Non-communicable diseases including cardiovascular disease, type 2 diabetes and obesity are the world’s leading causes of premature death and disability, with dietary intake being an important risk factor(1). In recent decades, a global dietary shift towards highly processed foods – including ultra-processed foods – has contributed to poor overall diet quality(1–Reference Moubarac, Parra and Cannon3). Ultra-processed foods are ‘formulations of food substances often modified by chemical processes and then assembled into ready-to-consume, hyper-palatable food and drink products using flavours, colours, emulsifiers, and a myriad of other cosmetic additives’(Reference Monteiro, Cannon and Levy2). These foods typically contain high amounts of Na, sugar, saturated or trans fats, leading to energy-dense, nutrient-poor food environments(Reference Monteiro, Cannon and Levy2–Reference Rauber, Louzada and Steele4).
Given that ultra-processed foods constitute more than half of energy intake in high-income countries including Canada, the USA and the UK(Reference Monteiro, Cannon and Levy2–Reference Baraldi, Martinez Steele and Canella5), and between one-fifth to one-third of energy intake in middle-income countries such as Mexico and Brazil(Reference Monteiro, Cannon and Levy2,Reference Louzada, Ricardo and Steele6,Reference Marrón-Ponce, Sánchez-Pimienta and Louzada7) , governments have adopted policy measures, such as nutrition labelling, to support healthy eating(Reference Feteira-Santos, Fernandes and Virgolino8). Nutrition labels are found on packaged foods and provide consumers with nutrient information at the point-of-purchase to aid informed decision-making in an increasingly processed food landscape(Reference Moubarac, Parra and Cannon3,Reference Feteira-Santos, Fernandes and Virgolino8) , while also incentivising the food industry to reformulate towards healthier nutritional profiles(1,Reference Feteira-Santos, Fernandes and Virgolino8,Reference Brambila-Macias, Shankar and Capacci9) . Nutrition labels implemented to date include back- or side-of-package nutrition facts tables (NFt) and front-of-package (FOP) labelling systems. NFt feature quantitative information on nutrient amounts, whereas FOP labels focus on simplified, interpretive information, often using symbols instead of numeric information to promote comprehension(10,Reference Hammond, Goodman and Acton11) .
In most cases, NFt implemented in different countries have a similar appearance and information content(10). In contrast, most FOP labels are voluntary and differ across countries. They may be nutrient-specific or interpretive summary indicator systems(Reference Hammond, Goodman and Acton11). Nutrient-specific FOP labelling systems highlight select nutrients of public health concern in the product, such as Mexico’s former Guideline Daily Amount label, which reinforces information also in the NFt, including calories, total sugars, saturated fats and Na(Reference Hammond, Goodman and Acton11). Summary indicator systems summarise nutrient content and interpret product healthfulness using algorithms to provide a score or ordinal ranking of the overall product(Reference Ikonen, Sotgui and Aydinli12). For example, Australia’s Health Star Rating assigns 0·5 to 5 stars to a food product, with higher star ratings corresponding with healthier options(Reference Hammond, Goodman and Acton11), whereas the UK has adopted a nutrient-specific Traffic Lights system indicating amounts of total fat, saturated fats, total sugars and Na in a product using colour-coding (high = red, medium = yellow and low = green)(Reference Hammond, Goodman and Acton11).
FOP labelling policies may be voluntarily implemented or mandatory in a given jurisdiction. The Mexican Guideline Daily Amount system was initially industry-led and later made mandatory by government, unlike the Health Star Rating and Traffic Light systems, which are government-led and voluntary. Voluntary policies provide food manufacturers with an option to opt out of implementing FOP labels. For example, the Health Star Rating appears on less than one-third of packaged food products(13), whereas in countries such as Chile and Mexico, FOP labels are mandatory and must be displayed on packaged products that exceed nutrient thresholds.
