People in Western countries generally

Transcription

1 Article Traffic lights and health claims: a comparative analysis of the nutrient profile of packaged foods available for sale in New Zealand supermarkets Abstract Objective: To assess the application of Multiple Traffic Light (MTL) nutrition labels and the Food Standards Australia New Zealand (FSANZ) Nutrient Profiling Standard Calculator (NPSC) to foods and their agreement in classification of foods. Methods: MTL and NPSC criteria were applied to selected packaged food products available in New Zealand (NZ) supermarkets in 2011: 157 breakfast cereals; 128 cereal bars; 40 pizzas; 69 sausages and hotdogs; and 13 burgers. A points-based system combined MTL criteria into an overall score to enable comparison of food classification by the two systems. Results: Based on NPSC criteria, 156 (38%) NZ packaged food products were eligible to carry a health claim and thus classified as healthy ; 251 were less healthy. Based on MTL criteria most healthy products (61%) would also receive one red light. Using the MTL score, only 14% of products were classified as healthy. Percentage agreement between the two systems was 73%. Conclusions: More than half of NZ packaged foods are less healthy according to criteria underpinning two front-of-pack labelling systems. A number of foods eligible to carry health claims would display at least one red traffic light if MTL were introduced. Key words: Nutrient profiling, nutrition labelling, multiple traffic lights, health claims Aust NZ J Public Health. 2013; 37: doi: / Simone C. Rosentreter, Helen Eyles, Cliona Ni Mhurchu National Institute for Health Innovation, School of Population Health, The University of Auckland, New Zealand People in Western countries generally consume excessive dietary energy, are largely sedentary, undertake insufficient exercise and rely increasingly on nutritionally poor pre-packaged snacks and meals. 1 In New Zealand (NZ), about 8,000 9,000 deaths a year have been attributed to poor diet. 2 Nutrition labels provide consumers with information about the nutrient content of packaged foods at the point of purchase. Front-of-pack nutrition labels aim to provide simple, at-a-glance nutrition information to help consumers make healthier food choices. Interpretive front-of-pack nutrition labels in particular are receiving increasing attention as a potentially important tool in a broad and multi-faceted approach to enable healthier population diets. 3,4 Empirical evidence for the impact of front-of-pack labelling systems on food purchases and consumption is lacking. 5 However, such labels are an appealing strategy to improve consumer food purchases since they provide guidance on the nutritional value of products without limiting consumer choice. In NZ, three main labelling systems are in current use: the Nutrition Information Panel (NIP), the Heart Foundation Tick programme, and the Food Industry-led Daily Intake Guide (DIG). 6 Furthermore, the NIP typically displayed on the back of packaging has been mandatory for all packaged foods sold in NZ and Australia since December Foods are required to provide information on levels of energy, total fat, saturated fat, protein, carbohydrate, sugars and sodium per serving, and per 100 g or 100 ml. In addition, if a claim is made on the product packaging about another nutrient, information on this nutrient must also be provided on the NIP. There are no mandatory front-of-pack labelling system requirements in NZ. However, the Heart Foundation Tick 7 and the Food Industry-endorsed DIG 8 labels are displayed on some products. In 2011, an independent review panel set up by the Australia and New Zealand Food Regulation Ministerial Council (ANZFRMC) recommended introduction of a multiple traffic light (MTL) front-of-pack labelling system. 9 The ANZFRMC supported development of an interpretive front-of-pack labelling system, and proposed a collaborative design process with the aim of reaching a consensus on a standard system by the end of MTL is an interpretative front-of-pack system where colours (red, amber, or green) are used to indicate about the amount of total fat, saturated fat, sugar, and salt present in a food. MTL labelling is the front-of-pack format developed and recommended in the UK by the Food Standards Agency (FSA). (Most food labelling work has now moved from the FSA to the UK Department of Environment, Food and Rural Affairs.) 10 The UK FSA advice regarding foods with red lights is it s fine to eat this food occasionally or as a treat, but think about how often you choose it and how much of it you eat, and Submitted: August 2012 Revision requested: October 2012 Accepted: January 2013 Correspondence to: Associate Professor Cliona Ni Mhurchu, National Institute for Health Innovation, School of Population Health, University of Auckland, Private Bag 92019, Auckland Mail Centre, Auckland 1142, New Zealand; c.nimhurchu@nihi.auckland.ac.nz 278 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2013 vol. 37 no. 3

