Study population

The NutriNet-Santé study (https://info.etude-nutrinet-sante.fr) is an ongoing web-based prospective observational cohort study launched in France in May 2009 with a scheduled follow-up of 10 years. It aims to investigate the relationship between nutrition and chronic disease risk, as well as the determinants of dietary behavior and nutritional status. The study was implemented in the French general population (internet-using adult volunteers, aged ≥18years). The rationale, design and methodology of the study have been fully described elsewhere [29]. In brief, to be considered as included into the study, participants have to complete a baseline set of self-administered, web-based questionnaires assessing dietary intake, physical activity, anthropometric characteristics, lifestyle, socioeconomic conditions and health status. As part of the follow-up, participants are asked to complete the same set of questionnaires each year. Moreover, each month, participants are invited by e-mail to fill in optional questionnaires related to dietary intake, determinants of eating behaviors, nutritional and health status. This study is conducted in accordance with the Declaration of Helsinki, and all procedures were approved by the Institutional Review Board of the French Institute for Health and Medical Research (IRB Inserm n°0000388FWA00005831) and the Commission Nationale de l’Informatique et des Libertés (CNIL n°908450 and n°909216). All participants provided informed consent with an electronic signature. This study is registered in EudraCT (n°2013-000929-31).

Label formats

Four labels (Fig 1) were tested in the study, providing simple or more detailed nutritional information along with a simple positive, neutral, or negative opinion of the product. In the introduction to the questionnaire, the different labels were presented and briefly explained to the participants. A detailed description of the labels used in the present study is provided below.

Nutrient-specific formats:

1. Guideline Daily Amounts (GDA): this label indicates the kilocalories and the amount of fat, saturated fatty acids, sugars and sodium in gram per portion as well as their corresponding contribution (in percentages) to the guideline-based daily intakes [22]. The guideline-based daily intakes indicate the recommended intake in kilocalories (kcal/day) and nutrients (g/day). The GDA information was calculated by using the Food and Drink Federation’s guiding principles and was based on the average nutrient requirements for an adult woman. This label can be found on most of the food packaging on the French market, following the voluntary initiative of manufacturers.
2. Multiple Traffic Lights (MTL): this label introduced by the UK Food Standard Agency (FSA), provides an evaluation of nutrient content regarding fat, saturated fatty acids, sugars and sodium. Depending on the quantity of the nutrient in the product (high, medium, low), a color is attributed to each nutrient (red, amber, green, respectively). The colors reflect the concentration in grams per 100 g or per 100 ml of product and the criteria of the FSA were applied to assign the color codes [23].

Summary formats:

1. 3. Five-Color Nutrition Label (5-CNL): this label was proposed to be introduced to the French market to guide consumer food choices [26]. Based on the FSA nutrient profiling system [30], used by the Office of Communication (OfCom) for regulation of advertising to children in the UK, and with specific adaptations for cheese and added fat, this label provides information about the overall nutritional quality of a given food item. The label is represented by a scale of five colors (from green to red) with corresponding letters (from A to E) [26]. The colors were attributed depending on the score of each food given by the FSA nutrient profiling system (S1 Fig) (26): ‘Green’ (-15 to -2 points), ‘Yellow’ (-1 to 3 points), ‘Orange’ (4 to 11 points), ‘Pink’ (12 to 16 points) and ‘Red’ (> = 17 points) [26].
2. 4. Green Tick (Tick): this label was derived from the “Keyhole” and “Pick the Tick” symbols, developed by the Swedish Food Administration and the Heart Foundation in Australia and New Zealand, respectively [25,31]. It reflects the overall nutritional quality of the food item and appears only on the healthier products within a food family. The Tick label was attributed to products assigned to the green or yellow categories of the 5-CNL.

Reference

1. 5. No label: a situation without any FOP labels was used as reference.

Data collection

Objective understanding and acceptability of the different labels were assessed in July 2014 via a web-based questionnaire.

Statistical analyses

We performed analyses using data from participants of the NutriNet-Santé cohort who had completed the questionnaire on FOP labels. We excluded participants who had responded “I don’t know” to more than two thirds of the cases. We compared included and excluded participants using Student’s t-tests or chi-square tests, as appropriate.

