Childhood obesity and poor dietary habits are pressing matters in Mexico and elsewhere. The prevalence of overweight and obesity in school-age children in Mexico is 35·6 %, one of the highest rates globally and in Mexican history(Reference Shamah-Levy, Vielma-Orozco and Heredia-Hernández1). One of the most important factors associated with this public health problem is the increasing shift in consumption of energy-dense, ultra-processed foods high in sugar, salt and fat(Reference Marrón-Ponce, Tolentino-Mayo and Hernández-F2). In Mexico, 26 % of the calories in the diet of adults come from ultra-processed foods; meanwhile, the caloric share of these products among children reaches 34–38 % of the total intake(Reference Marrón-Ponce, Sánchez-Pimienta and Da Costa Louzada3). Soft drinks significantly contribute to energy intake in Mexico, with an average consumption of around 150 l/person annually(4).
Front-of-package labelling is a public health strategy designed to inform consumers and promote healthier choices(Reference Roberto, Ng and Ganderats Fuentes5). In October 2020, a new front-of-package warning labelling system was implemented in Mexico(Reference White and Barquera6). Warning labels were initially introduced in Chile in 2016(Reference Kanter, Vanderlee and Vandevijvere7). They informed the Mexican regulations, which consist of octagon-shaped black warnings to draw consumers’ attention to ‘excess(es) in’ five nutrients of concern in packaged foods: calories, added sugars, saturated fat, trans fat and Na(Reference White and Barquera6). The Mexican regulation required additional elements compared with Chile, including precautionary disclaimers displayed as black rectangles on the front of packs to discourage child consumption of products with caffeine (‘Contains caffeine, avoid in children’) or non-sugary sweeteners (‘Contains sweeteners, not recommended in children’) (Fig. 1). A similar labelling system promoting disclaimers for caffeine and sweeteners was also implemented in Argentina in 2022; however, the results of this change are not yet known(8).
Fig. 1 Mexican warning label system for packaged foods and beverages approved in 2020
Non-sugary sweeteners (hereafter, sweeteners) are increasingly prevalent in the food supply, and their consumption has increased in the past decade(Reference Russell, Grimes and Baker9). However, sweeteners have significant public health consequences. These sugar substitutes promote several physiological mechanisms contributing to adverse metabolic effects, including altered sweetness perceptions that lead to greater caloric intake(Reference Shum and Georgia10). Furthermore, consumption of non-caloric sweeteners increases gut microbiota dysbiosis and disrupts glucose homoeostasis, which may lead to increased insulin resistance, hyperinsulinaemic response and increased adiposity(Reference Shum and Georgia10–Reference Sylvetsky, Jin and Mathieu13). Recent evidence also suggests a possible link between some types of sweeteners, particularly aspartame and acesulfame-K, and cancer(Reference Debras, Chazelas and Srour14). International health organisations and national health and nutrition associations recommend that parents be advised about the risk of sweeteners in children’s diets and discourage their use(Reference Baker-Smith, De Ferranti and Cochran15). Despite this, it was found that sweeteners were used to replace sugars after warning labels were introduced in Chile to circumvent the application of sugar warnings on products making the sweeteners prevalent in products marketed to children and increasing the risk of exceeding the acceptable daily intake of some sweeteners in paediatric populations(Reference Sambra, López-Arana and Cáceres16). Therefore, sweeteners were included in the Mexican warning label system to prevent manufacturers from reformulating their products in ways that circumvent the sugar warning as occurred in Chile.
Caffeine consumption in children has also been associated with multiple adverse health effects, including sleep disruption, poor decision-making, increased risk-taking, impaired growth and development, high blood pressure, decreased cerebral blood flow, cardiovascular complications, addiction, stress, anxiety and depression(Reference Rodak, Kokot and Kratz17,Reference Soós, Gyebrovszki and Tóth18) . There is no safe dose of caffeine for children(Reference Soós, Gyebrovszki and Tóth18). Furthermore, caffeine in already highly palatable sugary drinks, such as soda drinks, increases their hedonic properties(Reference Yeomans, Durlach and Tinley19), which may promote consumption. The Mexican population highly consumes soft drinks(4), including infants, toddlers and children(Reference González-Castell, González de Cosío and Rodríguez-Ramírez20). These drinks are the primary source of added caffeine intake in children and adolescents(Reference Reyes and Cornelis21), justifying the need to prevent youth consumption of soft drinks and associated detrimental health effects. The effect of caffeine on children’s health justified the inclusion of this disclaimer in the Mexican warning label system.
