumers towards healthier choices in situations were free snacks are offered, like many modern workplaces. Changing snacking behavior is challenging and naturalistic field experiments like this one are needed to translate the theory from previous laboratory studies to real-life settings.

Using machine learning models to predict the quality of plant-based foods.

Plant-based foods (PBFs) are considered healthy, especially, minimally processed whole foods, fruits, whole grains, and legumes while highly processed PBFs maybe less nutritious. Educating consumers on nurient quality will help to guide their choices. This study was aimed at estimating and predicting the nutrient quality of PBFs using their Nutri-Score and micronutrient content. The NHANES (2017-2020) data shows the output for foods consumed in the US and their nutrient composition based on a 24-h recall. Though the Nutri-Score label has been used to discriminate food quality, it still needs to be implemented in most countries. It computes mostly macronutrients with less consideration for micronutrients which also contributes to product quality. ML methods used in this study combine the Nutri-Score grade and micronutrient content in predicting food quality. The FNDDS data of PBFs for 2017-2020 were split into training (n = 300) and testing (n = 74) datasets. Eight ML models were used to predict the Nutri-Score and the Nutri-Score grade of PBFs. Random forest (RF) and light gradient boost model (LightGBM) performed best with accuracy and coefficient of determination (R2) scores of 0.88 and 0.96, respectively, while DT had the least scores in predicting the Nutri-Score grade (0.81) and Nutri-Score (0.93). These results suggest that ML can be effectively leveraged to predict PBFs quality.