Consumer awareness and use of nutrition labels are key indicators of the visibility and effectiveness of labelling policies and related nutrition education initiatives. Awareness is indicative of consumers’ attention and exposure to labelling policy, thus precedes label use(Reference Miller and Cassady14). Label understanding is critical to – but does not guarantee – label use(Reference Miller and Cassady14). Label awareness, understanding and use are influenced by a range of factors, which have largely been explored via experimental or ‘pre-implementation’ studies(Reference Hammond, Goodman and Acton11). A growing number of pre-implementation studies suggest FOP labels are easier to understand than NFt, particularly among consumers with lower education and income(Reference Brambila-Macias, Shankar and Capacci9,Reference Campos, Doxey and Hammond15) . In comparison, greater use of NFt has been observed among women and those with higher income and education(Reference Campos, Doxey and Hammond15,Reference Sinclair, Goodman and Hammond16) . Moreover, consumers with specific motivation (i.e. diet-or weight-related goals), dietary behaviours (i.e. vegetarianism) and with prior nutrition knowledge have been associated with higher NFt label awareness and use(Reference Ikonen, Sotgui and Aydinli12,Reference Miller and Cassady14,Reference Malloy-Weir and Cooper17–Reference Christoph, Larson and Laska19) . Given the relative dearth of post-implementation research and recency of FOP labelling policies, research is needed to understand whether consumers who use FOP labels are similar to those who use NFt.
There is also little post-implementation data that compare use and awareness of FOP labels across different countries, or NFt to FOP label use within countries with both label types. These evidence gaps limit our ability to evaluate the uptake and effectiveness of different labelling policies across subgroups (i.e. among consumers with higher v. lower health literacy status) and countries, which may inform policy adoption or dissemination strategies in countries considering FOP labelling systems. This study thus aimed to examine differences in nutrition label awareness and use across five countries (Australia, Canada, Mexico, the UK and the USA), three of which have government-led FOP labelling policies in place (Australia, Mexico and the UK). In particular, this study explored between-country differences in NFt use and awareness; correlates of NFt and FOP label use and awareness, including sociodemographic, dietary and knowledge-related characteristics; and NFt v. FOP label use and awareness in countries with both.
Methods
Study design and participants
This study used cross-sectional data from the 2018 wave of the International Food Policy Study(Reference Hammond, White and Rynard20). Respondents aged 18 years and over and were recruited in Australia, Canada, Mexico, the UK and the USA via Nielsen Consumer Insights Global Panel and their partners’ panels and completed web-based surveys between November and December 2018. The Nielsen panels use probability and non-probability recruitment methods in each country. Email invitations were sent to a random sample of panellists after targeting for age and sex in each country. Quotas were applied to facilitate recruitment of a diverse sample that approximated known proportions in each country for males and females in four age groups: 18–29, 30–44, 45–64, and 65 years and over. Respondents were queried about a range of topics related to nutrition and the food environment, including food purchasing, dietary behaviours, nutrition knowledge and perceptions of national-level food policies. Surveys were conducted in English in Australia and the UK; Spanish in Mexico; English or French in Canada; and English or Spanish in the USA(Reference Hammond, White and Rynard20).
All respondents provided consent prior to completing the survey and received remuneration in accordance with their panel’s usual incentive structure (e.g. points-based, monetary rewards or chances to win prizes).
Of the 22 824 respondents who completed the 2018 IFPS survey, a subsample of 21 586 respondents from Australia (n 3901), Canada (n 4107), Mexico (n 4012), the UK (n 5121) and the USA (n 4445) were included in the current study. Those with missing data for ethnicity (n 296), income adequacy (n 182), education (n 69), food shopping role (n 29), dietary efforts (n 122), health literacy status (n 29); self-reported NFt awareness (n 157) and use (n 184); self-reported FOP label awareness (n 201) and use (n 201); and Food Processing Knowledge (FoodProK) (n 17) were excluded from analyses. Respondents with missing data were not different with respect to label awareness and use compared with the rest of the sample (data not shown). The median time to complete the survey across all countries was 40 min.
Measures
Self-reported awareness and use of food labels
Label awareness was measured by showing respondents country-specific NFt (Table 1) and asking, ‘Have you seen this type of food label on packages or in stores?’ (response options were never/rarely/sometimes/often/all the time). Label use was measured by asking, ‘How often do you use this type of food label when deciding to buy a food product?’ (never/rarely/sometimes/often/all the time). These measures were adapted from the 2014 US Food and Drug Agency Health and Diet Survey(Reference Lin, Zhang and Carlton21). After answering questions about the NFt, respondents from Australia, Mexico and the UK were shown images of the FOP labels in place in their countries at the time, including voluntary Health Star Ratings, mandatory Guideline Daily Amounts, and voluntary multiple Traffic Lights, respectively (Table 1), and asked to respond to the same measures of label awareness and use. All labelling variables were queried using a five-point response scale and analysed as continuous variables. Potential correlates of label awareness and use were identified from the literature and included nutrition knowledge, consumer dietary behaviours, BMI and sociodemographic characteristics.