To assess label acceptability, the percentage of support for each label was calculated for all variables related to liking, attractiveness and perceived cognitive workload. To compare label performance with respect to the participants’ understanding of nutritional quality, the percentage of correct answers was calculated and chi-square tests were performed. Differences across product categories and across subgroups (e.g. sex, age) were also assessed by using chi-square tests.

The data were weighted according to the French population socio-demographic distribution. Weighting was calculated separately for each sex using an iterative proportional fitting procedure and the 2009 national Census data on age, educational level, area of residence and whether or not the household included any children [32].

All tests of significance were two-sided, and a P value <0.05 was considered significant. Statistical analyses were performed using SAS software (version 9.3; SAS Institute Inc.).

Characteristics of the sample

A total of 15,002 participants completed the questionnaire. A total of 772 participants were excluded because in more than two thirds of the cases they had responded “I don’t know”; and 652 participants were excluded because of missing data required for the weighted analyses. This left 13,578 participants available for analyses. Characteristics of included participants are presented in Table 1. Compared with excluded participants, included participants were more often women (p<0.0001), younger (p<0.0001), and had a higher educational level (p<0.0001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Socio-demographic characteristics of the sample (n = 13,578) (The NutriNet-Santé study, 2014).

https://doi.org/10.1371/journal.pone.0140898.t001

Acceptability of front-of-pack labels

The results on the acceptability of the different FOP labels are presented in Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Acceptability of the different labels (n = 13,578) (The NutriNet-Santé study, 2014).

https://doi.org/10.1371/journal.pone.0140898.g002

Objective understanding of front-of-pack labels

Results regarding the objective understanding, tests for differences across product categories, and socio-demographic sub-groups are presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Percentage of correct answers by label, across product categories and socio-demographic characteristics (n = 13,578) (The NutriNet-Santé study, 2014).

https://doi.org/10.1371/journal.pone.0140898.t002

Acceptability of front-of-pack labels

The GDA was declared to be the preferred label. Previous studies have reported mixed results on consumers’ perception of GDA. In line with our findings, one study conducted in Canada showed that consumers preferred GDA compared to other labels including MTL and simple summary labels [33]. By contrast, other previous studies conducted among shoppers from New-Zealand and Europe found that GDA was the least preferred label [3,13]. A potential explanation for this discrepancy is that the GDA label had been in use on the Canadian and French markets for a long time when studies were performed. The GDA might therefore be preferred because it is familiar. In our study, the GDA label was the one contributing the most to needed information and also the most reliable according to the subjects, which could also explain why it was liked. Despite these strengths, the GDA label presented some weaknesses already highlighted in the literature: it was perceived by the participants as the one requiring the highest cognitive workload in terms of complexity and needed processing time [3,5,15]. In addition, subjects indicated that it was the least easy to identify.

By contrast, the 5-CNL was found to be the easiest label to identify, requiring the lowest cognitive workload in terms of time and complexity. Consistent with these results, other studies have shown that consumers’ processing time was faster when they were exposed to summary labels compared with nutrient-specific labels [3,5,14,15]. In particular, one study including a graded labeling system found that such a label allowed the consumer to identify the healthier product faster than did the other labels [3]. Another explanation to account for the higher performance of the 5-CNL might be its color-coded format and the associated letters which have been shown to focus attention [34] and to help consumers process FOP labels more quickly [5,35]. Focusing the consumer’s attention is of major importance because it is the first step when leading individuals to use nutritional information and making healthier food purchases [4]. Moreover, the label’s ease of identification is all the more important given that in a real shopping situation, consumers are making decisions quickly: European consumers spend, on average, between 25 and 47 seconds when choosing a product in a supermarket [36].

In contrast with previous studies showing that MTL was appreciated by consumers because it was rapid and easy to identify, understand and use, compared with both complex formats (such as GDA) and simple summary labels (such as Tick) [4,12,17], in our study, MTL showed no greater performance compared to other labels. One hypothesis to explain this result is that few studies have compared MTL with graded label formats (such as 5-CNL). Among the tested labels, the MTL and 5-CNL entailed both positive and negative opinions. When comparing these labels, a potential explanation for the fact that the 5-CNL label performed better than did the MTL could be that consumers appreciated simplification [37,38].

However, in their review, Grunert and Wills highlighted that even if consumers appreciate simplification, they would like to know what the simplified information reflects [4], which could explain why Tick was the least preferred label.