To date, several studies have investigated the effects of warning labels on the perceived healthfulness of foods and their potential impact on purchasing decisions(Reference Taillie, Hall and Popkin22). However, to our knowledge, no evidence exists on their potential impact due to the novelty of caffeine and sweetener disclaimers. The present study examined the awareness and use of sweetener disclaimers among Mexican youth and adults 1 month after the new warning labels were implemented. We also investigated the potential impact of sweetener and caffeine disclaimers on the perceived healthfulness of industrialised beverages and whether socio-demographic characteristics modified this effect.
Methods
Study design
This article presents results from two online between-subjects experiments embedded in the Mexican arm of the 2020 International Food Policy Study (IFPS). The IFPS is a repeated cross-sectional online survey of youth (10–17 years) and adults (≥18 years) from six and five countries, respectively, including Mexico, to assess the impact of food policies at the national level. Data were collected in November 2020, 1 month after the full implementation of the Mexican warning labels. Using programming in the online survey software, Alchemer, adult participants were randomly assigned to one of a series of controlled experiments, including the two described in this article. Experiment 1 examined the potential impact of a sweetener disclaimer on a sweetened fruit drink on the perceived healthfulness of that type of drink for a child. Experiment 2 examined the potential impact of a caffeine disclaimer of a caffeinated lemon-flavoured soda on the perceived healthfulness of that type of drink for a child. All participants in the youth survey were also asked to complete Experiment 1. Youth were not asked to complete Experiment 2 to accommodate shorter attention spans.
Recruitment
Participants were recruited through the Nielsen Consumer Insights Global Panel and their partners’ panels using probability and non-probability sampling methods. Data were collected via self-completed web-based surveys conducted in November–December 2020 with youth (10–17 years) and adults (≥18 years). Email invitations with unique survey access links were sent to a random sample of panellists after targeting for demographics; panellists known to be ineligible were not invited. For the adult survey, random samples were drawn from online panels, with quotas for age and sex proportional to the general population. Respondents provided consent before completing the survey. For the youth survey, invitations were sent to adult panellists and those who confirmed they had a child aged 10–17 years in their household were informed about the study and provided consent for their child’s participation. Only one child per household was invited. Children were subsequently screened to confirm eligibility, given study information and provided assent before questionnaire commencement. The child’s parent/guardian and adult participants received compensation according to their panel’s usual incentive structure (e.g. points-based or monetary rewards, chances to win prizes). The survey was conducted in Spanish. The adult and youth studies were reviewed and approved by the University of Waterloo Research Ethics Committee (ORE# 30829 and ORE #41477, respectively) and the National Institute of Public Health of Mexico (CI: 1641 and 1122). A full description of the study methods can be found in the International Food Policy Study Technical Reports(Reference Hammond, White and Rynard23,Reference Hammond, White and Rynard24) .
For Mexico, a total of 135 508 (adults: 81 234; youth: 54 274) email invitations were sent to a random sample of panellists. The American Association for Public Opinion Research cooperation rate #2 was 86·1 % for youth and 49·8 % for adults, calculated as the percentage of participants who completed the survey and were retained in the sample (adults: 4284; youth: 1823) out of that eligible who accessed the survey link (adults: 8708; youth: 2118)(25).
Measures
Disclaimer awareness and use
Youth and adult participants were asked: ‘How often have you seen this type of food label on packages or in stores?’, along with an image of the sweetener disclaimer. Response options were (1) Never, (2) Rarely, (3) Sometimes, (4) Often, (5) All the time, (6) Don’t know and (7) Refuse to answer. All participants choosing options 2–5 were further asked: ‘How often do you use this type of food label when deciding to buy a food product?’, with the same response options as the previous question.