Table 1 Food labels by country in the 2018 International Food Policy Study survey
NFt, nutrition facts table; FOP, front-of-package.
Functional nutrition knowledge
Prior nutrition knowledge may influence consumers’ ability and motivation to use nutrition labels(Reference Miller and Cassady14,Reference Hess, Visschers and Siegrist18) . The survey assessed consumer nutrition knowledge using the FoodProK score, a functional test based on level of processing(Reference Bhawra, Kirkpatrick and Hall22). Respondents viewed and rated images of three food products within four categories: fruits (apple, apple juice and apple sauce), meat (chicken breast, deli chicken slices and chicken nuggets), dairy products (1 per cent milk, cheese block and processed cheese slices) and grains (oats, cereal and cereal bar). Products in each category were selected based on availability in multiple international contexts and to represent different levels of processing according to the NOVA system(Reference Monteiro, Cannon and Levy2). The twelve product images and corresponding NFt and ingredients lists were displayed one at a time, in random order. For each product, respondents were asked, ‘Overall, how healthy is this food product?’ and answered using a scale of 0 to 10, with 0 representing ‘not healthy at all’ to 10 indicating ‘extremely healthy’.
FoodProK scores were calculated based on the concordance of healthiness ratings within each food category to NOVA classification rankings, with less processed foods representing higher healthiness. Respondents received a full score of 2 if their food product ratings corresponded with the order of NOVA food processing groups (e.g. apple > apple sauce > apple juice). If the respondent ranked 2 of 3 products in a given category in accordance with NOVA (e.g. apple > apple juice > apple sauce), they received a score of 1. Zero was assigned if the respondent’s rankings did not align with those based on NOVA. Scores were summed across the four food categories to create the total FoodProK score (hereon referred to as nutrition knowledge score), ranging from 0 to 8(Reference Weiss, Mays and Martz23).
Health literacy status
Respondents completed an adapted version of the Newest Vital Sign in which an ice cream container NFt was shown, and respondents answered six questions that assessed their ability to make mathematical calculations (numeracy), read and apply label information (prose literacy), and understand the label information (document literacy)(Reference Blitstein and Evans24). Based on the number of correct answers, respondents were sorted into one of three literacy categories where a score of 0–1 suggested ‘high likelihood (50 % or more) of limited literacy’; a score of 2–3 indicated ‘possibility of limited literacy’; and a score of 4–6 indicated ‘high likelihood of adequate literacy’(Reference Blitstein and Evans24). The Newest Vital Sign served not only as a proxy measure of health and nutrition literacy, but also as a functional measure of consumer NFt understanding. The NFt images were adapted to include NFt design and layout specific to each country(Reference Hammond, White and Rynard20). A score between 0 and 6 was calculated based on the number of correct answers, with higher scores corresponding with greater NFt understanding.
Consumer dietary behaviours
Diet modification efforts, another possible predictor of label awareness and use(Reference Hess, Visschers and Siegrist18,Reference Christoph, Larson and Laska19) , were measured by asking, ‘Have you made an effort to consume more or less of the following in the past year?’ Respondents answered, ‘consume less’, ‘consume more’ or ‘no effort made’, to a list of nutrients and food categories. This study focused on efforts in five categories that have received increasing attention in policies such as dietary guidelines: ‘trans-fats’, ‘sugar/added sugars’, ‘salt/sodium’, ‘calories’, and ‘processed foods’(1,Reference Monteiro, Cannon and Levy2,Reference Rauber, Louzada and Steele4) . A value of -1 was assigned to ‘consume less’, +1 to ‘consume more’, and 0 for ‘no effort made’ for each of the five categories. Five points were added to the sum of the five categories to create a scale ranging from 0 to 10, with 0 representing ‘consume less’ responses to all categories, 10 representing ‘consume more’ responses to all categories and the range between reflecting all other response combinations.
Consumers with specific dietary practices, as well as those with a primary food shopping role in their households, were hypothesised to have greater interest in and exposure to labels(Reference Hess, Visschers and Siegrist18,Reference Christoph, Larson and Laska19,25) . Respondents indicated whether they followed any dietary practices (vegetarian/vegan/pescatarian/a religious practice for eating). Responses were recoded to indicate no specific dietary practices or one or more dietary practice. Food shopping role was captured by asking, ‘Do you do most of the food shopping in your household?’ (Yes/No/Share equally with others)(Reference Hammond, White and Rynard20).