2 Comparative analysis of front-of-pack nutrition labels there is evidence to suggest that consumers have a stronger aversion to red lights than to amber lights. 11,12 Over the past decade, Food Standards Australia New Zealand (FSANZ) has also been developing a Standard to determine whether foods are eligible to make health claims. Underpinning this Standard is the FSANZ Health Claims Nutrient Profiling Standard Calculator (NPSC), a nutrient profiling system that has undergone extensive development over the past decade and been tested on more than 10,000 NZ and Australian food products. 13 It is possible that front-of-pack labels such as MTL and health claims could coexist in the retail setting in NZ and Australia at some point in the future and thus consumers may be faced with using these two systems together to interpret the healthiness of a packaged food. Therefore, research on the agreement of these systems in classifying foods as healthy and less healthy is much needed. This study aimed to assess the combined application of MTL labels and the FSANZ NPSC criteria to NZ packaged foods, and evaluate agreement in classification between the two front-ofpack labelling systems. Methods Food and nutrient data for selected food products Food and nutrient data were sourced from the Nutritrack supermarket database, a University of Auckland branded food and nutrient database containing package and nutrient information for the majority of processed foods for sale in NZ supermarkets (n>6000). 14 Nutritrack data are collected from the mandatory product NIP on food packages and are updated annually (2011 data were used for this analysis). Source and selection of food products Research by the UK FSA found that consumers thought front-ofpack labelling would be most useful on composite and processed foods. 10 Therefore, the packaged food product categories selected for this analysis were chosen from the categories recommended for MTL labelling by the FSA. In addition, to be selected, food categories also had to be commonly consumed in NZ and as such have relevant data available for analysis in the NZ Nutritrack database. Selected food categories were: breakfast cereals (n=157), cereal bars (n=128), sausages and hotdogs (n=69), burgers (n=13), and pizzas (n=40). Application of UK FSA MTL criteria MTL nutrition criteria (Table 1) were applied to products based on their nutrient contents as recorded in the Nutritrack food database. These criteria were based on those recommended by the UK FSA but adapted for the Australian/NZ environment. 10,15 The major difference was that the criteria for added sugars was not taken into account because data on added sugars is not readily available on the NIP in NZ. A MTL aggregate score, previously developed and published 15,16 to assess the overall healthiness of foods was also applied to products: one point was assigned for every green light, two for every amber light, and three for every red light, giving a total possible score of between four and 12. Foods were considered healthy if the total score based on the level of four key nutrients was less than seven. 15 Application of the FSANZ NPSC Criteria The NPSC system is a continuous points-based model where points are allocated based on the nutritional composition per 100 g of food. 13 First, baseline points are allocated based on energy, saturated fat, total sugar and sodium content. Then modifying points are awarded based on the percentage of fruit and vegetables and fibre content in the product. Some foods (depending on baseline points and fruit and vegetable content) can score further modifying points from protein. A final score is calculated by subtracting the modifying points from the baseline points. Three categories or types of foods are defined: (1) beverages; (2) all other foods; and (3) oils, oil spreads, and high calcium cheese; each with a different cut-off point for determining whether a food is eligible to carry a health claim. All foods selected for this analysis fell into category 2 (all other foods), which has a threshold of 3 points. Points according to the NPSC model (version 1.2.7) were allocated to selected foods