You are here

  • Home
  • Archive
  • Volume 13, Issue 4
  • Effects of front-of-package nutrition labelling systems on understanding and purchase intention in Jamaica: results from a multiarm randomised controlled trial

Article Text

Article menu
Public health
Original research
Effects of front-of-package nutrition labelling systems on understanding and purchase intention in Jamaica: results from a multiarm randomised controlled trial
  1. Vanessa White-Barrow1,
  2. Fabio S Gomes2,
  3. Sheerin Eyre1,
  4. Gaston Ares3,4,
  5. Audrey Morris5,
  6. Deonne Caines6,
  7. David Finlay1
  1. 1College of Health Sciences, University of Technology Jamaica, Kingston, Saint Andrew Parish, Jamaica
  2. 2Noncommunicable Diseases and Mental Health, Pan American Health Organization, Washington DC, District of Columbia, USA
  3. 3Departamento de Ciencia y Tecnología de Alimentos, Facultad de Química, Universidad de la Republica Uruguay, Montevideo, Uruguay
  4. 4Centro de Investigación Básica en Psicología, Facultad de Psicología, Universidad de la Republica Uruguay, Montevideo, Uruguay
  5. 5Noncommunicable Diseases and Mental Health, Pan American Health Organization, Kingston, Jamaica
  6. 6Ministry of Health & Wellness of Jamaica, Government of Jamaica, Kingston, Jamaica
  1. Correspondence to Dr Fabio S Gomes; gomesfabio{at}paho.org

Abstract

Objective To assess the effects of three different front-of-package labelling (FOPL) schemes on objective understanding and intention to purchase of products, in Jamaica.

Setting Supermarkets in Jamaica.

Participants Adult supermarket shoppers in Jamaica (n=1206) aged 18 years old or older were included in the study, except for those visually impaired, or unable to give informed consent.

Design Multiarm parallel-group randomised trial.

Interventions Participants were randomly allocated to one of the three intervention groups or the control group. They were exposed to two-dimensional images of 12 mock-up products presented in random and balanced order. Participants assigned to the intervention groups were exposed to one FOPL scheme: black octagonal warning labels (OWL), magnifying glass high-in single icon (MGG) or traffic-light labelling (TFL). The control group was exposed to the nutrition facts up front.

Outcome measures OR for correctly understanding nutritional information (correctly selecting the least harmful option, correctly identifying sugars, sodium and/or saturated fats found to be in excess) and choosing to purchase the least harmful option (purchase intention), more often.

Results Compared with the control group, the odds for correctly selecting the least harmful option more often were 107% higher in the OWL group (OR 2.07, 95% CI 1.54 to 2.78; p<0.001), whereas the MGG (1.18, 95% CI 0.89 to 1.57; p=0.24) and the TFL (1.13, 95% CI 0.85 to 1.51; p=0.39) were inefficacious in improving such odds. OWL also resulted in the highest odds for correctly identifying a product with excessive amounts of sugars, sodium and/or saturated fats and for deciding to purchase the least harmful option or none of the options.

Conclusions Octagonal warning labels performed best at improving the ability of adult shoppers in Jamaica to understand the nutrition information and at encouraging them to purchase the least harmful option more often.

  • health policy
  • public health
  • nutrition & dietetics

Data availability statement

No data are available.

https://creativecommons.org/licenses/by-nc/3.0/igo/

This is an open access article distributed under the terms of the Creative Commons Attribution Non-commercial IGO License (https://creativecommons.org/licenses/by-nc/3.0/igo/), which permits use, distribution, and reproduction for non-commercial purposes in any medium, provided the original work is properly cited. In any reproduction of this article there should not be any suggestion that PAHO or this article endorse any specific organization or products. The use of the PAHO logo is not permitted. This notice should be preserved along with the article's original URL. Disclaimer: The author is a staff member of the Pan American Health Organization. The author alone is responsible for the views expressed in this publication and they do not necessarily represent the views, decisions or policies of the Pan American Health Organization.

https://doi.org/10.1136/bmjopen-2022-065620

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    STRENGTHS AND LIMITATIONS OF THIS STUDY

    • The study used a between-subject design which allows the results to be attributed to front-of-package labelling (FOPL) schemes and avoids differential carry-over effects that are more likely to happen in within-subject studies.

    • The exposure of participants to two-dimensional mock-up products with different FOPL schemes was standardised to match real products’ label sizes and sizes and proportions of FOPL, resembling real-life conditions.

    • The study was conducted using fictious brands, products were not associated with price information and participants did not purchase the products, which has strengthened the attribution of the effects to the FOPL schemes, but conversely, limited the analysis of the relative importance of these other factors.

    Background

    The prevalence of overweight, obesity and non-communicable diseases (NCDs) continues to increase in the Americas among all age groups.1 This region has the highest prevalence of overweight and obesity in the world: the global prevalence of overweight and obesity in adults is 39%, whereas in the Americas it is 64.1% among men and 60.9% among women. NCDs are the major cause of disability and premature death, and in 2016, they were responsible for 80.7% of all deaths in the region.2 Thirty-four per cent of these NCD-related deaths occurred prematurely in people between the ages of 30 and 69 years, when people have the most economically productive time of their life.3 In particular, Jamaica is among the countries with the highest rates of overweight and obesity and NCDs in the Americas. In 2016–2017, the latest estimate available, the prevalence of overweight and obesity in male and female adults was 38.8% and 67.6%, respectively.4 In 2016, 80% of deaths in Jamaica were caused by NCDs.2

    Unhealthy eating has been identified as one of the main modifiable causes of this situation,5 and the expansion of unhealthy diets has been largely driven by the consumption of ultra-processed products and processed products that are energy-dense and contain excessive amounts of nutrients associated with NCDs (ie, sugars, sodium, fats, saturated fats and trans fats).6–11 For these reasons, the development and adoption of policies that create supportive food environments that enable the population to improve their diets are paramount.5 Provision of information about the content of products is usually regarded as a core policy for encouraging healthier food decisions.12 Although nutrient declarations have been compulsorily implemented in several countries worldwide, a large body of literature has shown that people find it difficult to detect and understand this type of information and that they seldom use it for making their food purchases.13–18