Experiments embedded in the survey
To test the effect of disclaimers, we used beverage images of real drink packages containing sweeteners or caffeine, correspondingly, and digitally altered to display the disclaimers (as applicable). The caffeine disclaimer and the sweetener disclaimer were displayed at the top front of the container, as stipulated in the Mexican regulation (Fig. 2). The regulation indicates that these disclaimers should be included when the food or beverage contains added caffeine or sweeteners, as tested products. We only tested the effect of these disclaimers because we did not want to test the complete warning label system.
Fig. 2 Images with front-of-package labels displayed during experimental tasks. Experiment 1 = sweetener experiment. Experiment 2 = caffeine experiment. Note: each participant was only shown one image, corresponding to their assigned condition (control or disclaimer)
Participants’ allocation and intervention
Experiment 1
Participants were randomised to view on-screen one of two images (4·5 cm × 9 cm) of a sweetened fruit drink (agüitas) with differing labelling: no disclaimer (control) or the disclaimer ‘CONTIENE EDULCORANTES, NO RECOMENDABLE EN NIÑOS’ label (Fig. 2, Panel A).
Experiment 2
Participants assigned to the caffeine experiment were randomised to view on-screen one of two images of a lemon-flavoured soda with differing labelling: no disclaimer (control) or ‘CONTIENE CAFEÍNA – EVITAR EN NIÑOS’ label (Fig. 2, Panel B). The image used corresponded to a soda that is not common to people in Mexico.
Outcomes
In both experiments, adult participants were asked, ‘In your opinion, how unhealthy or healthy is this type of drink for a child?’, along with an image of the control drink or a sweetener or caffeine disclaimer. The adult survey contained nine response options: (1) Very unhealthy, (2) Unhealthy, (3) A little unhealthy, (4) Neither unhealthy nor healthy, (5) A little healthy, (6) Healthy, (7) Very healthy, (8) Don’t know and (9) Refuse to answer.
Meanwhile, for Experiment 1 the youth were asked, ‘Is this type of drink unhealthy or healthy?’, along with an image of the control drink or a sweetener disclaimer. The youth survey included seven response options: (1) Very unhealthy, (2) A little unhealthy, (3) A little healthy, (4) Healthy, (5) Very healthy, (6) Don’t know and (7) Refuse to answer. Responses were analysed as continuous data, excluding those who answered, ‘Don’t know’ or ‘Refuse to answer’.
Covariates
Demographic information included sex (male; female), age (youth: 10–13, 14–17 years; adults: 18–29, 30–39, 30–39, 40–49, 50–59 or 60 years and over), ethnicity (Indigenous v. Non-indigenous), frequency of using a nutrition facts table (Never/Rarely, Sometimes, Often/All the time) and self-reported weight (in kg) and height (in cm). Participants were also queried about their nutrition knowledge because it is one of the first mechanisms (cognitive) by which front-of-pack labels may function to influence purchasing behaviour(Reference Roberto, Ng and Ganderats Fuentes5). Adults were asked, ‘How would you rate your nutrition knowledge?’ Response options for adult participants were recoded as ‘Not at all or A little knowledgeable,’ ‘Somewhat knowledgeable’ and ‘Very or Extremely knowledgeable’. Youth were asked, ‘How much do you know about healthy eating and nutrition’ and asked to answer on a scale ranging from 0 (Nothing) to 10 (A lot). Perceived income adequacy was measured with the following question for adults and child parents/guardians: ‘Thinking about your total monthly income, how difficult or easy is it for you to make ends meet?’ (Response options recoded as ‘Very difficult/Difficult’; ‘Neither easy nor difficult’; ‘Easy/Very easy’)(Reference Litwin and Sapir26). Adult participants were asked about their education level, which was recoded as ‘Low’ (i.e. completed secondary school or less), ‘Medium’ (i.e. some post-secondary qualifications) or ‘High’ (i.e. university degree or higher) according to country-specific criteria related to the highest level of formal education attained. Their household shopping role was measured with the following question: ‘Do you do most of the food shopping in your household?’ (’Most of the shopping’; ‘Share equally’; ‘Share some of the shopping’; ‘None’). Adults were also asked, ‘Do you have any children (including step-children or adopted children)?’ (Response options were ‘Yes’, ‘No’, ‘Don’t know’, ‘Refuse to answer’). Finally, youth and adult participants were asked about their agreement with the following statement: ‘Boys and girls should not consume foods or drinks with sweeteners (sugar substitutes)’ with response options recorded as: (1) Strongly agree, (2) Agree, (3) Neither agree nor disagree, (4) Disagree, (5) Strongly disagree, (6) Don’t know and (7) Refuse to answer. Using self-reported weight and height, BMI categories were estimated according to WHO criteria for adults (kg/m2): underweight (<18·5 kg/m2) normal weight (18·5–24·9 kg/m2), overweight (25·0–29·9 kg/m2) and obesity (≥30 kg/m2) and using the WHO macro for youth (BMI for age Z-scores)(27). Further details on the IFPS study methods are available elsewhere(Reference Hammond, White and Rynard23,Reference Hammond, White and Rynard24) .