Sociodemographic variables and BMI
To capture differences in nutrition label awareness and use based on sociodemographic characteristics, age group (18–29, 30–44, 45–59 and ≥60 years), sex at birth (female or male), country (Australia, Canada, Mexico, the UK and the USA), and derived variables for education and ethnicity were included in analyses. Given that less than 1 % (n 113) of respondents reported a gender different than their biological sex, only the variable ‘sex at birth’ was used in analyses. Education level was categorised in accordance with country-specific criteria, with respondents classified as having ‘low’ (high school completion or lower), ‘medium’ (some post-secondary school qualifications, including some university) or ‘high’ (university degree or higher) levels of education(Reference Hammond, White and Rynard20). Ethnicity was treated as a binary variable to enable between-country comparisons, with respondents categorised as ‘majority’ in Mexico if they identified themselves as ‘non-Indigenous’, and ‘majority’ in Australia, Canada, the UK and the US if they identified themselves as ‘White’, predominantly English-speaking, or non-Indigenous based on country-specific ethnicity questions(Reference Hammond, White and Rynard20). Income adequacy, which refers to whether an income is enough to support an individual or household, was assessed by asking, ‘Thinking about your total monthly income, how difficult or easy is it for you to make ends meet?’ (Very difficult/Difficult/Neither easy nor difficult/Easy/Very easy)(Reference Hammond, White and Rynard20). Income adequacy was used instead of household income to ensure relevance of this measure across countries.
Weight status may play a role in consumers’ use or interest in nutrition labels, particularly among those with weight-related goals(Reference Christoph, Larson and Laska19). Categorisation of BMI followed WHO criteria(Reference Benjamini and Hochberg26), with self-reported height and weight used to classify respondents based on BMI < 18·5 kg/m2, 18·5 to 24·9 kg/m2, 25·0 to 29·9 kg/m2 and ≥30 kg/m2. Given the large number of cases with missing height and weight data – including those who selected ‘don’t know’ or ‘refuse to answer’ – a separate category for ‘missing’ BMI was created and retained as a response category for analyses.
Statistical analysis
Descriptive statistics were used to summarise the sample profile and labelling outcomes by country. Three multiple linear regression models were fitted to examine NFt/FOP use and NFt awareness across the five countries. All models were adjusted for sociodemographic characteristics (age, sex, country, income adequacy, education level and ethnicity), consumer dietary behaviours (dietary practices, modification efforts and food shopping role) and BMI. Due to the moderate correlation between the nutrition knowledge score and Newest Vital Sign (r s = 0·37, P < 0·0001), nutrition knowledge score was added to the main model in a subsequent step to assess the association of nutrition knowledge with the labelling outcomes.
Multiple comparisons were conducted to assess all pairwise contrasts for categorical variables. The Benjamini–Hochberg procedure was applied to decrease the false detection rate following multiple exploratory tests(Reference Jukes27). All statistically significant pairwise contrasts were reported after applying the Benjamini–Hochberg procedure, assuming a false discovery rate of 10 %. The models tested two-way interactions between country and the covariates age, sex, ethnicity, education, income adequacy, BMI, health literacy status, dietary practices, dietary efforts, and food shopping role, as research has shown differences in label awareness and use based on these characteristics(Reference Miller and Cassady14,Reference Campos, Doxey and Hammond15) .
Generalised linear mixed models were run separately for Australia, the UK and Mexico to test awareness of NFt v. FOP labels, and use of NFt v. FOP labels. A repeated-measures analysis was used to account for the correlated data within individuals for these measures. Each model included two-way interactions for the individual-level variables above to assess whether awareness/use differed for NFt v. FOP labels among these subgroups. Finally, Spearman’s rank correlations tested the correlation between the four self-reported labelling outcomes (NFt awareness and use, FOP label awareness and use).
Statistical analyses were conducted using SAS Studio (SAS Institute). Parameter estimates were reported with 95 % CI. Data were weighted with post-stratification sample weights constructed using population estimates from respective country-based censuses based on age group, gender, region, ethnicity (except in Canada as the national census did not include a simple measure of ethnicity suitable for creating weights) and education (except in Mexico, where the proportion of respondents with lower educational attainment was much smaller than population estimates from census data)(Reference Hammond, White and Rynard20). All reported estimates are weighted.
Results
Sample characteristics are presented in Table 2.
Table 2 Sample characteristics (n 21, 586), International Food Policy Study, 2018*
* All reported estimates are weighted.
† Mean and sd reported for dietary efforts, health literacy score and nutrition knowledge score.