    Considering that people spend little time and cognitive effort when taking their food purchase decisions, the inclusion of simplified nutrition information schemes can improve their ability to find and understand nutrition information, encouraging informed food choices.17–20 For this reason, the adoption of front-of-package (FOP) nutrition labelling has been identified as a priority globally.21 FOP nutrition labelling intends to increase attention to nutrition information at the point of purchase and to subsequently improve consumer ability to understand it.22 Increased attention and understanding are expected to encourage consumers to take the nutritional composition of products into account for their food purchase decisions, acting like a nudge towards healthier choices.22 23 In particular, FOP nutrition labelling is expected to discourage purchase and consumption of products excessive in nutrients associated with NCDs, which would lead to changes in nutrient intake in the middle term.16 24 25 This sequence of steps is expected to contribute to reducing the prevalence of obesity and NCDs in the long term, as shown by several modelling studies.26–28

    Several FOP nutrition labelling schemes have been developed worldwide, but they differ in purpose and performance.29 Non-interpretive schemes (eg, guidelines daily amounts, facts up front) only intend to provide numerical information about the content of nutrients, whereas interpretive schemes aim at facilitating the interpretation of nutrition information by including graphical or textual information.30 Among interpretive schemes, two main types can be identified: nutrient-specific schemes (eg, traffic-light system or warning labels) that provide simplified information about specific nutrients of concern and summary indicators of overall nutritional quality (eg, Health Star Rating and Nutri-Score).30

    In the Americas, the Caribbean and in Jamaica, where almost half of deaths are caused by high fasting blood glucose, hypertension and overweight and obesity, the main objective sought to be met with FOP labelling is to allow consumers to easily and correctly identify products that are excessive in critical nutrients associated with those risk factors, including sugars, fats, saturated fats, trans fats and sodium. In addition, FOP labelling is expected to encourage consumers to make more informed and healthier food choices, leading to a reduction in the consumption of fats, sodium and/or sugars.

    Jamaica’s existing Standard Act and Regulation31 form the legislative framework for labelling, and the country is still to incorporate provisions on the application of FOP labelling to fulfil its commitments to national, subregional and regional mandates. As part of the country’s NCD response, a National Task Force was assembled under the auspices of the Minister of Health with the aim of achieving improvements in the food environment, including the adoption of FOP labelling for prepackaged food products to provide adequate and understandable product and nutrition information to consumers. In addition, Jamaica is also taking part in a subregional standard development process that the Caribbean Community (CARICOM) Regional Organisation for Standards and Quality (CROSQ) has been leading, since 2018, to incorporate FOP labelling specifications into a revision of CARICOM Regional Standard for Specification for labelling of prepackaged foods (CRS 5:2010).32 This has involved extensive consultations at country level through National Mirror Committees established by the National Bureaus of Standards with representation of government, industry, commerce and civil society organisations. During this process, the FOP octagonal warning labelling system was proposed as the best fit for purpose based on the evidence, and other proposals were put forward by representatives of industry during consultations.

    This research, which is the first study of its kind in the Caribbean, was designed to add another piece to the existing body of evidence to help inform policy decisions. The study compared the octagonal warning labels (OWL) included in the CROSQ proposal for adoption by the CARICOM, the ‘high in’ single icon FOP labelling system known as the magnifying glass (MGG) and the traffic-light labelling (TFL), using nutrition facts up front as a control. These other FOP labelling schemes were included in the study since they were suggested by industry representatives along the regulatory process as alternatives to OWL. The comparisons aimed at assessing the effect of these FOP labelling schemes on objective understanding (correct identification of the least harmful product and correct identification of excessive nutrient content), and intention to purchase of a series of products, in Jamaica. Results are also expected to contribute to the evaluation of the effect of graphical design on the effectiveness of warning labels, a topic that has been increasingly receiving more attention in the literature more recently.33–39 The hypotheses tested in the study include: (1) OWL, MGG and TFL improve consumers’ ability to correctly identify the least harmful product more often; (2) these FOP labelling schemes improve consumers’ ability to correctly identify the presence of excessive nutrient content in products more often; (3) and these FOP labelling schemes increase consumers’ intention to purchase the least harmful option or none of the options of a series of products, more often. These hypotheses have been tested both against the control condition, and between the FOP labelling schemes, to verify whether there is a scheme that performs differently from the others for each hypothesis.

    Methods

    Design

    A multiarm parallel-group randomised trial was conducted among adult shoppers at supermarkets across Jamaica. Participants were randomly allocated at equal rate (1/4) to the four study groups (three experimental and the control group). Participants in each group were exposed to either one of the experimental conditions or allocated to the control group.

    Participants

    Adult supermarket shoppers in Jamaica (n=1206) aged 18 years old or older were included in the study, except for those visually impaired, or unable to give informed consent. Participants were recruited and interviewed at popular supermarkets serving customers of varying socioeconomic status, between December 2020 and February 2021. The 24 supermarkets included in the study were located in different parishes across Jamaica (Clarendon n=2, Hanover n=2, Kingston and St Andrew n=4, St Ann n=2, St Catherine n=5, St Elizabeth n=2, St James n=3, St Mary n=2 and St Thomas n=2). Shoppers were individually invited by interviewers to participate in a survey about food packages while they were entering the supermarket. They were informed that the study aimed to find out how people in Jamaica perceive food packages, and that they would be presented with a series of pictures of food packages and asked to answer simple questions that would take about 15 min to respond in total. An information sheet providing more detailed information about the study was also read by the interviewer, if additional information was required by the recruit. Participants were interviewed consecutively by trained field researchers after providing their signed informed consent, and the interviews took place before they entered the supermarket, during the supermarket opening hours, on weekdays.