Statistical analysis
The IFPS study sample size was powered to examine differences in nutritional outcomes between countries over time and not for each task within the survey. Post-hoc analyses indicated that with a sample size of 500 participants in each experimental condition (control v. caffeine or sweetener experiments) and a SD of 1·5, this study had an estimated 84 % power to detect a 0·25 mean difference on the 7-point Likert scale. Data analysis was performed using STATA 17.
Of the 4284 participants who completed the adult survey, 1085 were assigned to Experiment 1 (caffeine experiment) and 1098 to Experiment 2 (sweetener experiment); the remaining 2101 were assigned to other experiments not discussed in this article. Between 1·7 % and 5·1 % of adult participants were excluded due to missing data in the outcome (Experiment 1: n 18; Experiment 2: n 56) or a technical glitch in Experiment 1 that resulted in them being shown both conditions (control and legend) (n 1), leaving 2108 adult participants for main analyses (Experiment 1: 1067; Experiment 2: 1041). Of the 1823 participants who completed the youth survey, 1·8 % were excluded due to missing data in Experiment 1 (n 29) or a technical glitch that resulted in them seeing both conditions (n 4); thus, the final sample size for youth was 1790. Demographic characteristics, awareness and use of the disclaimers and perceived healthfulness of sweeteners were described using proportions and means, correspondingly, along with 95 % CI.
Differences between experimental groups for the covariates were tested using Chi-squared tests for categorical variables and t-tests for continuous variables. For Experiment 1, differences between experimental conditions were observed in the belief that children should not consume foods or drinks with sweeteners among adults and in the percentage of participants self-identifying as indigenous among youth. For Experiment 2 (only among adults), differences between experimental conditions were observed in the distribution of self-reported nutrition knowledge categories.
Differences in the mean perceived healthfulness between experimental conditions were tested using t-tests for independent means among the full sample. To account for the unbalanced covariates between experimental conditions, we also used linear regression models introducing the beverage perceived healthfulness as the outcome and the experimental condition as the exposure and were adjusted for the unbalanced variable. These models were run among those with complete information in the outcome and the unbalanced variable (Experiment 1: adults: 1066, youth: 1749; Experiment 2: 1038).
We tested for possible interactions between experimental conditions and demographic characteristics (i.e. sex, age group, income adequacy, ethnicity, education (only for adults), nutrition knowledge, food shopping in the household (only for adults), frequency of using the nutrition facts table, BMI category and the belief that children should not consume foods or drinks with sweeteners). For this purpose, linear regression modelling was used to evaluate the effect of the labels on perceived healthfulness, introducing multiplicative interactions between each demographic variable and experimental condition and adjusting for the unbalanced variable, accordingly. Only significant interactions (P < 0·05) were retained. Associations within the demographic variables were presented in cases where demographic × label interactions were significant.
Results
A total of 3898 (2108 adults and 1790 youth) participants were analysed. Table 1 shows the characteristics of the study participants. In total, 80·7 % of adults and 67·3 % of youth agreed that boys and girls should not consume foods or drinks with sweeteners.
Table 1 Demographic characteristics of Mexican participants in the 2020 International Food Policy Study
* Smaller sample sizes are presented due to missing information.
Randomisation successfully created comparable groups for most measures across experimental groups (see online Supplemental Table 1). In total, 1067 adults (control: 531; sweetener disclaimer: 536) and 1790 youth (control: 917; sweetener disclaimer: 873) completed the sweetener experiment, and 1041 adult participants completed the caffeine experiment (control: 505; caffeine disclaimer: 536).