Besides, Tick was the least likely to elicit discomfort compared with other labels. Nonetheless, the overall discomfort provoked by the labels was low for all labels. Thus, according to our study, labels providing both positive and negative evaluation of foods (i.e. MTL or 5-CNL) do not cause any more discomfort to the consumer compared with those providing a neutral evaluation, such as the GDA.

Objective understanding of front-of-pack labels

In line with previous research, we found that FOP labels were an effective tool to help consumers identify food products with the best nutritional quality compared with a no-label condition [5,11,12,16,20,21] and that a graded label format, the 5-CNL, performed better than did other FOP labels [3]. As discussed above, a potential explanation for the better performance of the 5-CNL label might be that these formats include color and text, which have been shown to enhance consumers’ ability to compare the nutritional quality across different products [5,12,38].

On the whole, we observed the second higher percentage of correct answers for the MTL labeling system, followed by GDA and Tick labels. Consistent with our results, a recent review has shown that color-coded labels such as MTL performed better than did GDA to compare products in terms of their healthfulness [5].

In our study, the Tick label demonstrated low performance whereas in other contexts, simple summary labels have been identified as effective systems allowing the comparison across products in terms of their overall nutritional quality [3,13,14,39]. This discrepancy might be explained by the fact that most of the studies in the literature asked participants to compare only two products. In these cases, a summary label format indicating the healthiest product might easily allow subjects to compare products in terms of their healthfulness if only one of the two products has the label. In turn, the design of our study, including a set of three products, highlighted this limit inherent in simple summary labeling systems.

In line with a previous study which included three graded systems (Stars, Smileys and Health Protector Factor), along with MTL, GDA and a simple summary format, we found that the graded label (i.e. 5-CNL) was the label format which exhibited the best effectiveness across product categories [3].

The evaluation of label understanding across population subgroups (sex, age, educational level, occupational category) showed that the 5-CNL label was a particularly effective label allowing the comparison of products according to their nutritional quality. Among subjects with lower educational levels, who have been identified as having potentially more difficulties in understanding nutrition labels [2], the 5-CNL label was the one which demonstrated the greatest performance in terms of increasing the number of correct answers compared with other labels. These results are in line with the literature: in comparison with nutrient-specific labeling systems, simpler FOP labels have been shown to be more effective to help individuals with low formal education to identify healthier products [5,6,14,18].

Strengths and limitations

This study provides new insights on graded summary labels which have been rarely studied in comparison to nutrient-specific and simple summary formats, whereas the Institute of Medicine, after reviewing the literature, recommended the use of such graded summary labels [14]. Another strength of this study was its large sample size including subjects with various socio-demographic profiles. In addition, we used a weighting scheme to be able to apply these results to the general French population. To avoid any potential effects of the order in which the labels was presented, we set up a rotation system. Possible interactions with the product categories were also taken into account by using all label/product category combinations. Next, we used an objective measure of participant understanding and we developed an experiment including three different products, which is more similar to a real shopping situation than a choice between two products, and also limits the impact of random responses. This design allowed us to identify the limitations of a simple summary label, which provides information on the healthiest product but which failed to differentiate other products. Some limitations in the present study should be noted. First, because the subjects were volunteers in a nutritional study, they may have greater nutrition knowledge and be more interested in nutritional issues. Thus, even though our analyses were weighted on socio-demographic variables, caution is needed when generalizing our results. Secondly, self-reported data may not fully reflect real shopping situations where environmental parameters such as noise, time pressure, surge of commercial messages, prices and special offers may condition the perception, use and understanding of the labels [8,40,41]. Next, compared to other studies, we found a lower number of correct answers regarding nutritional quality. Two hypotheses might be formulated to explain this finding. First, as previously discussed, our test included three products whereas most of the studies asked the participants to compare only two products, which is an easier task. Further, it is possible that some individuals misunderstood the statement « from your point of view » and did not use the labels to order the products according to their nutritional quality leading to a lower rate of correct answers. In addition, when products are correctly ranked, this would be due to knowledge of product or nutrition rather than due to the label, and therefore do not reflect true understanding. Finally, label familiarity was not assessed although this is likely to influence participants’ understanding. However, the GDA label which is the only label in use on the French market and is estimated to be provided on about 20% of the front of food packages in 2010 [42] did not perform better than other, suggesting a limited effect of familiarity.