    Interventions

    Participants in the experimental groups were shown two-dimensional (2D) images of 12 different mock-up products presented in random and balanced order between and within categories of products. The images were printed in A3 size booklets. Figure 1 illustrates one of the pages of a booklet of images shown to participants of one of the experimental groups.

    Figure 1
    Figure 1

    Example of images of one category of products shown to participants assigned to one of the front-of-package labelling groups (traffic-light labelling). Images developed by coauthors, Carlos Felipe Urquizar Rojas and Carla Galvão Spinillo, and designed by Carlos Felipe Urquizar Rojas and Carla Galvão Spinillo.

    The mock-up products did not correspond to real commercial products available in the Jamaican market but had similar characteristics in terms of graphical design and nutritional composition. Four sets of mock-ups were designed. Each set included three (3) products from each of four (4) product categories of ultra-processed products commonly consumed (3×4=12 mock-up products). The product categories were breakfast cereal extrudates, chocolate milk, cream crackers and yoghurt.

    The same 12 mock-up products were used in each group;f the only difference across groups was the FOP labelling scheme they were featuring. Mock-ups shown to participants featured solely one of the following front-of-package labelling (FOPL) schemes, according to the group they were allocated to: black octagonal warning labels (OWL group), magnifying glass high-in single icon (MGG group), traffic-light labelling (TFL group) or the nutrition facts up front (control group).

    The application of TFL and the nutrition facts up front (control condition) followed the specifications developed by the UK Department of Health (DH), the Food Standards Agency (FSA), and devolved administrations in Scotland, Northern Ireland and Wales in collaboration with the British Retail Consortium.40 The nutrition symbol that features an MGG proposed in the public consultation launched by Health Canada was used to depict the high-in single icon FOP label.41 The specifications used for the application of black OWL followed the CROSQ proposed standard.32 For consistency, thresholds used to define the ‘high’ content of sugars, fats, saturated fats or sodium, was the same for all FOPL systems, and the Pan American Health Organization (PAHO) nutrient profile model criteria included in the CROSQ proposed standard was the one used.42 The thresholds for low and medium levels of nutrients used in the TFL were those established by the UK/DH/FSA, and devolved administrations in Scotland, Northern Ireland and Wales in collaboration with the British Retail Consortium.40 See the nutritional composition of products in the supplemental material (online supplemental table S1). All sets of mock-up products were identical except for the FOPL icons featured. Figure 2 illustrates one of the mock-up products with the FOPL schemes applied.

    Figure 2
    Figure 2

    Example of a product from each of the FOPL groups. (A) Nutrition facts up front (control condition); (B) single icon high in FOPL with magnifying glass; (C) traffic-light labelling scheme; (D) octagonal warning label. Images developed by coauthors, Carlos Felipe Urquizar Rojas and Carla Galvão Spinillo, and designed by Carlos Felipe Urquizar Rojas and Carla Galvão Spinillo. FOPL, front-of-package labelling.

    The order of the questions aimed at reducing potential response bias. Participants were first shown the three products from each of the four product categories and asked to indicate which product they would buy in each category. Participants could also indicate that they would not buy any of the options within each category. The four product categories were shown one by one. Both the categories and the products within the categories were presented in random and balanced order.

    Participants were then shown the same products in different order of categories, and different order within categories. For each set of three products within a category they were asked to indicate which of those was the least harmful for health.

    For the last task, participants were shown one product of each category, in a random and balanced order, and asked to indicate if the product had an amount of sugars, sodium, fat, saturated fat and/or trans fat, that was higher than the recommended for a healthy diet.

    Finally, participants answered questions about their socio-demographic and health statuses.

    Outcomes

    The primary outcomes of the study included the contribution of the different FOPL schemes to improving the decision of participants to buy the least harmful option more often, the selection of the least harmful option more often and the correct identification of sugars, sodium and/or saturated fats found to be in excess in the products more often.

    For the first task participants had four opportunities to indicate which product, out of a set of three, they would buy, or whether they would not buy any of the products, which served the estimation of the frequency with which participants intended to buy the least harmful option or none of the options. The second task also provided four opportunities for participants to identify the least harmful product, out of a set of three, and this data served the estimation of the frequency with which participants made a correct identification. The third task allowed participants to verify four products, one at a time, and to indicate whether they contained amounts of sugars, sodium and/or saturated fats found to be higher than the recommended for a healthy diet. With this data the number of correct answers and the proportion of participants with zero, one, two, three or four correct answers for the total of four products was estimated.

    The estimated frequencies with which participants chose to buy the least harmful product or none of the products, with which participants correctly identified the least harmful product and the frequencies with which they correctly identified the nutrients in excess in the four products, were compared between the experimental and control groups to assess how systems performed according to the primary outcomes described above.

    Table 1 summarises the outcomes, the questions asked to gather the data on each outcome in the order they were asked and the responses.

    View this table:
    Table 1

    Outcomes, survey questions and responses

    Sample size

    The sample size was estimated based on calculations to detect a difference between two proportions. The most conservative criterion was used, assuming that the proportion of participants who correctly identify products with nutrient above nutritional recommendations for the control condition would be 50%. The number of participants needed to detect an absolute increase of 12% (which is smaller than what has been previously reported43 44) with a confidence level of 95% and a power level of 80% was estimated in 265 participants in each of the experimental groups (comparisons and control). The total resulting sample size was of 1200 (300×4) participants.

    Randomisation

    Shoppers were selected using quota sampling to meet a composition of age, gender and educational level within each group that resembles the one found for the population of Jamaica.45 A similar number of participants were randomly allocated to one of the three intervention groups or the control group: black OWL (n=303); MGG (n=301); TFL (n=301); and the control group which was exposed to the nutrition facts up front (n=301). The randomisation of the experimental conditions and groups was completed adopting a Williams design46 to ensure the order of categories of products and the order of products within categories was random and balanced for all groups. This randomisation produced four random and balanced sequences of categories of products and products within categories used for each of the four groups, resulting in 16 possible combinations of groups and sequences (4×4), equally balanced and order within each group.