Disclaimer awareness and use
Over 80 % of youth and adults reported seeing disclaimer-type labels on packages or in the stores at least rarely (Fig. 3), and, among those, around 30 % and 20 % reported that they used the disclaimers sometimes or frequently, respectively, when deciding what to eat or buy. Regarding the use of disclaimers to make buy choices, 67·0 % of adults and 60·4 % of youth participants mentioned using them sometimes, frequently or all the time (Fig. 3).
Fig. 3 Awareness and use of disclaimer-type labels among Mexican adults and youth
Experiment 1. Effect of the sweetener disclaimer on the perceived healthfulness of a sweetened fruit drink
Among adults, the sweetener disclaimer led to a lower perceived healthfulness of the sweetened drink for a child. Participants assigned to the no-disclaimer condition reported a higher mean perceived healthfulness of the sweetened fruit drink (3·17 ± 1·54) compared with those assigned to the sweetener disclaimer condition (2·74 ± 1·44; t = 4·66, P value: <0·001). The linear regression model adjusting for the belief that children should not consume foods or drinks with sweeteners (unbalanced variable across experimental groups) showed similar differences (β = −0·39, 95 % CI (−0·56, −0·21); Fig. 4, Panel A). This effect differed across age groups (Fig. 4, Panel B). The sweetener disclaimer was more effective in reducing the perceived product healthfulness for a child among adults aged 60 years and over compared with those aged 18–29 years (interaction β = −1·24, 95 % CI (−2·05, −0·44)). No other differences in sweetener label effects were observed among adults across key demographic characteristics (i.e. sex, income adequacy, ethnicity, education, nutrition knowledge, food shopping in the household, frequency of using the nutrition facts table, BMI category or the belief that children should not consume foods or drinks with sweeteners).
Fig. 4 Perceived healthfulness of a sweetened fruit drink with and without the sweetener disclaimer among adults. Estimations in the total sample (a) and across age categories (b) (n 1020). Linear predictions in the total sample were derived from linear regression models adjusted for the belief that children should not consume foods or drinks with sweeteners (unbalanced variable across experimental conditions). Liner predictions across age categories were derived from linear regression models introducing multiplicative interactions between age categories and experimental condition and adjusted for the same covariate
Similar results to the ones from adults were observed among youth. Youth assigned to the no-disclaimer condition reported a higher mean perceived healthfulness of the sweetened fruit drink (2·32 ± 0·96) compared with those assigned to the sweetener disclaimer (2·04 ± 0·96; t = 6·18, P value: <0·001). The linear regression model adjusted for ethnicity (unbalanced variable across experimental groups) showed similar differences between experimental conditions (β = −0·28, 95 % CI (−0·37, −0·19); Fig. 5, Panel A). This effect differed across sex (Fig. 5, Panel B) and the belief that children should not consume foods or drinks with sweeteners (Fig. 5, Panel C). The sweetener disclaimer was more effective in reducing the perceived beverage healthfulness among boys compared with girls (interaction β = −0·24, 95 % CI (−0·42, −0·06)). Meanwhile, the sweetener disclaimer led to a higher mean perceived healthfulness of the drink among those who disagreed or strongly disagreed that children should not consume foods or drinks with sweeteners (interaction β = 0·67, 95 % CI (0·21, 1·13), Fig. 5 Panel C) compared with those who agreed with this belief. No other differences in label effects across key demographic characteristics (i.e. age category, income adequacy, ethnicity, nutrition knowledge, frequency of using the nutrition facts table or BMI category) were observed among youth.
Fig. 5 Perceived healthfulness of a sweetened fruit drink with and without the sweetener disclaimer among youth. Estimations in the total sample (a), across sex (b) and the belief that children should not consume sweeteners (c) (n 1790). Linear predictions in the total sample were derived from linear regression models adjusted for ethnicity (unbalanced variable across experimental conditions). Liner predictions across categories were derived from linear regression models introducing multiplicative interactions between categorical variables and experimental condition and adjusted for the same covariate
Experiment 2. Effect of the caffeine disclaimer on the perceived healthfulness of a caffeinated drink
No differences in the perceived healthfulness of a caffeinated drink for a child were observed between adults assigned to the no-disclaimer condition (2·05 ± 1·30) v. the caffeine disclaimer (1·98 ± 1·32; t = 0·861, P value = 0·389) (Fig. 6). Similar results were observed in a linear regression model adjusted for nutrition knowledge (unbalanced variable across experimental groups) (β = −0·06, P value = 0·442). No differences in label effects were observed across key demographic characteristics (i.e. sex, age category, income adequacy, ethnicity, education, nutrition knowledge, food shopping in the household, frequency of using the nutrition facts table, BMI category or the belief that children should not consume foods or drinks with sweeteners).
