A five-country evaluation of nutrition labelling policies: consumer use, understanding, and knowledge of processed foods

Abstract

Background: Poor dietary intake is a critical risk factor for non-communicable diseases – the world’s leading cause of premature death and disability. Globally, consumption of highly processed foods has increased in recent decades. Population-health interventions, such as nutrition labelling, have the potential to promote healthy eating behaviours. Nutrition Facts tables (NFts) and front-of-package (FOP) labelling systems provide consumers with essential nutrition information at the point-of-purchase to aid healthy decision-making and encourage healthier product reformulation. An increasing number of countries are implementing FOP labelling systems, ranging from ‘high in’ labels in Chile to Health Star Ratings in Australia. There is a need to examine consumers’ knowledge of levels of food processing, as well as awareness, understanding, and use of nutrition labels to better understand the impact of labelling policies. Objectives: This dissertation explored patterns and correlates of nutrition label awareness, understanding, and use, as well as functional nutrition knowledge across five countries with different nutrition labeling systems. Canada and the US currently have NFts only, which were compared to NFt and FOP labelling systems in Australia (voluntary Health Star Rating FOP labels), the UK (voluntary multiple Traffic Light FOP labels) and Mexico (mandatory Guideline Daily Amount FOP labels). The four primary aims of this study were to: 1) assess face and content validity of a new functional nutrition knowledge measure based on level of food processing – the Food Processing Knowledge (FoodProK) score; 2) determine functional nutrition knowledge levels (FoodProK scores) and associated correlates; 3) identify and compare patterns and correlates of self-reported versus functional label understanding; and 4) explore patterns and correlates of label awareness and use across countries. Methods: This dissertation consisted of four sub-studies: Study 1 developed and tested a new functional measure of nutrition knowledge which was based on respondents' ability to understand and apply the concept of food processing in a functional task; Study 2 examined patterns and correlates of functional nutrition knowledge across countries; Study 3 explored self-reported (NFt, FOP label) and functional (NFt) label understanding across countries; and Study 4 examined patterns of NFt and FOP label use and awareness across countries. Cross-sectional data from the 2018 International Food Policy Study were used. Respondents aged ≥18 years (n = 22,824) from Australia (n = 4103), Canada (n = 4397), Mexico (n = 4135), the UK (n = 5549), and the US (n = 4640) were recruited through Nielsen Consumer Insights Global Panel and their partners’ panels. Respondents completed web-based surveys answering questions about food policies, dietary behaviours, health literacy, and other factors related to food environment. The primary outcomes were functional nutrition knowledge; self-reported label (NFt, FOP) awareness, understanding, and use; and functional NFt understanding. Sociodemographic factors (age, sex, ethnicity, country, education level, income adequacy), body mass index, dietary behaviours (dietary practices, diet modification efforts, food shopping role), and knowledge-related characteristics (health literacy, FoodProK score) were included in all analyses. In Study 1, content validity of the newly developed FoodProK score was examined by surveying Registered Dietitians in Canada (n = 64). Dietitians completed the FoodProK measure, which required rating the healthiness of three food products in four categories (fruit, dairy, grain, meat). Thereafter, dietitians answered several open-ended survey questions about the measure. One-way repeated-measure ANOVA models tested whether dietitians’ product ratings were significantly different between products and food categories. Multiple linear regression models were fitted to examine between-country differences in functional nutrition knowledge in Study 2. Studies 3 and 4 also used multiple linear regression models to assess correlates of label understanding and use/awareness, respectively. All analyses adjusted for sociodemographic, dietary behaviours, and knowledge-related characteristics. Interaction terms with country and sociodemographic characteristics were tested to examine how patterns differed across countries. Results: Study 1 – Overall, 70.3% of dietitians scored 7 and above on the 8-point FoodProK measure. The majority of dietitians rated food products in congruence with level of processing (85.9% of dietitians correctly ordered products in the fruit and dairy categories; 73.4% correctly ordered grain products). The meat category was an exception, with approximately half of dietitians (54.7%) rating meat products in accordance with level of processing. Open-ended responses showed dietitians did not perceive meaningful differences between the processed meat products. Overall, 80% of dietitians reported level of processing as an important indicator of the healthiness of foods. Preliminary content validity evidence suggests knowledge of food processing levels as one indicator of general nutrition knowledge. Study 2 –The highest FoodProK scores were reported in Canada (mean: 5.1) and Australia (5.0), followed by the UK (4.8), Mexico (4.7), and the US (4.6). Health literacy and self-rated nutrition knowledge were positively associated with FoodProK scores (p<.0001). FoodProK scores were higher among those who reported specific dietary practices such as vegetarianism (p<.0001); made efforts to consume less sodium, trans fats, sugars, processed foods, or calories (p<.0001); respondents classified as having ‘adequate health literacy’ (p<.0001); respondents who self-reported being ‘very knowledgeable’ or ‘somewhat knowledgeable’ about nutrition (p<.0001); those who were 60+ years old (p=0.0023), women (p<.0001), and respondents who belonged to the ‘majority’ ethnic group in their respective countries (p<.0001). Education, income adequacy, and food shopping role were not significantly associated with FoodProK scores. Study 3 – Self-reported and functional NFt understanding was highest in the US and Canada, followed by Australia, the UK, and Mexico. Functional and self-reported NFt understanding were weakly correlated (rs=0.18, p<.0001). In adjusted analyses, functional NFt understanding was higher among women (p<.0001), ‘majority’ ethnic groups (p<0.0001), respondents with higher education levels (p<.0001), and those making efforts to consume less sodium, sugar, fat, calories or processed food (p<.0001). Similar correlates were significant for self-reported NFt and FOP label understanding, with some differences in diet behaviour correlates between self-reported and functional NFt understanding. Self-reported FOP label understanding was higher for interpretative labelling systems in Australia (Health Star Ratings) and the UK (Traffic Lights) compared with Mexico’s Guideline Daily Amounts (p<.0001). Mean self-reported FOP label understanding was higher than NFt understanding, with the exception of Mexico where self-reported NFt understanding was higher. Study 4 - Respondents from the US, Canada, and Australia reported significantly higher NFt use and awareness than respondents in Mexico and the UK. Mexican respondents reported the highest level of FOP label awareness, followed by the UK and Australia, whereas UK respondents reported the highest FOP label use followed by Mexico and Australia. In countries with both NFt and FOP labelling systems, use and awareness was higher for NFts in Australia and Mexico, with UK respondents reporting higher FOP label than NFt use and awareness. Correlates of NFt and FOP label use were similar, with the exception of health literacy where NFt use was higher among respondents with ‘adequate literacy,’ but FOP use was lower among this group compared to those with a ‘high likelihood of limited literacy.’ Food processing knowledge, sex, and ethnicity were only significantly associated with NFt use. Conclusions: Cross-country differences in labelling outcomes provide an opportunity to examine differences in nutrition labelling policies across countries. Nutrition labels requiring greater numerical skills (i.e., NFt, GDA) were more difficult for consumers to understand than interpretive FOP labels, and mandatory labelling policies (NFt, GDA) had the highest levels of awareness. These findings highlight the importance of mandatory FOP labelling policies to maximize reach, particularly among consumers with lower literacy, nutrition knowledge, and education who reported using nutrition labels less. This study also provides further evidence for the use of functional measures of knowledge and label understanding for multi-country, population-based studies. Overall, these findings support the need for mandatory labelling policies and national health promotion efforts that are accessible to all populations to minimize nutrition-related health disparities.