    For the allocation of participants into each of the groups and each of the four sequences within the group, a simple randomisation technique was adopted using a Pareto randomisation procedure. Random numbers falling between 0 and 1 were generated for each of the 16 combinations of groups and sequences, then these combinations were sorted by the random numbers and used to sequentially randomise participants’ allocation to 1 of these 16 combinations of groups (experimental or control) and random sequences of categories of products and of products within categories.

    The study was single-blinded, since participants were not aware they had been assigned to an intervention or control group. Although field researchers had no prior knowledge of which intervention would be assigned to a participant, once, and only once, the exposure of the participant to an intervention or the control was initiated, researchers could tell which intervention or control participants were assigned to.

    Statistical analyses

    Descriptive statistics on the sample included proportions (for categorical variables), means (for numeric variables) and their respective CIs. To examine the contribution of FOPL schemes to improving the frequency with which consumers would correctly identify the least harmful options of products, ordered logistic regression models were used to estimate the OR of correctly identifying the least harmful products more often. The number of times participants correctly identified the least harmful option was used as the ordinal dependent variable, and FOPL schemes the independent covariates. Following a similar procedure, the number of times participants correctly identified when products contained excessive amounts of sugars, fats, saturated fats, trans fats and/or sodium was analysed as the ordinal dependent variables to verify the contribution of FOPL schemes in helping consumers to correctly identify those more often. To analyse the contribution of FOPL schemes to improving the frequency with which consumers would choose to buy the least harmful options of products, a similar model was adjusted. The number of times participants indicated they would purchase the least harmful option was the ordinal dependent variable, in this case. The ORs respective 95% CIs were also calculated, and Wald test was used to verify the significance of the contribution of FOPL schemes against the control condition and between schemes for the three outcomes listed above.

    Models were also adjusted for age, gender, education and reported NCD and related risk factors. A similar set of models were also adjusted using logistic regression with logit link function for each single category of products as subset analyses.

    All tests were two-sided and we considered p<0.05 to be statistically significant. The analyses were conducted in R language and environment for statistical computing V.4.0.1.47

    Patient and public involvement

    No patient involved.

    Results

    Table 2 provides a description of the socio-demographic characteristics of participants as well as reported NCDs conditions and related risk factors. Similar to the estimates made by the United Nations for the Jamaican population in 2021,48 the majority of the respondents were under 50 years of age, with the proportion of women and men around half, and most of them with an education level of up to secondary/vocational to Grade 11–13. In addition, one-fourth indicated they had been informed by a health professional to have overweight or obesity and the most common reported NCD condition and risk factor was hypertension, followed by diabetes, high cholesterol and heart disease. Table 2 also confirms the random allocation resulted in an even and proportional distribution of participants in the different groups.

    View this table:
    Table 2

    Socio-demographic characteristics and reported non-communicable diseases conditions and related risk factors of the sample

    Identification of the least harmful option

    The odds of participants correctly identifying the least harmful option more often more than doubled when they were exposed to the OWL (OR 2.07, 95% CI 1.54 to 2.78; p<0.001), whereas the MGG (1.18, 95% CI 0.89 to 1.57; p=0.24) and TFL (1.13, 95% CI 0.85 to 1.51; p=0.39) were inefficacious in improving such odds, compared with the control (table 3).

    View this table:
    Table 3

    Effect of different FOPL schemes on correct identification of the least harmful option, understanding of the nutritional content and intention to purchase products, in Jamaica, compared with the control condition.* Values are ORs (95% CIs)

    When analysing the results separately by product category, the OWL significantly improved the capacity of participants to identify the least harmful option for all product categories, except for flavoured milks. The MGG was only able to significantly help participants to correctly complete this task when applied to breakfast cereal extrudates, being inefficacious in all the other products, whereas the TFL was inefficacious in all products (table 3).

    Understanding about the nutritional content

    The chances of participants correctly identifying when a product contained excessive amounts of critical nutrients (sodium, sugars and saturated fats) more often were significantly the highest when they were exposed to the OWL (4.57, 95% CI 3.41 to 6.15; p<0.001), followed by MGG (2.69, 95% CI 2.01 to 3.62; p<0.001) and the TFL (2.09, 95% CI 1.57 to 2.79; p<0.001), which did not differ from each other. Similar results were found for the analysis of each product category separately (table 3). OWL outperformed the other two systems, except for the flavoured milk category, where it did not significantly differ from MGG (table 3).

    Intention to purchase

    When compared with the control, the chances of participants choosing to purchase the least harmful option or none of the options more often also doubled when they were exposed to the OWL (OR 2.03, 95% CI 1.51 to 2.72; p<0.001), and the odds for MGG (1.58, 95% CI 1.18 to 2.11; p=0.002) were lower than but not significantly different from OWL’s, whereas TFL (1.25, 95% CI 0.93 to 1.67; p=0.13) was inefficacious in improving such odds and did not differ significantly from the MGG. Similar results were found for the intention to purchase the least harmful options, whereas for the intention to not purchase any of the options, the OWL was the only one significantly increasing the odds for such decision (1.77, 95% CI 1.27 to 2.47; p<0.001) (table 3).

    The effects exerted by OWL were similar for almost all product categories when they were analysed separately, except for yoghurts. In only two instances schemes different from the OWL improved the intention to purchase the least harmful or none of the options: MGG in breakfast cereal extrudates and the TFL in crackers (table 3). For yoghurts and flavoured milks both MGG and TFL were inefficacious in increasing the intention to purchase the least harmful product or none of the products.