Fig. 6 Perceived healthfulness of a caffeinated drink with and without the caffeine disclaimer among adults. n 1041. Linear predictions derived from linear regression models adjusted for nutrition knowledge
Discussion
This study found that shortly after implementing the new warning labels in Mexico, over 80 % of adult and youth participants were aware of the sweetener disclaimers. Over 60 % reported using this type of disclaimer sometimes or frequently. These findings are consistent with those from a study among Uruguayan adults showing that 77 % had seen nutritional warnings when making food purchases during the first month of full compliance(Reference Ares, Antúnez and Curutchet28) and studies in Chile reported that the regulation was well known in the country after the first year of implementation(Reference Correa, Fierro and Reyes29,Reference Dourado, Ramires and Flores30) . In Mexico, some retailers started applying warning labels and disclaimer stickers to products already in stock immediately after the regulation was approved in April 2020(Reference White and Barquera6). These stickers were progressively replaced by labels printed on the front of food packages by the end of March 2021. Therefore, consumers were exposed to the new warning labels and disclaimers well before manufacturers were required to comply with the regulation in October 2020 fully, either by a sticker (until 31 March 2021) or printed in the package.
Our randomised experiments showed that a sweetener disclaimer led youth and adults to perceive a sweetened drink as less healthy than a drink without a label. Although research on the health impacts of artificial sweeteners is novel, evidence has consistently questioned their safety for children(Reference Shum and Georgia10–Reference Sylvetsky, Jin and Mathieu13) and, more recently, for adults(Reference Debras, Chazelas and Srour14). Studies among adults(Reference Tang, Mars and James31), parents(Reference Sylvetsky, Greenberg and Zhao32,Reference Bucher and Siegrist33) and children(Reference Bucher and Siegrist33) have suggested that sweeteners are generally perceived as unhealthful due to their synthetic origin, carcinogenic and addictive potential. In line with these perceptions, our data showed that a high proportion of adults and adolescents believed children should not consume products with sweeteners. The inclusion of these disclaimers in the recently implemented warning label systems in Mexico and Argentina was based on this evidence. The findings of this study support the potential effect of this additional disclaimer in the warning label system to help adults and young consumers identify the harmful properties of industrialised drinks consumed by children and promote the purchase and consumption of beverages that do not have sweeteners. To further elucidate the potential impact of these warning labels on health outcomes, future experimental or real-life studies may wish to explore how sweetened drinks are substituted (e.g. by plain water or other types of processed drinks) among those who indeed modify their purchasing intentions due to the sweetener disclaimers.
Our findings also suggest that the effect of the sweetener disclaimer may differ across demographic characteristics. Among adults, sweetener disclaimers led to a lower perceived healthfulness of the sugary drink among older compared with younger participants. Similarly, a larger effect of warning labels on the perceived healthfulness ratings of a sugary drink was observed among older adults than younger adults from the UK(Reference Jáuregui, White and Vanderlee34). It has been stated that older adults have greater health risk perceptions (how individuals subjectively estimate and feel about the risks they face) compared with their younger counterparts(Reference Kim, Park and Kang35,Reference Dawood, Eweida and Rashwan36) . Others have suggested that older age is associated with an increased tendency to use nutrition labels because of an increased interest in healthy eating as individuals get older(Reference Ogundijo, Tas and Onarinde37) and/or increased efforts to meet specific dietary requirements as the metabolism declines with age(Reference Jeruszka-Bielak, Kollajtis-Dolowy and Santoro38). However, the measure in our study was about whether the beverage with the sweetener disclaimer was healthy for children, not for themselves. It is possible that this higher risk perception could also increase adults’ responsiveness to sweetener disclaimers on products for children.