Patterns and correlates of nutrition facts table use and awareness
Figure 1 shows mean NFt use and awareness across countries (categorical responses can be seen in Supplemental Tables 1 and 2). The cross-country data showed that respondents from the USA, Canada and Australia reported significantly higher NFt use than respondents from the UK, and respondents from Mexico reported the lowest use among all countries (see Table 3). Similarly, NFt awareness was highest among respondents from the USA, followed by Canada, Australia, the UK and Mexico. A Spearman rank correlation indicated a moderate correlation between self-reported NFt use and awareness across all countries (r s = 0·41, P < 0·0001).
Fig. 1 Nutrition facts table and front-of-package label awareness and use by country. Mean levels of awareness and use are shown with 95 % CI. A mean of 1 indicates no awareness/use, and 5 indicates the highest level of self-reported awareness/use. The asterisk denotes significantly different label awareness/use than the reference country (Mexico) at P < 0·05. NFT, nutrition facts table; FOP, front-of-package
Table 3 Sociodemographic and behavioural correlates of nutrition facts table and front-of-package label use, International Food Policy Study, 2018†
β, parameter estimate; NFt, nutrition facts table; FOP, front-of-package.
* Variables are significant (P < 0·05) after post hoc adjustment using Benjamini–Hochberg procedure.
† All reported estimates are weighted.
The pattern of sociodemographic characteristics is shown in Table 3. Age, sex, ethnicity, education and BMI were significantly associated with NFt use. Younger respondents reported higher NFt use compared with older respondents. Females reported higher NFt use than males, and respondents from ‘majority’ ethnic groups in their respective countries reported lower NFt use than ‘minority’ ethnic groups. Respondents categorised as having ‘high’ education levels reported higher NFt use than those with ‘medium’ or ‘low’ education, and NFt use was higher with higher income adequacy. Finally, NFt use was lower among respondents with BMI over 30 compared with those with BMI between 18·5–24·9, 25–29·9 and the ‘missing’ category.
Consumers’ dietary behaviours were also associated with NFt label use, as respondents who were primary food shoppers or shared this responsibility equally with others reported higher NFt use than those who were not primary food shoppers in their households. Respondents engaging in vegetarian or other dietary practices, as well as those making efforts to reduce calories, Na, sugars, trans-fats or processed food intake, reported higher use of NFt than those not engaging in specific dietary practices or efforts.
Knowledge-related factors including health literacy status and functional nutrition knowledge were significantly associated with NFt use. Respondents with ‘adequate health literacy’ reported higher NFt use compared with those with a ‘possibility of limited health literacy’ and a ‘high likelihood of limited health literacy’. NFt use was also higher among respondents with higher nutrition knowledge (β: 0·07, 95 % CI (0·05, 0·07), P < 0·0001).
When comparing NFt use to awareness, a similar pattern of correlates was observed, with the exception of education for which respondents with ‘high’ education reported lower NFt awareness than those with ‘low’ education levels (Table 4). When functional nutrition knowledge was added to this model, NFt awareness was higher among respondents with higher nutrition knowledge scores (β: 0·06, 95 % CI (0·05, 0·07), P < 0·0001).
Table 4 Sociodemographic and behavioural correlates of nutrition facts table awareness, (n 21 586), International Food Policy Study, 2018†
β, parameter estimate; NFt, nutrition facts table.
* Variables are significant (P < 0·05) after post hoc adjustment using Benjamini–Hochberg procedure.
† All reported estimates are weighted.
There were differential patterns across countries for NFt use based on age, sex, ethnicity, education level, income adequacy, health literacy and dietary efforts (see online Supplemental Table 3). Women in Mexico reported lower NFt use than UK women; however, Mexican respondents with ‘high’ education and income adequacy reported higher NFt use compared with respondents in the UK with similar education and income adequacy. Australian respondents with ‘adequate health literacy’ reported higher NFt use than ‘adequate health literacy’ respondents in the UK. For NFt awareness, Canadian respondents with ‘high’ education reported lower NFT awareness than those with similar education in the UK. Mexican respondents with ‘adequate literacy’ and a ‘possibility of limited literacy’ reported lower NFT awareness than the corresponding health literacy groups in the UK.