    Discussion

    The present study contributes to the growing body of evidence suggesting that FOP nutrition labelling improves consumer understanding of the nutritional composition of packaged foods.22 23 The three schemes including interpretational aids (OWL, MGG and TFL) increased participants’ ability to correctly identify products with high content of sugars, fat and sodium. However, results showed that the OWL significantly outperformed the MGG and the TFL in helping consumers to (a) correctly identify the least harmful option and (b) correctly identify when products contained excessive amounts of critical nutrients and (c) choose to purchase the least harmful option or none of the options in Jamaica, regardless of the populations’ age, gender, education and reported NCD and related risk factors. In addition, TFL was inefficacious in helping consumers to correctly identify the least harmful option and to choose to purchase the least harmful option or none of the options (except for crackers), and the MGG was inefficacious in helping consumers to correctly identify the least harmful option (except for breakfast cereals).

    These findings contribute to the growing body of evidence showing that warning labels outperform the TFL in improving understanding and encouraging consumers to make healthier food choices. Some of the reasons include the fact that warning labels are easier and quicker to find on the labels and to understand than TFL due to their simplicity and higher salience from the background.33 34 43 44 49 50 The TFL classifies the content of target nutrient content into low/medium/high. This information is expected to require more time and cognitive effort to interpret compared with the OWL or MGG, as reported in previous studies based on response times.43 49 Previous studies have also shown that the use of green colour found in systems such as the TFL and Nutri-Score may drive consumers to misperceive a product as healthier and undesirably raise their appetite for such products, which may explain their lower effect on understanding of nutritional information and purchase intention.49 51–61 In addition, the use of red colour, also found in such systems, although intended to communicate a higher harmfulness level, may trigger an opposite effect in some products. Lemos et al have shown, using objective measures of brain activities, that the red colour triggers in consumers a positive emotional motivation towards sweet ultra-processed products.62 Reviews and meta-analysis have also documented that warning labels are more consistently successful and perform best in reducing purchase intention of products,23 63 and a unique real-life interrupted time series study conducted in Chile has shown that warnings contributed to actual reductions in the purchase of products high in calories, sodium, saturated fats and sugars.64

    The research also compared two warning labels with different graphical design: octagonal warnings and magnifier glass. Results showed that octagonal warnings were more effective at improving understanding of nutritional information and at encouraging healthier food choices, in agreement with results reported by Deliza et al and Goodman et al.33 34 The difference can be attributed to the learnt association between octagons and danger in the context of traffic signs.65 In addition, the literature on warnings for other unhealthy products have demonstrated that there is a dose-response between the proportion of the package surface occupied by warnings and the response of consumers, that is, the higher is the proportion of the warnings labels on the label the better is the response of consumers to avoid the product.66 OWL provides an implicit dose-response, as the surface occupied by this FOP labelling scheme increases as the number of nutrients in excess increase. Conversely, single icon schemes, such as the MGG, do not count with such mechanism since they occupy the same proportion of the label surface area regardless of the number of nutrients in excess found in the product. Prates et al have tested such hypothesis and found that in products with one single nutrient in excess, OWL and MGG performed similarly, however when products had two nutrients in excess, the OWL outperformed the MGG in reducing the healthfulness perceptions (all categories of products) and the purchase intention (cereal bars and cookies) of these products.39 The superiority of octagonal signs supports the graphical design of warning labels implemented in most countries in the Americas, and can help informing policymaking in the Caribbean, as OWL outperformed other systems suggested by industry representatives along the regulatory process, such as the MGG and TFL.

    This is the first study of its kind to be completed in the Caribbean region and one of the few that have compared different warning labelling schemes (multiple-icon vs single-icon scheme), thus helping to fill the local, regional and global knowledge gaps on the matter. One of its major strengths is the robust between-subject design which allows the results to be attributed to FOPL schemes, and avoids differential carry-over effects that are more likely to happen in within-subject studies. The exposure of participants to 2D mock-up products with different FOPL schemes was standardised to match real products’ label sizes and sizes and proportions of FOPL, approaching real-life conditions, also strengthened its external validity. However, it should be noted that the study was conducted using fictitious brands, products were not associated with price information and participants did not purchase the products, which has strengthened the attribution of the effects to the FOPL schemes, but conversely, limited the analysis of the relative importance of these other factors. Additional studies should be conducted to assess the effect of octagonal warnings on actual purchases under real-life conditions.

    Conclusions and policy implications

    The findings of this study indicate that the OWL included in the CROSQ proposal32 to be adopted by CARICOM Member States is the most effective option in meeting the regulatory objective of helping the population in Jamaica to correctly identify the least harmful option and the presence of excessive amounts of critical nutrients, and to choose to purchase the least harmful more often, regardless of the populations’ age, gender, education and reported NCD and related risk factors. Once the system is implemented in Caribbean countries, future research to evaluate the impact on actual purchases of products and on dietary changes will be needed to keep track of the changes expected to be exerted by OWL in the short-terms and mid-terms.

    Data availability statement

    No data are available.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study was approved by the Advisory Panel on Ethics and Medico-Legal Affairs in the Ministry of Health of Jamaica (Reference number: SRD/ETH/20), the Research Ethics Committee of the University of Technology of Jamaica (Reference number: 2019/10/COHS/263) and the Pan-American Health Organization Ethics Review Committee (Reference number: PAHOERC.0105.01). In line with the International Committee of Medical Journal Editors’ recommendation, the study was not required to be registered in a public trial registry since it does not measure health outcomes.

    Acknowledgments

    We would like to thank Carla G Spinillo and Carlos Felipe U Rojas, (Federal University of Paraná, Brazil) for their support in designing the mock-up products used in this study, and Tamu Davidson (formerly with the Jamaica Ministry of Health & Wellness) for her contribution to the conceptualisation of the project and for providing the helpful contextual information on the Jamaican setting.