based on their nutrient composition per 100 g. Energy, saturated fat, total sugar, protein, and sodium values were available for all foods. Fruit and vegetable points are assigned for foods if they have at least 25% concentrated (or dried) or more than 40% non-concentrated fruit and vegetable content. No foods in this analysis were eligible for any fruit and vegetable points. In addition, fibre data were not available for 10 breakfast cereals, 32 pizzas, all burgers (n=13), and all sausages (n=69) in the database, as it is not mandatory to include this information on the NIP. Fibre values based on the most similar product within the same category in the Nutritrack database were therefore assigned to products with missing values to build the NPSC score. Statistical analysis The proportion of products with nutrient levels that met the criteria for green, amber and red lights within each of the five food categories was calculated. Then the number of products that would carry zero, one, two, three, or four red lights were calculated. The number of products classified as healthy under the NPSC systems was also calculated. The MTL labelling system was compared with the Health Claims NPSC in two ways. First, a descriptive comparison of products that would carry red traffic lights and their classification as healthy or less healthy using the NPSC criteria was undertaken. Second, the percentage agreement between classification of products as healthy Table 1: The Multiple Traffic Light label nutrition criteria. 11,16 Green (/100 g) Amber (/100 g) Red (/100 g) Total fat 3 g g 20 g Saturated fat 1.5 g g 5 g Sugar (total sugars) 5 g g 15 g Sodium 120 mg g 600 mg 2013 vol. 37 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 279

3 Rosentreter, Eyles and Ni Mhurchu Article and less healthy using an MTL healthy threshold score of <7 and the NPSC criteria was calculated. Two assumptions were made for analyses: (1) That consumers interpret red traffic lights to mean less healthy 11,12 and use the colour-coded information about the four key nutrients to make a decision about the overall healthiness of the food; and (2) that consumers interpret the presence of a health claim to mean a food is healthy. Red lights were compared with healthy products under the NPSC criteria because a specific aim of these analyses was to determine where lack of agreement between the two nutrient profiling systems could cause consumer confusion. The cut-off of seven as healthy under MTL labels was previously used by Sacks et al. to assess the impact of such labels on food purchases in an on-line trial in Australia. 15 A sensitivity analysis used an MTL healthy threshold score of <8 because the use of <7 as the threshold by Sacks et al. was largely arbitrary. 15,16 Further, a reliability analysis using the kappa statistic was performed. Kappa is a conservative estimate of inter-rater agreement compared with percentage agreement. Kappa values range from -1.0 to 1.0 where 1.0 indicates perfect agreement. Standard interpretation of kappa is based on a scale of six categories from less than chance agreement (0) to almost perfect agreement (0.81 to 0.99). 17 All analyses were performed in SPSS for Mac statistical software package version 20 (SPSS Inc., Chicago, IL, USA) and/or Microsoft Excel 2010 (Microsoft Corporation, Redmond, WA, USA). The lead researcher (SR), who has a background in public health, selected the products and undertook all statistical analyses. She was supported by qualified nutritionists CNM and HE. Results Proportions of green, amber and red lights The analysis determining the number of red, amber, and green lights that would be present on selected foods demonstrated that if MTL were introduced in NZ, a large proportion of breakfast cereals would receive a red light for sugar (110/157, 70%), an amber light for total fat (55%), and green lights for saturated fat (66%) and sodium (52%) see Figure 1. Most cereal bars would receive a red light for sugar (95%), and amber lights for total fat (67%) and sodium (65%). Almost all sausages and hotdogs would receive a red light for sodium (97%); most would receive a red light for saturated fat (88%) and just over half would receive a red light for total fat (54%). Most burgers would receive an amber light for total fat and sodium, and a red light for saturated fat (92% for each). All pizzas would receive an amber light for total fat, and most would receive a green light for sugar (70%), and amber lights for sodium and saturated (90% and 85%, respectively). Number of red lights Overall, red lights would be most common for sausages (100% would receive at least one with 61% receiving two or three red lights), followed by cereal bars (98%), burgers (92%), breakfast cereals (71%) and pizzas (18%). Two or more red lights would again be most common for sausages (88%), followed by cereal Figure 1: Proportion of breakfast cereals with green, amber, and red traffic lights by nutrient (n=157 breakfast cereals). 