Patterns of front-of-package labelling and correlates
Cross-country data in the three countries with FOP labelling policies found that respondents from Mexico reported the highest awareness of FOP labels (mean 4·0), followed by the UK (mean 3·9) and Australia (mean 3·5) (Fig. 1). In addition, respondents in the UK reported the highest FOP label use (mean 3·2) and Australia the lowest (mean 2·9). FOP label use and awareness were moderately correlated (r s = 0·39, P < 0·0001). Correlates of FOP label use were similar to NFt use, with a few exceptions (Table 3). A review of sociodemographic and other correlates found that sex and ethnicity were not significantly associated with FOP label use, and respondents with ‘adequate literacy’ reported lower FOP label use compared with those with a ‘high likelihood of limited literacy’. Nutrition knowledge score was not significantly associated with FOP label use (β: 0·01, 95 % CI (0·00, 0·02), P = 0·1978).
Use and awareness of nutrition facts table v. front-of-package labels
Significant differences were found between NFt and FOP label use and awareness within countries with both label types. In Australia, respondents reported higher use and awareness of NFt compared with voluntary FOP Health Star Ratings. As shown in Table 4, respondents aged ≥60 years were more likely to be aware of and use NFt than Health Star Ratings compared with 18–29-year-olds. Female respondents and those with higher income adequacy were also more likely to use NFt than Health Star Ratings. Respondents with ‘adequate literacy’ were more likely to report higher NFt than FOP label use and awareness compared with those with a ‘high likelihood of limited literacy’. Respondents with higher nutrition knowledge were more likely to use and be aware of NFt than FOP labels. Specific dietary practices or efforts to consume less of specific nutrients (i.e. sugar, Na and trans fat) were associated with higher NFt than FOP label use, and primary food shoppers were less likely to be aware of NFt than FOP labels compared with respondents who were not primary food shoppers in their households.
In the UK, respondents reported lower NFt use and awareness compared with the voluntary FOP Traffic Light labels. Older age groups (60+, 45–59 and 30–44 years compared with 18–29 years) were more likely to be aware of or use NFt compared with FOP Traffic Lights. Respondents who identified as belonging to the ‘majority’ ethnic group in the UK were more likely to report higher FOP label than NFt use and awareness compared with those from ‘minority’ ethnic groups. Respondents with ‘high’ education levels were significantly more likely to be aware of NFt than FOP labels compared with respondents with ‘medium’ education levels. Similarly, respondents with ‘adequate literacy’ were more likely to report higher use and awareness of NFt than FOP labels compared with respondents with a ‘high likelihood of limited literacy’. Respondents engaging in efforts to consume less Na, sugar, trans-fat, processed food or calories were more likely to use FOP labels than NFt.
In Mexico, respondents reported higher NFt use and awareness compared with Guideline Daily Amount labels. Older age groups and females were more likely to report higher NFt than FOP (Guideline Daily Amount) label awareness compared with 18–29-year-olds and males, respectively. Respondents who reported higher nutrition knowledge and those with ‘adequate literacy’ were more likely to report higher FOP label than NFt awareness compared with those with lower nutrition knowledge scores or a ‘high likelihood of limited literacy’, respectively. Dietary efforts to consume less of specific nutrients were also associated with higher FOP label than NFt use. There were no significant differences between NFt and FOP label use among the subgroups tested in Mexico. Interactions between country and BMI were not significant for NFt awareness or use in Australia, the UK or Mexico.
Discussion
Multi-country, population-level studies are important for ascertaining which labelling policies are most effective. Country-specific differences in label awareness and use provide insights into which labels have the greatest reach among consumers from varied subgroups. Evidence has shown greater uptake for mandatory labelling policies(10), consistent with findings from this study demonstrating significantly higher NFt use compared with voluntary FOP labelling systems (with the exception of Traffic Lights in the UK), and higher awareness of the mandatory FOP Guideline Daily Amount label in Mexico compared with voluntary FOP labelling systems. NFt have been a long-standing policy in all five countries(Reference Wartella28–Reference Van Camp, De Souza Monteiro and Hooker32); hence, high levels of use and awareness were not surprising. Significantly higher levels of NFt awareness and use observed in this study were in the USA. As the first country to enact nutrition labelling policies, as well as a greater reliance on processed, packaged foods, higher levels of NFt awareness among Americans may be attributed, in part, to these factors(Reference Baraldi, Martinez Steele and Canella5).
Among countries with FOP labelling systems, Guideline Daily Amount labels in Mexico had the lowest level of reported use, despite high levels of awareness. Mexico is the only country in this study with a mandatory FOP label, so greater awareness of Guideline Daily Amounts may have stemmed from relatively higher exposure to this label on the FOP compared with voluntary FOP labels(10). The voluntary nature of FOP labels in Australia and the UK may account for lower levels of awareness and use compared with mandatory NFt and may have resulted in lower FOP label exposure as Traffic Light labels and Health Star Ratings are estimated to appear on approximately 8 % and 30 % of food products, respectively(13,Reference Nieto, Jauregui and Contreras-Manzano33) . Lower awareness of voluntary labels supports consideration of mandatory FOP labelling policies and also reiterates the importance of closely monitoring policy implementation.