    References

    1. ↵
      1. World Health Organization
      . Global health observatory data repository. 2022. Available: http://www.who.int/gho
    2. ↵
      1. Pan American Health Organization
      . Noncommunicable diseases in the region of the Americas: facts and figures. Washington, DC: Pan American Health Organization, 2019.
    3. ↵
      1. Pan American Health
      . ENLACE: data portal on noncommunicable diseases, mental health, and external causes. n.d. Available: https://www.paho.org/en/enlace
    4. ↵
      1. Jamaica Ministry of Health
      . Jamaica health and lifestyle survey III (2016-2017). Ministry of Health & Wellness, 2018.
    5. ↵
      1. GBD 2017 Diet Collaborators
      . Health effects of dietary risks in 195 countries, 1990-2017: a systematic analysis for the global burden of disease study 2017. Lancet 2019;393:1958–72. doi:10.1016/S0140-6736(19)30041-8
      OpenUrlPubMed
    6. ↵
      1. Swinburn BA,
      2. Kraak VI,
      3. Allender S, et al
      . The global syndemic of obesity, undernutrition, and climate change: the Lancet commission report. Lancet 2019;393:791–846. doi:10.1016/S0140-6736(18)32822-8
      OpenUrlCrossRefPubMed
    7. ↵
      1. Monteiro CA,
      2. Cannon G,
      3. Lawrence M, et al
      . Ultra-processed foods, diet quality, and health using the NOVA classification system. Rome: FAO, 2019.
    8. ↵
      1. Pan American Health Organization
      . Ultra-processed food and drink products in Latin America: trends, impact on obesity, policy implications. Washington, DC: Pan American Health Organization, 2015.
    9. ↵
      1. Pan American Health Organization
      . Ultra-processed food and drink products in Latin America: sales, sources, nutrient profiles, and policy implications. Washington, DC: Pan American Health Organization, 2019.
    10. ↵
      1. Stanton RA
      . Food retailers and obesity. Curr Obes Rep 2015;4:54–9. doi:10.1007/s13679-014-0137-4
      OpenUrl
    11. ↵
      1. Story M,
      2. Kaphingst KM,
      3. Robinson-O’Brien R, et al
      . Creating healthy food and eating environments: policy and environmental approaches. Annu Rev Public Health 2008;29:253–72. doi:10.1146/annurev.publhealth.29.020907.090926
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Mazzocchi M,
      2. Cagnone S,
      3. Bech-Larsen T, et al
      . What is the public appetite for healthy eating policies? Evidence from a cross-European survey. Health Econ Policy Law 2015;10:267–92. doi:10.1017/S1744133114000346
      OpenUrl
    13. ↵
      1. Cowburn G,
      2. Stockley L
      . Consumer understanding and use of nutrition labelling: a systematic review. Public Health Nutr 2005;8:21–8. doi:10.1079/phn2005666
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Cheftel JC
      . Food and nutrition labelling in the European Union. Food Chemistry 2005;93:531–50. doi:10.1016/j.foodchem.2004.11.041
      OpenUrl
    15. ↵
      1. Feunekes GIJ,
      2. Gortemaker IA,
      3. Willems AA, et al
      . Front-of-pack nutrition labelling: testing effectiveness of different nutrition labelling formats front-of-pack in four European countries. Appetite 2008;50:57–70. doi:10.1016/j.appet.2007.05.009
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Grunert KG,
      2. Wills JM
      . A review of European research on consumer response to nutrition information on food labels. J Public Health 2007;15:385–99. doi:10.1007/s10389-007-0101-9
      OpenUrlCrossRef
    17. ↵
      1. Sharf M,
      2. Sela R,
      3. Zentner G, et al
      . Figuring out food labels. young adults’ understanding of nutritional information presented on food labels is inadequate. Appetite 2012;58:531–4. doi:10.1016/j.appet.2011.12.010
      OpenUrlCrossRefPubMed
    18. ↵
      1. Machín L,
      2. Curutchet MR,
      3. Gugliucci V, et al
      . The habitual nature of food purchases at the supermarket: implications for policy making. Appetite 2020;155:S0195-6663(20)30548-1. doi:10.1016/j.appet.2020.104844
    19. ↵
      1. Hawley KL,
      2. Roberto CA,
      3. Bragg MA, et al
      . The science on front-of-package food labels. Public Health Nutr 2013;16:430–9. doi:10.1017/S1368980012000754
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Kleef EV,
      2. Dagevos H
      . The growing role of front-of-pack nutrition profile labeling: a consumer perspective on key issues and controversies. Crit Rev Food Sci Nutr 2015;55:291–303. doi:10.1080/10408398.2011.653018
      OpenUrlPubMed
    21. ↵
      1. World Health Organization
      . Tackling NCDs: “best buys” and other recommended interventions for the prevention and control of noncommunicable diseases. Geneva: World Health Organization, 2017.
    22. ↵
      1. Roberto CA,
      2. Ng SW,
      3. Ganderats-Fuentes M, et al
      . The influence of front-of-package nutrition labeling on consumer behavior and product reformulation. Annu Rev Nutr 2021;41:529–50. doi:10.1146/annurev-nutr-111120-094932
      OpenUrlCrossRefPubMed
    23. ↵
      1. Temple NJ
      . Front-of-package food labels: a narrative review. Appetite 2020;144:S0195-6663(19)30408-8. doi:10.1016/j.appet.2019.104485
    24. ↵
      1. Crockett RA,
      2. King SE,
      3. Marteau TM, et al
      . Nutritional labelling for healthier food or non-alcoholic drink purchasing and consumption. Cochrane Database Syst Rev 2018;2:CD009315. doi:10.1002/14651858.CD009315.pub2
    25. ↵
      1. Ares G,
      2. Antúnez L,
      3. Curutchet MR, et al
      . Immediate effects of the implementation of nutritional warnings in Uruguay: awareness, self-reported use and increased understanding. Public Health Nutr 2021;24:364–75. doi:10.1017/S1368980020002517
      OpenUrl
    26. ↵
      1. Egnell M,
      2. Crosetto P,
      3. d’Almeida T, et al