42.7% 18.5% 14.6% Total fat 2.5% Sugar 54.8% 66.9% 66.2% 52.2% Saturated fat 4.5% Sodium 5.1% 29.3% 42.7% bars (47%), burgers (8%), and breakfast cereals (6%). No pizzas examined would receive more than one red light. Number of healthy products (eligible to carry a healthy claim) Overall, 156 NZ food products (39%) were classified as healthy and 251 (62%) were classified as less healthy using the NPSC criteria (Table 2). Breakfast cereals were most likely to be classified as healthy by NPSC (70%), followed by pizzas (38%) and cereal bars (25%). Red lights versus healthy The classification of products as healthy or less healthy using the NPSC system versus the number of red lights received is shown in Table 2. The majority of products classified as healthy would also receive one red light under the MTL system. Two breakfast cereals classified as healthy would receive two red lights. No products classified as healthy using the NPSC criteria would receive three or more red lights. MTL aggregate scores versus NPSC Health claim criteria Overall, a greater proportion of products were classified as healthy using the NPSC criteria compared with a MTL healthy threshold score of <7 (39% vs. 14%) see Table 3. Agreement between the MTL and NPSC scores was 73%, but varied by food category. Agreement was highest for sausages (99%) and lowest for breakfast cereals (59%). 280 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2013 vol. 37 no. 3

4 Comparative analysis of front-of-pack nutrition labels Table 2: Classification of products by Nutrient Profiling Standard Calculator (NPSC) criteria versus number of red traffic lights received using Multiple Traffic Light label criteria. NPSC healthy NPSC less healthy No red lights One red light 2 red lights* Breakfast cereals (n=157) 42 (27%) 65 (41%) 2 (1%) 48 (31%) Cereal bars (n=128) 3 (2%) 29 (23%) 0 96 (75%) Sausages (n=69) (100%) Burgers (n=13) (100%) Pizzas (n=40) 14 (35%) 1 (2%) 0 25 (63%) All (n=407) 59 (14%) 95 (23%) 2 (<1%) 251 (62%) *note: no products were rated as 3 or 4 With a threshold of <8 points (sensitivity analysis; Table 3) for the MTL system, the percentage of products classified as healthy increased from 14% to 34%. The agreement between the MTL and NPSC systems overall also increased from 73% to 78%. For breakfast cereals, cereal bars, and burgers agreement between the two systems increased, whereas for pizzas, and sausages agreement decreased. The kappa statistic (Table 3) indicated a fair level of agreement between the two systems using a traffic light aggregate score of <7 (k=0.35) and a moderate level of agreement (0.52) using the higher threshold. 18 Discussion To the best of our knowledge this is the first study to analyse the MTL profile of foods available for sale in NZ supermarkets, and to assess the agreement in classification of this front-of-pack labelling system with the proposed health claims eligibility assessment system. The FSANZ health claims standard is a well-developed, robust nutrient profiling system. 13,19 The purpose of the standard is to determine eligibility of foods to carry health claims, whereas the purpose of MTL is to guide consumer food choices based on the amounts of key nutrients present in a product. Assuming that consumers interpret red lights on MTL labels to mean less healthy, and health claims to mean healthy, 11,12 then agreement of the health claims and MTL systems in classifying foods is important to ensure that the consumer is presented with a consistent overall message regarding the healthiness of a food. There was good percentage agreement and fair inter-rater agreement between the two systems (73% and k=0.35, respectively) using an aggregate MTL score ranging from four to 12, and a cut-off of seven for healthy. Moreover, agreement increased with a cut-off of eight for the aggregate MTL score (78% and k=0.52, respectively). Less than perfect agreement between models highlights the differences in how the systems were built. The NPSC score incorporates modifying points for positive nutrients such as protein, fibre, and fruit and vegetable content, meaning a product could Table 3: Percentage agreement between the Multiple Traffic Light healthy