Existing evidence also highlights that not all FOP labels are equal. The finding that self-reported use of the mandatory Guideline Daily Amount label in Mexico was significantly lower than voluntary FOP label use in Australia and the UK is consistent with literature documenting consumers’ difficulty understanding these labels(Reference Hammond, Goodman and Acton11,Reference Arrúa, Vidal and Antúnez34–Reference White and Barquera37) . Indeed, the Mexican government is replacing Guideline Daily Amounts with mandatory FOP ‘high-in’ labels similar to Chile(Reference Taillie, Hall and Popkin38), as emerging evidence demonstrates ease of use and greater understanding of this simple, interpretative label format(Reference Grummon, Smith and Golden39,Reference Kanter, Reyes and Vandevijvere40) . While the present study did not examine the impact of mandatory v. voluntary nutrition labels on product reformulation, evidence has shown that labels have the potential to incentivise healthier reformulation of packaged foods that exceed thresholds for nutrients of public health concern(Reference Feteira-Santos, Fernandes and Virgolino8,Reference Brambila-Macias, Shankar and Capacci9,Reference Taillie, Bercholz and Popkin41) . There is also encouraging evidence of positive impacts of mandatory labelling on consumers’ food purchasing, with one longitudinal study in Chile finding increased healthy food and decreased sugar-sweetened beverage purchases after the implementation of FOP warning labels(Reference Mejean, Macouillard and Peaneau42).
Differences in nutrition label use and awareness based on consumer knowledge-related factors (i.e. health literacy status, functional nutrition knowledge) and sociodemographic characteristics have important policy implications, particularly regarding the design of accessible nutrition labels. Consistent with the literature(Reference Campos, Doxey and Hammond15,Reference Sinclair, Goodman and Hammond16) , this study found that respondents with higher education, health literacy status and functional nutrition knowledge reported higher NFt use and awareness, likely reflecting better numeracy skills and ability to understand label information(Reference Blitstein and Evans24,Reference Arrúa, Vidal and Antúnez34) . Evidence has shown that consumers with lower literacy or nutrition knowledge may be at a disadvantage for applying nutrition information from labels, which could limit their ability to make healthier purchasing decisions(Reference McKinnon, Giskes and Turrell43–Reference Lassen, Lehmann and Andersen45). In response, FOP labels were designed to make nutrition information simpler to interpret; thus, higher self-reported FOP label use among those with lower health literacy status suggests greater accessibility of interpretative FOP label information compared with NFt.
Despite different rates of usage, there were many similarities in the correlates of NFt and FOP label use, with generally higher use among primary food shoppers, respondents with specific dietary practices or diet modification efforts, respondents with BMI under 30, and females compared with males. Research has shown that consumers following specific dietary practices or with diet-related goals have increased motivation to seek out nutrition information, which likely drives higher label use(Reference Hess, Visschers and Siegrist18,Reference Christoph, Larson and Laska19,25) . While primary food shoppers may not necessarily be making specific dietary efforts, they may be making food choices for others in their household (i.e. children), potentially motivating greater use of labels than those who are not primary food shoppers(25). Moreover, studies have found women to be more health conscious than men, leading to greater use of nutrition information(Reference Ikonen, Sotgui and Aydinli12,Reference Moon, Balasubramanian and Rimal46) . In order for the general public to take an interest in nutrition labelling policy, health promotion campaigns which aid consumers in identifying how good nutrition fits in the broader context of their health may encourage label use. Pairing complementary initiatives together (i.e. nutrition labelling and school-based curricula) can increase awareness and accessibility of nutrition labelling information beyond groups that have a vested interest in nutrition information.