      . Modelling the impact of different front-of-package nutrition labels on mortality from non-communicable chronic disease. Int J Behav Nutr Phys Act 2019;16:56. doi:10.1186/s12966-019-0817-2
    27. ↵
      1. Labonté M-E,
      2. Emrich TE,
      3. Scarborough P, et al
      . Traffic light labelling could prevent mortality from noncommunicable diseases in Canada: a scenario modelling study. PLoS One 2019;14:e0226975. doi:10.1371/journal.pone.0226975
    28. ↵
      1. Basto-Abreu A,
      2. Torres-Alvarez R,
      3. Reyes-Sánchez F, et al
      . Predicting obesity reduction after implementing warning labels in Mexico: a modeling study. PLoS Med 2020;17:e1003221. doi:10.1371/journal.pmed.1003221
    29. ↵
      1. Hodgkins C,
      2. Barnett J,
      3. Wasowicz-Kirylo G, et al
      . Understanding how consumers categorise nutritional labels: a consumer derived typology for front-of-pack nutrition labelling. Appetite 2012;59:806–17. doi:10.1016/j.appet.2012.08.014
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Jones A,
      2. Neal B,
      3. Reeve B, et al
      . Front-of-pack nutrition labelling to promote healthier diets: current practice and opportunities to strengthen regulation worldwide. BMJ Glob Health 2019;4:e001882. doi:10.1136/bmjgh-2019-001882
    31. ↵
      1. Jamaica. Bureau of Standards
      . JS CRS 5 2010 Jamaican standard specification for labelling of pre-packaged foods. Kingston: Bureau of Standards, 2017.
    32. ↵
      1. CARICOM Regional Standard
      . Labelling of foods – pre-packaged foods – specification (revision). DCRS 5: 202X. Specification for voting. Warrens: CROSQ, 2020.
    33. ↵
      1. Deliza R,
      2. de Alcantara M,
      3. Pereira R, et al
      . How do different warning signs compare with the guideline daily amount and traffic-light system? Food Quality and Preference 2020;80:103821. doi:10.1016/j.foodqual.2019.103821
    34. ↵
      1. Goodman S,
      2. Vanderlee L,
      3. Acton R, et al
      . The impact of front-of-package label design on consumer understanding of nutrient amounts. Nutrients 2018;10:1624. doi:10.3390/nu10111624
    35. ↵
      1. Khandpur N,
      2. Mais LA,
      3. de Morais Sato P, et al
      . Choosing a front-of-package warning label for Brazil: a randomized, controlled comparison of three different label designs. Food Res Int 2019;121:854–61. doi:10.1016/j.foodres.2019.01.008
      OpenUrl
    36. ↵
      1. Bandeira LM,
      2. Pedroso J,
      3. Toral N, et al
      . Performance and perception on front-of-package nutritional labeling models in Brazil. Rev Saude Publica 2021;55:19. doi:10.11606/s1518-8787.2021055002395
    37. ↵
      1. Silva ARCS,
      2. Ni Mhurchu C,
      3. Anastácio LR
      . Comparison of two front-of-pack nutrition labels for Brazilian consumers using a smartphone APP in a real-world grocery store: a pilot randomized controlled study. Front Nutr 2022;9:898021. doi:10.3389/fnut.2022.898021
    38. ↵
      1. Khandpur N,
      2. Amaral Mais L,
      3. Bortoletto Martins AP
      . A comparative assessment of two different front-of-package nutrition label designs: a randomized experiment in Brazil. PLoS One 2022;17:e0265990. doi:10.1371/journal.pone.0265990
    39. ↵
      1. Prates SMS,
      2. Reis IA,
      3. Rojas CFU, et al
      . Influence of nutrition claims on different models of front-of-package nutritional labeling in supposedly healthy foods: impact on the understanding of nutritional information, healthfulness perception, and purchase intention of Brazilian consumers. Front Nutr 2022;9:921065. doi:10.3389/fnut.2022.921065
    40. ↵
      1. Department of Health
      . Food standards agency. guide to creating a front of pack (fop) nutrition label for pre-packed products sold through retail outlets. London: UK-FSA, 2016.
    41. ↵
      1. Health Canada
      . Consultation on proposed front-of-package labelling. n.d. Available: www.canada.ca/en/health-canada/programs/consultation-front-of-package-nutrition-labelling-cgi.html
    42. ↵
      1. Pan American Health Organization
      . Pan American Health Organization nutrient profile model. Washington, DC: Pan American Health Organization, 2016.
    43. ↵
      1. Arrúa A,
      2. Machín L,
      3. Curutchet MR, et al
      . Warnings as a directive front-of-pack nutrition labelling scheme: comparison with the Guideline daily amount and traffic-light systems. Public Health Nutr 2017;20:2308–17. doi:10.1017/S1368980017000866
      OpenUrl
    44. ↵
      1. Acton RB,
      2. Rynard VL,
      3. Adams J, et al
      . Awareness, use and understanding of nutrition labels among adults from five countries: findings from the 2018-2020 international food policy study. Appetite 2023;180:S0195-6663(22)00402-0. doi:10.1016/j.appet.2022.106311
    45. ↵
      1. Jamaica. Ministry of Labour and Social Security. Planning, Research and Monitoring Unit
      . National employment report Jamaica. Kingston: Ministry of Labour and Social Security, 2006.
    46. ↵
      1. Wang B-S,
      2. Wang X-J,
      3. Gong L-K
      . The construction of a williams design and randomization in cross-over clinical trials using sas. J Stat Soft 2009;29:1–10. doi:10.18637/jss.v029.c01
      OpenUrl
    47. ↵
      1. R Core Team
      . R: A language and environment for statistical computing. vienna, R foundation for statistical computing. 2021. Available: www.R-project.org
    48. ↵
      1. United Nations, Department of Economic and Social Affairs, Population Division
      . World population prospects. 2019. Available: https://population.un.org/wpp
    49. ↵
      1. Cabrera M,
      2. Machín L,
      3. Arrúa A, et al
      . Nutrition warnings as front-of-pack labels: influence of design features on healthfulness perception and attentional capture. Public Health Nutr 2017;20:3360–71. doi:10.1017/S136898001700249X