threshold score of <7 or <8 and Nutrient Profiling Standard Calculator (NPSC) classification. Both healthy Both unhealthy TL healthy (<7), NPSC unhealthy NPSC healthy, TL unhealthy Overall agreement kappa (95%CI) Breakfast cereals 47 (30%) 45 (29%) 3 (2%) 62 (39%) 92 (59%) 0.27* (0.17, 0.38) (n=157) Cereal bars (n=128) 2 (2%) 94 (73%) 2 (2%) 30 (23%) 75 (96%) 0.06 (-0.06, 0.18) Sausages (n=69) 0 68 (99%) 8 (1%) 0 99 (68%) - Burgers (n=13) 0 12 (92%) 0 1 (8%) 12 (92%) - Pizzas (n=40) 3 (7.5%) 25 (62.5%) 0 12 (30%) 28 (70%) 0.24** (0.01, 0.47) All (n=407) 52 (13%) 244 (60%) 6 (1%) 105 (26%) 296 (73%) 0.35* (0.27, 0.43) Both healthy Both unhealthy TL healthy (<8), FSANZ unhealthy FSANZ healthy, TL unhealthy Overall agreement kappa (95%CI) Breakfast cereals 78 (50%) 31 (20%) 17 (11%) 31 (20%) 109 (69%) 0.33* (0.18, 0.49) (n=157) Cereal bars (n=128) 11 (9%) 90 (70%) 6 (5%) 21 (16%) 101 (79%) 0.33* (0.14, 0.52) Sausages (n=69) 0 67 (97%) 2 (3%) 0 67 (97%) - Burgers (n=13) 1 (8%) 12 (92%) (100%) - Pizzas (n=40) 12 (30%) 14 (35%) 11 (27.5%) 3 (7.5%) 26 (65%) 0.33** (0.06, 0.59) All (n=407) 102 (25%) 214 (53%) 36 (9%) 55 (14%) 316 (78%) 0.52* (0.43, 0.60) *p<0.001, **p< vol. 37 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 281

5 Rosentreter, Eyles and Ni Mhurchu Article be unhealthy for some nutrients but redeemed by others. What is perhaps most relevant in the real-world setting is the comparison between the number of red lights received by products and their classification using the NPSC health claims criteria, as products with health claims and red lights are potentially confusing for consumers. Our analyses indicated that out of the four food categories assessed, such inconsistencies were most likely for breakfast cereals (42% of healthy products would carry at least one red light), and least likely for sausages, pizzas, and burgers (0 to 1% of products healthy products would carry one more red lights). A summary indicator, such as an overall traffic light colour, would make it easier for consumers to determine whether foods were considered healthy, and to determine agreement with the NPSC system. Coexistence of the two systems could be viewed as a better representation of a food s nutrient profile however by offering whole food (in the form of health claims) and key nutrient specific (traffic light) information. However, consideration should be given to potential discrepancies between current and future nutrient profiling systems that may co-exist in the marketplace. As the NPSC system will underpin the ability of products to carry health claims in NZ and Australia, the new interpretive front-of-pack label recommended by the independent review panels should not conflict with this system. A core nutrient profiling system including common elements such as standard scoring, nutrients and food categories would alleviate confusion for consumers and other stakeholders. 20 Two previous studies described the traffic light nutrient profile of products available for sale in Australian supermarkets. 21,22 The first found that of 1,070 snack foods analysed, less than 1% received four green lights, 21 and the second found that few yoghurts and dairy drinks (1-2%) were healthy based on the same criteria despite many claiming health or nutritional benefits. 22 The second study also found that based on the Ofcom nutrient profiling system (the system on which NPSC is based), 21% and 28% of products were healthy in 2005 and 2008, respectively. In contrast, this study found that based on NPSC 39% of products were healthy. Similar to the current findings, previous research has found good percentage agreement between the MTL and NPSC schemes; using a definition of three or more green lights and no red lights for healthy, percentage agreement was 88% for bread, 90% for snacks and desserts, 78% for dairy, and 63% for processed meat and fish, although only 13% for vegetable oil spreads. 23 The MTL system is simple to use and our analysis found it was largely consistent with the NPSC. Therefore, it may be useful as brief interpretive tool to assess whether or not a product is broadly healthy. However, further research is needed to ensure the MTL system and points-based aggregate threshold agrees with the NPSC across a wider range of food categories. Our analysis indicates a threshold of eight may be more appropriate than a threshold of seven to indicate healthy. However, this threshold needs to be validated with a larger number of food groups. Further research is also needed to determine how consumers perceive the healthiness of products using MTLs and whether they would use labels to guide their purchasing. A previous UK study used a willingness-to-pay survey to assess consumers relative tolerance for different trafficlight labelled nutrients. 