Similarly, the effect of the sweetener disclaimer among youth differed by sex, with larger effects among males compared with females. To date, few studies have investigated the effect of front-of-pack labels among youth(Reference VanEpps and Roberto39–Reference Babio, Vicent and López41), and, to the authors’ knowledge, none has reported sex differences in their effects. Differences in usual beverage consumption patterns may help explain these differential findings by sex. Studies in Mexico and other countries suggest that male adolescents are more likely to consume unhealthy beverages, including soda(Reference Piernas, Barquera and Popkin42,Reference Kim, Yun and Oh43) . Therefore, males may be more susceptible to this disclaimer because they are heavier consumers of these sorts of drinks.
Among youth, the effect of the sweetener disclaimer also differed by the belief that children should not consume sweeteners. Opposite to the expected effect, the sweetener disclaimer led to a higher beverage perceived healthfulness among youth who disagreed that children should not consume foods or drinks with sweeteners. These results suggest that, among this population group, the disclaimer is misinterpreted as a positive attribute of the beverage. Mixed-methods studies by our research team have suggested that Mexican children and adolescents do not understand what sweeteners are and are generally interpreted as colour additives perhaps because these two words are very similar in Spanish: sweetener in Spanish is edulcorante, while colour additive is translated as colorante in this language (unpublished data). Additionally, although the Mexican regulation considers caffeine and sweetener disclaimers targeting children as defined by the United Nations (‘a person under the age of 18 years’)(44), this word in Mexico is generally interpreted as school-age children. Therefore, youth participants may not consider themselves children and may believe this artificially sweetened drink is healthier than a regular sugary drink. It is also possible that this population group could not identify sweeteners as harmful, possibly because the food industry promotes them as a positive ingredient, producing a potential misunderstanding of their health impact. Marketing used in some ‘diet,’ ‘organic’ or ‘healthier’ food products promotes sweeteners as a beneficial ingredient, frequently placed in front of the package as a desirable characteristic (e.g. using logos), indirectly indicating that it is better than sugar.
Our study also suggests that a caffeine disclaimer does not influence the perceived healthfulness of a caffeinated soda for children among Mexican adults. This null finding may be explained by the type of drink used to test the disclaimer. Soda is commonly assumed to be an unhealthy beverage, leaving little place for the caffeine disclaimer to modify this perception. However, we used this beverage because sodas comprise the vast majority of products labelled with the caffeine disclaimer in Mexico(Reference Contreras-Manzano, Cruz-Casarrubias and Munguía45). Another potential explanation for this finding may be related to the fact that caffeine is highly consumed among the Mexican population, including infants, children and youth, which could contribute to the normalisation of the high and extensive consumption of caffeinated drinks. In Mexico, introducing tea or coffee to the diet of infants is a common practice(Reference González-Castell, González de Cosío and Rodríguez-Ramírez20,Reference Swigart, Bonvecchio and Théodore46) , and coffee and soda are highly consumed by children and youth(Reference Piernas, Barquera and Popkin42). These consumption practices are deep-seated social norms and essential to Mexican cultural and culinary traditions, on special occasions (e.g. holidays) and daily life. However, we did not measure the perceived healthfulness of caffeine among participants. Despite the null effect of the caffeine disclaimer alone on the perceived healthfulness of a caffeinated soda observed in our study, future studies should explore how the caffeine disclaimer may contribute to the overall effect of warning labels and how the social normalisation of caffeine consumption in Mexico may contribute to a low perceived risk of caffeine. It is desirable to assess the effect of the caffeine disclaimer on youth since they are likely to consume coffee, sodas and energy drinks that contain caffeine(Reference Piernas, Barquera and Popkin42), and they may be unaware that sodas are unhealthy and of the health effects of caffeine(Reference Jain, Srivastava and Verma47).