      OpenUrl
    50. ↵
      1. Khandpur N,
      2. de Morais Sato P,
      3. Mais LA, et al
      . Are front-of-package warning labels more effective at communicating nutrition information than traffic-light labels? A randomized controlled experiment in a Brazilian sample. Nutrients 2018;10:688. doi:10.3390/nu10060688
    51. ↵
      1. Franco-Arellano B,
      2. Vanderlee L,
      3. Ahmed M, et al
      . Influence of front-of-pack labelling and regulated nutrition claims on consumers’ perceptions of product healthfulness and purchase intentions: a randomized controlled trial. Appetite 2020;149:S0195-6663(19)30909-2. doi:10.1016/j.appet.2020.104629
    52. ↵
      1. Schuldt JP
      . Does green mean healthy? Nutrition label color affects perceptions of healthfulness. Health Commun 2013;28:814–21. doi:10.1080/10410236.2012.725270
      OpenUrlCrossRefPubMed
    53. ↵
      1. Huang L,
      2. Lu J
      . The impact of package color and the nutrition content labels on the perception of food healthiness and purchase intention. J Food Product Market 2016;22:191–218. doi:10.1080/10454446.2014.1000434
      OpenUrl
    54. ↵
      1. Nyilasy G,
      2. Lei J,
      3. Nagpal A, et al
      . Colour correct: the interactive effects of food label nutrition colouring schemes and food category healthiness on health perceptions. Public Health Nutr 2016;19:2122–7. doi:10.1017/S1368980016000483
      OpenUrl
    55. ↵
      1. Talati Z,
      2. Pettigrew S,
      3. Dixon H, et al
      . Do health claims and front-of-pack labels lead to a positivity bias in unhealthy foods? Nutrients 2016;8:787. doi:10.3390/nu8120787
    56. ↵
      1. Machín L,
      2. Aschemann-Witzel J,
      3. Curutchet MR, et al
      . Traffic light system can increase healthfulness perception: implications for policy making. J Nutr Educ Behav 2018;50:668–74. doi:10.1016/j.jneb.2018.03.005
      OpenUrl
    57. ↵
      1. Singh S
      . Impact of color on marketing. Management Decision 2006;44:783–9. doi:10.1108/00251740610673332
      OpenUrl
    58. ↵
      1. Spence C
      . On the psychological impact of food colour. Flavour 2015;4:21. doi:10.1186/s13411-015-0031-3
    59. ↵
      1. Garber LL,
      2. Hyatt EM,
      3. Starr RG
      . The effects of food color on perceived flavor. J Market Theory Pract 2000;8:59–72. doi:10.1080/10696679.2000.11501880
      OpenUrl
    60. ↵
      1. Hock K,
      2. Acton RB,
      3. Jáuregui A, et al
      . Experimental study of front-of-package nutrition labels’ efficacy on perceived healthfulness of sugar-sweetened beverages among youth in six countries. Prev Med Rep 2021;24:101577. doi:10.1016/j.pmedr.2021.101577
    61. ↵
      1. Jáuregui A,
      2. White CM,
      3. Vanderlee L, et al
      . Impact of front-of-pack labels on the perceived healthfulness of a sweetened fruit drink: a randomised experiment in five countries. Public Health Nutr 2022;25:1094–104. doi:10.1017/S1368980021004535
      OpenUrl
    62. ↵
      1. Lemos TC,
      2. Almo A,
      3. Campagnoli RR, et al
      . A red code triggers an unintended approach motivation toward sweet ultra-processed foods: possible implications for front-of-pack labels. Food Quality Preference 2020;79:103784. doi:10.1016/j.foodqual.2019.103784
    63. ↵
      1. Croker H,
      2. Packer J,
      3. Russell SJ, et al
      . Front of pack nutritional labelling schemes: a systematic review and meta-analysis of recent evidence relating to objectively measured consumption and purchasing. J Hum Nutr Diet 2020;33:518–37. doi:10.1111/jhn.12758
      OpenUrlCrossRefPubMed
    64. ↵
      1. Taillie LS,
      2. Bercholz M,
      3. Popkin B, et al
      . Changes in food purchases after the Chilean policies on food labelling, marketing, and sales in schools: a before and after study. Lancet Planet Health 2021;5:e526–33. doi:10.1016/S2542-5196(21)00172-8
      OpenUrlCrossRefPubMed
    65. ↵
      1. Mazis MB,
      2. Morris LA
      . Channel. In: Wogalter MS, Dejoy DM, Laughery KR, eds. Warnings and risk communication. London: Taylor & Francis, 2005: 13–21.
    66. ↵
      1. Hammond D
      . Health warning messages on tobacco products: a review. Tob Control 2011;20:327–37. doi:10.1136/tc.2010.037630
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors VW-B, FSG, SE, GA, AM, DC and DF provided substantial contributions to the conception and design of the study. DF led the sample design and sampling procedures. VW-B and SE led the data acquisition. FSG performed the statistical analyses and provided preliminary interpretation of findings. FSG drafted the first version of the paper and VW-B, FSG, SE, GA, AM, DC and DF revised it critically for important intellectual content. VW-B, FSG, SE, GA, AM, DC and DF approved the final version to be published. VW-B, FSG, SE, GA, AM, DC and DF agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. VW-B and FSG are the guarantors. FSG and AM are staff members of the Pan American Health Organization. Authors hold sole responsibility for the views expressed in the manuscript, which may not necessarily reflect the opinion or policy of the Pan American Health Organization.

    • Funding This work was supported by the Global Health Advocacy Incubator and funded by Bloomberg Philanthropies (grant number not available). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

    • Competing interests All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declared no financial relationships with any organisations that might have an interest in the submitted work and no other relationships or activities that could appear to have influenced the submitted work.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.