12 Consumers were willing to pay more for a change from red to green labels, compared with a change from amber to green, suggesting an aversion to red labels or desire for a larger change (red to green). Only two studies using sales data have been performed. 15,16 Neither found any correlation between application of MTL to products and change in sales. Both, however, examined a small number of products (n=18 in the UK and n=53 in Australia) and the timeframe assessed was arguably shorter (four and 10 weeks respectively) than required to see evidence of behaviour change. Strengths Our findings demonstrate that most packaged foods available for sale in NZ would be classified as less healthy and/or carry at least one red light. However, there was generally a range of healthy and less healthy products within the five categories analysed. If consumers are driven to purchase healthy options, it may stimulate manufacturers to reformulate foods to remain competitive. Further, manufacturers may decide to pre-emptively reformulate products in order to receive more favourable labels for their products. A second strength of this analysis is the assessment of the agreement of MTL with a credible nutrient profile model as a basis for comparison. While the NPSC and MTL systems differ in their purpose and construction, combined they would give greater front-of-pack information at point of purchase, by permitting health claims and providing standardised nutrient information. The agreement between the two systems appears to be acceptable given the systems are intended to fulfil different purposes; however, the MTL system is based on the absence of high levels of negative nutrients and does not consider the presence of positive nutrients. Limitations A limitation of this study is that the nutrient profiles described for products were not linked to sales data; therefore it was not possible to determine the volumes of products typically purchased by consumers. Knowledge of sales data would enable precise estimation of consumer exposure to volumes of healthy and less healthy foods and thus a better indication of likely impact of frontof-pack labels on consumer food choices and overall diets. A further limitation is that consumer perception of healthiness of products based on the MTL system is unclear. However, it is known the MTL system is preferred to and understood better than other front-of-pack systems, particularly among low-income and priority ethnic groups. 24,25 To date, most research has been limited to simple artificial choice situations and hasn t looked at real-life scenarios. Future research should focus on the choices consumers make in complex situations, such as when two products display labels with red lights for different nutrients or where one product carries two red and two green lights and another carries four amber lights. 282 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2013 vol. 37 no. 3

6 Comparative analysis of front-of-pack nutrition labels Conclusion There is considerable room to improve the nutritional profile of many foods available for sale in NZ supermarkets. Implementation of front-of-pack labelling systems such as MTL and health claims that provide interpretative information on the nutritional profile of foods could provide impetus for healthier consumer choices and industry reformulation of products. The MTL labels and NPSC Health Claims criteria applied to the same products across five packaged food categories are broadly consistent, although some foods eligible to carry health claims could display one or more red traffic lights. This was the case for more than 40% of breakfast cereals, and more than 20% of cereal bars. However, when combined, the two systems potentially offer a better representation of a food s nutrient profile by providing whole food (in the form of health claims) and nutrient specific (traffic light) information. Consideration should be given to potential discrepancies between current and future nutrient profiling and front-of-pack labelling systems that may co-exist in the marketplace. Acknowledgements Simone Rosentreter is now an employee of Network Communication, a communications company with a number of food industry clients. All research and analyses performed were independent of this role and preceded her employment. Cliona Ni Mhurchu is a member of a Front-of-Pack Labelling Expert Advisory