The findings of this study may provide relevant insights for policy- and decision-makers regarding front-of-pack labelling systems and may be used to monitor and improve the warning label policy implemented in Mexico. Results clearly suggest that communication strategies should be implemented in Mexico, and possibly in other countries using similar warning label systems, to change the social norm around the consumption of caffeinated beverages by children and promote awareness of the negative impacts of caffeine on children’s health(Reference Jain, Srivastava and Verma47). In addition, these findings also offer valuable information for better targeting communication campaigns. These campaigns can include personalised messages about the adverse health effects of sugar, sweeteners and caffeine. By targeting different audiences appropriately, such campaigns can have a more profound and positive impact on healthy decision-making in the population. Despite plausible explanations for the counterintuitive findings on the effect of the sweetener disclaimer among specific population groups, our results emphasise the need to implement communication strategies that clearly explain what sweeteners are and how to interpret this disclaimer correctly, specifically among children. It is also necessary to apply more robust policies to the food industry regarding the ‘placement’ of sweeteners as something good for health and to regulate their use in foods and beverages and marketing, especially in those made for children.
To our knowledge, this is the first study investigating the potential impact of caffeine and sweetener disclaimers on the perceived healthfulness of industrialised beverages consumed by children. Strengths of this study include using a randomised design, limiting the influence of confounding from observed and unobserved factors, and a national sample of Mexican adults and youth. Nonetheless, results should be interpreted within the context of several limitations. Respondents were recruited using non-probability-based sampling; therefore, the findings do not provide nationally representative estimates. This study examined the potential impact of disclaimers on perceived healthfulness using a single-item measure. To expand evidence on the effectiveness of labels in communicating the relative healthfulness of products, future studies should explore the effect of labels using multiple measures across a range of healthy products, including direct comparisons between their healthfulness and the likelihood of purchase. Further, these disclaimers were evaluated alone on the front of the pack of the product, and the experiment was not performed in a store. In real-life situations, caffeine and sweetener disclaimers are generally presented in combination with other warnings (e.g. regular cola drinks are also labelled with warnings for calories, sugar, sweeteners and caffeine). Therefore, the effect may differ in a real-life situation or shopping environment.
Conclusion
There were high awareness and use of the disclaimers among study participants shortly after implementing the warning label regulation in Mexico. A sweetener disclaimer appears to be an effective intervention for helping consumers modify their perceived healthfulness of industrialised beverages consumed by children. Meanwhile, the caffeine disclaimer did not affect the perceived healthfulness of a caffeinated drink for a child among adults. Findings may be used by decision-makers to improve the front-of-pack labelling regulation in Mexico or other countries considering similar warning label systems. Our results highlight the need to implement communication strategies to promote awareness of the negative impacts of caffeine and sweeteners on children’s health and the correct interpretation of warning labels and provide relevant data for better targeting these efforts.
Acknowledgements
None.
Financial support
This work was supported by a Canadian Institutes of Health Research (CIHR) Project Grant (PJT-162167), with additional support from Health Canada, the Public Health Agency of Canada (PHAC) and a CIHR–PHAC Applied Public Health Chair (Hammond). Analysis and paper writing were supported by The National Institute of Diabetes, Digestive and Kidney Diseases of the National Institutes of Health under award number R01 DK128967 (JFT). CN was funded by Bloomberg Philanthropies (Award number 43003, SB) and was awarded with the healthy food policy fellowship from Vital Strategies. Funders had no role in designing the study, collecting, analysing and interpreting the data, drafting the manuscript nor the decision to publish findings. The supporting sources had no involvement or restrictions regarding publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Canadian Institutes of Health Research.
Conflicts of interest
There are no conflicts of interest.
Authorship
L.P.A.G. contributed to the statistical analysis and data interpretation and drafted the manuscript. A.J. conceived the study, participated in the design of the study, analysed and interpreted the data and helped draft the manuscript. C.N. and A.C.M. helped conceive and design the study, interpreted the data and substantively revised the manuscript. K.L.Q. helped interpret the data and substantively revised the manuscript. C.M.W., J.F.T., R.E.D., D.H. and S.B. helped conceive of the study, participated in the design of the study and substantively revised the manuscript. All authors read and approved the final manuscript and have agreed both to be personally accountable for their own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which they were not personally involved, are appropriately investigated, resolved and the resolution documented in the literature.
Ethics of human subject participation
This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved, reviewed by and received ethics clearance through a University of Waterloo Research Ethics Committee (ORE# 30829 and 41477) and the Research Ethics Committee from the National Institute of Public Health in Mexico (CI # 1641). Written informed consent was obtained from all subjects/patients.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980023002100
Open access