Several sociodemographic characteristics were significantly associated with nutrition label use, with generally lower use among older age groups and those with lower income adequacy. One potential explanation for lower label use among older individuals may be brand or product familiarity(Reference Ikonen, Sotgui and Aydinli12). Studies on product health claims have shown that consumers who are familiar with a product are less likely to read labels or claims(Reference Ikonen, Sotgui and Aydinli12,Reference French, Tangney and Crane47) ; hence, NFt and FOP labels may not be used by habitual consumers unless they are considering a new brand or product(Reference Gorton, Ni Mhurchu and Chen48). Households with low incomes report prioritising accessibility and affordability when making food purchasing and consumption decisions(Reference Turrell and Kavanagh44,Reference Feng and Fox49) . As a result, these consumers may report using nutrition label information less often due to other priorities aside from nutrition quality. Understanding the sociodemographic characteristics that drive nutrition label use is critical for policymakers to consider, particularly in shaping the broader nutrition environment. For example, consumers prioritising factors such as affordability over product healthiness reiterates the need for a system-wide approach to improving the food system. Prompted by mandatory nutrition labelling(13,Reference Taillie, Bercholz and Popkin41) , research has shown promising improvements in healthier product reformulation, a system-level strategy which is necessary given the prominence of ultra-processed foods(1,Reference Moubarac, Parra and Cannon3) .
There are very few studies exploring label use differences by ethnicity(Reference Merill and Richardson50), and this study found higher use of NFt among ‘minority’ respondents. The dearth of literature exploring disparities in the reach of nutrition labelling policy limits our ability to unpack why some ethnic groups use labels more or less. For example, in a study of menu label use, Feng & Fox (2018) found that Black and Hispanic groups used labels more than their White counterparts at sit-down restaurants(Reference Merill and Richardson50), whereas another study in New Zealand found lower use of NFt labels among minority ethnic groups(Reference Mejean, Macouillard and Peaneau42). More research is required to better understand what other factors may be driving NFt use, and whether immigrant status or cultural food preferences may play a role in product familiarity and label use.
This study compared label awareness and use between five countries, with a large sample that enabled consideration of a range of covariates. A limitation is the non-probability-based sampling strategy, which does not enable the generation of nationally representative population estimates. Although analyses included post-stratification weights to make the sample more similar to the age, sex, region and ethnicity distributions in each country, the Mexico sample had higher educational attainment than in the Mexican population, while self-reported BMI was lower than national estimates in each of the five countries(Reference Hammond, White and Rynard20). Moreover, the primary outcomes (NFt/FOP label awareness/use) and several other correlates (nutrition knowledge score, BMI) are subject to social desirability bias given the use of self-reported measures. In addition, the Newest Vital Sign has been tested across a variety of age and ethnic groups in different countries but has not yet been validated for online, self-administration(Reference Malloy-Weir and Cooper17,Reference Blitstein and Evans24) ; however, that pattern of results we found provide evidence of its construct validity across countries. Lastly, the cross-sectional study design limits possible conclusions about the direction of variable relationships such as label use and nutrition knowledge.
Conclusions
The study findings are relevant as an increasing number of countries adopt voluntary or mandatory FOP labelling policies. Lower use of Guideline Daily Amount labels compared with voluntary FOP labelling systems provides further support for Mexico’s decision to switch to mandatory FOP ‘high-in’ symbols. Sociodemographic and other subgroup differences in label use are important as they indicate the reach of various labelling policies, which can potentially translate to dietary choices. Considering the varied use and awareness of nutrition labels among consumers in each country, accessible nutrition labelling policies (i.e. use of FOP labels) as well as broader nutrition promotion initiatives which address the diversity of consumers (i.e. label promotion campaigns in different languages) can aid efforts to improve diet quality. Future research should investigate the implications of labelling policies on consumers’ eating patterns over time, including countries with mandatory FOP labelling policies, as well the impact of nutrition labelling policies on food reformulation.
Acknowledgements
Acknowledgements: The authors would like to acknowledge Christine M. White for her role in the development of measures for the International Food Policy Study and support of this paper. Financial support: Funding for the International Food Policy Study was provided by a Canadian Institutes of Health Research (CIHR) Project Grant (#PJT-162167), with additional support from an International Health Grant, the Public Health Agency of Canada (PHAC), and a CIHR-PHAC Applied Public Health Chair held by David Hammond. K01HL147713 from the National Heart, Lung, and Blood Institute of the NIH supported Marissa Hall’s time working on the paper. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Authorship: All authors contributed to development of the survey and data collection. J.B. and D.H. conducted analyses and wrote the first draft with contributions from S.I.K., M.G.H., L.V., J.F.T. and A.J.M. All authors reviewed and commented on subsequent drafts and provided final approval of the manuscript. 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 by the University of Waterloo Research Ethics Committee (ORE# 30829). Written informed consent was obtained from all participants.
Conflicts of interest:
The authors have no conflicts of interest to declare.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980022002257
Open access