Group convened by the New Zealand Government (Ministry for Primary Industries) in All research and analyses performed were independent of this role. Helen Eyles holds a Heart Foundation of New Zealand research fellowship (Grant 1463). Cliona Ni Mhurchu holds the Heart Foundation of New Zealand Senior Fellowship (Grant 1380). References 1. University of Otago and Ministry of Health. A Focus on Nutrition: Key Findings of the 2008/09 New Zealand Adult Nutrition Survey. Wellington (NZ): Government of New Zealand; Ministry of Health and University of Auckland. Nutrition and the Burden of Disease: New Zealand Wellington (NZ): Government of New Zealand; Draper AK, Adamson AJ, Clegg S, Malam S, Rigg M, Duncan S. Front-of-pack nutrition labelling: are multiple formats a problem for consumers? Eur J Public Health Dec 2. PubMed PMID: Food Standards Agency. Citizens Forum Findings on Nutrition Labelling. London (UK): FSA; 2010 [cited 2012 May 15]. Available from: food.gov.uk/news/newsarchive/2010/jan/citizenforumfop 5. Vyth EL, Steenhuis IHM, Brandt HE, Roodenburg AJ, Brug J, Seidell JC. Methodological quality of front-of-pack labelling studies: a review plus identification of research challenges. Nutr Rev. 2012;70(12); Ni Mhurchu C, Gorton D. Nutrition labels and claims in New Zealand and Australia: A review of use and understanding. Aust N Z J Public Health. 2007;31(2): Heart Foundation. Heart Foundation Tick. Auckland (NZ): Heart Foundation of New Zealand; 2012 [cited 2012 Jun 10]. Available from: heartfoundation.org.nz/healthy-living/healthy-eating/heart-foundation-tick 8. Australia Food and Grocery Council. Daily Intake Guide Labelling. Canberra (AUST): AFGC; 2010 [cited 2012 Oct 27]. Available from: org.au/daily-intake.html 9. Blewett N, Goddard N, Pettigrew S, Reynolds C, Yeatman H. Labelling Logic - Review of Food Labelling Law and Policy. Canberra (AUST): Commonwealth Department of Health and Ageing; United Kingdom Food Standards Agency. Front-of-pack Traffic Light Signpost Labelling Technical Guidance. Issue 2. London (UK): FSA; United Kingdom Food Standards Agency. Signposting and Traffic Light Labelling. Belfast (UK): FSA Northern Ireland; 2007 [cited 2012 Nov 2]. Available from: niyoungpeople/survivorform/bestreadbefore/signposting#.ujmpjkisrfi 12. Balcombe K, Fraser I, Falco SD. Traffic lights and food choice: a choice experiment examining the relationship between nutritional food labels and price. Food Policy. 2010;35(3): Food Standards Australia New Zealand. Health Claims Nutrient Profilling Calculator. Canberra (AUST): FSANZ; National Institute for Health Innovation. Nutritrack: Reformulation of Processed Foods to Promote Health. Auckland (NZ): NIHI; 2011 [cited 2012 Nov 5]. Available from: Sacks G, Tikellis K, Millar L, Swinburn B. Impact of traffic-light nutrition information on online food purchases in Australia. Aust N Z J Public Health. 2011;35(2): Sacks G, Rayner M, Swinburn B. Impact of front-of-pack traffic-light nutrition labelling on consumer food purchases in the UK. Health Promot Int. 2009;24(4): Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med. 2005;37(5): Landis J, Koch G. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1): Coalition on Food Advertising to Children. Submission to the Australian Communications and Media Authority on the Draft Children s Television Standards Canberra (AUST): Australian Communications and Media Authority; 2008 October. 20. Sacks G, Rayner M, Stockley L, Scarborough P, Snowdon W, Swinburn B. Applications of nutrient profiling: potential role in diet-related chronic disease prevention and the feasiblity of a core nutrient-profiling system. Eur J Clin Nutr. 2011;65(3): Walker K, Woods J, Rickard C, Wong C. Product variety in Australian snacks and drinks: how can the consumer make a healthy choice? Public Health Nutr. 2007;11(10): Walker K, Woods J, Ross J, Hechtman R. Yoghurt and dairy snacks presented for sale to an Australian consumer: are they becoming less healthy? Public Health Nut.r 2009;13(9): Truong Q. An Evaluation of Food Classification in Nutrient Profiling Systems for Use in New Zealand. Palmerston North (NZ): Massey University; Gorton D, Ni Mhurchu C, Mei-hua C, Dixon R. Nutrition labels: a survey of use, understanding and preferences among ethnically diverse shoppers in New Zealand. Public Health Nutr. 2009;12(9): Signal L, Lanumata T, Robinson JA, Tavila A, Wilton J, Mhurchu CN. Perceptions of New Zealand nutrition labels by Maori, Pacific and low-income shoppers. Public Health Nutr. 2008;11(7): vol. 37 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 283