See discussions, stats, and author profiles for this publication at: [Link]

net/publication/279967206

How many foods in the UK carry health and nutrition claims, and are they
healthier than those that do not?

Article in Public Health Nutrition · July 2015

DOI: 10.1017/S1368980015002104 · Source: PubMed

CITATIONS READS

68 326

6 authors, including:

Asha Kaur Peter Scarborough

University of Oxford University of Oxford
45 PUBLICATIONS 2,165 CITATIONS 236 PUBLICATIONS 30,032 CITATIONS

SEE PROFILE SEE PROFILE

Anne Matthews Sarah Payne Riches

University of Oxford UK Health Security Agency
19 PUBLICATIONS 784 CITATIONS 16 PUBLICATIONS 410 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Asha Kaur on 15 October 2015.

The user has requested enhancement of the downloaded file.

 Public Health Nutrition: page 1 of 10 doi:10.1017/S1368980015002104

How many foods in the UK carry health and nutrition claims,

and are they healthier than those that do not?
Asha Kaur1, Peter Scarborough1, Anne Matthews1, Sarah Payne2, Anja Mizdrak1
and Mike Rayner1,*
1
British Heart Foundation Health Promotion Research Group, Centre on Population Approaches for
Non-Communicable Disease Prevention, Nuffield Department of Population Health, University of Oxford,
Old Road Campus, Oxford OX3 7LF, UK: 2University of Oxford School of Public Health, Oxford, UK

Submitted 4 November 2014: Final revision received 1 June 2015: Accepted 8 June 2015

Abstract
Objective: The present study aimed to measure the prevalence of different types of
health and nutrition claims on foods and non-alcoholic beverages in a UK sample
and to assess the nutritional quality of such products carrying health or nutrition
claims.
Design: A survey of health and nutrition claims on food packaging using a newly
defined taxonomy of claims and internationally agreed definitions of claim types.
Public Health Nutrition

Setting: A national UK food retailer: Tesco.

Subjects: Three hundred and eighty-two products randomly sampled from those
available through the retailer’s website.
Results: Of the products, 32 % (95 % CI 28, 37 %) carried either a health or nutrition
claim; 15 % (95 % CI 11, 18 %) of products carried at least one health claim and
29 % (95 % CI 25, 34 %) carried at least one nutrition claim. When adjusted for
product category, products carrying health claims tended to be lower in total fat
and saturated fat than those that did not, but there was no significant difference in
sugar or sodium levels. Products carrying health claims had slightly higher fibre
levels than products without. Results were similar for comparisons between
products that carry nutrition claims and those that do not.
Conclusions: Health and nutrition claims appear frequently on food and beverage
products in the UK. The nutrient profile of products carrying claims is marginally
healthier than for similar products without claims, suggesting that claims may have Keywords
some but limited informational value. The implication of these findings for guiding Health claims
policy is unclear; future research should investigate the ‘clinical relevance’ of these Nutrition claims
differences in nutritional quality. Food labelling

A poor diet is a major modifiable risk factor for many defined as, ‘any claim that states, suggests or implies that a
diseases including CVD and cancer(1–3). Poor diet and low food has particular beneficial nutritional properties due to
physical activity levels, combined, were estimated to the energy, nutrients or other substances it contains,
account for 14·3 % of disability-adjusted life years in the UK contains in reduced or increased proportions or does not
in 2010, exceeding even the impact of tobacco (11·8 % of contain’(9). The EU regulation(9) that contains these defi-
disability-adjusted life years)(4). Health and nutrition claims nitions separates nutrition claims into two sub-categories:
could potentially help consumers make healthier food and (i) ‘nutrient content claims’ that refer to the amount of a
beverage purchases, and there is some research to show nutrient in a product (e.g. ‘low in fat’); and (ii) ‘nutrient
that the presence of health and nutrition claims can influ- comparative claims’ that compare the nutrient levels
ence the perceived healthiness of products(5) and can lead between two or more products (e.g. ‘lower in fat than …’).
to an increase in the sales of products bearing claims(6–8). An additional sub-category of nutrition claims, i.e. ‘health-
In the UK, food labelling law is determined by the related ingredient claims’ that refer to substances other
European Union (EU). A health claim is defined by the EU than nutrients or energy, can also be distinguished.
as, ‘any claim which states, suggests or implies that a The EU regulation separates health claims into three
relationship exists between a food category, a food or one different sub-categories, here referred to as: (i) ‘nutrient
of its constituents and health’(9). A nutrition claim is and other function claims’, i.e. health claims that describe

*Corresponding author: Email [Link]@[Link]

© The Authors 2015. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://
[Link]/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original
work is properly cited.
 2 A Kaur et al.
the ‘role of a nutrient or other substance in the growth, not been clear whether those products that carry health or
development and the functions [both physiological and nutrition claims in the UK have a better nutritional profile
psychological] of the body’ (Article 13 claims); (ii) ‘reduc- than those that do not carry them and the current study
tion of disease risk claims’; and (iii) ‘general health claims’. therefore investigates that question.
Full definitions for all the different types of health and The research questions for the present study are
nutrition claims analysed in the present paper are provided therefore:
in the online supplementary material. These definitions are
1. Are health and nutrition claims easily identified and
those proposed by the International Network for Food and
categorised?
Obesity/non-communicable disease Research, Monitoring
2. What is the prevalence of health and nutrition claims
and Action Support (INFORMAS), which in turn are based
for packaged food and non-alcoholic beverage pro-
on the definitions for the different types of claim proposed
ducts sold by Tesco in the UK?
by the Codex Alimentarius Commission (Codex)(10). The
3. Do products that carry a health or nutrition claim have
online supplementary material also shows the equivalent
a better nutritional profile than products that do not
EU definitions where they exist.
carry such claims?
While some types of claim e.g. reduction of disease risk
or nutrient and other function claims are relatively easy to Questions 2 and 3 are the focus of the current paper.
identify and categorise, some types of health claim are not.
For example, claims such as ‘full of goodness’ or ‘to be
enjoyed as part of a healthy, active lifestyle’ would be
categorised by some as general health claims but not by Methods
others. The current study assesses how easily different
Public Health Nutrition

types of claims can be identified and categorised. Sampling

It has been suggested that health and nutrition claims The EU Food Information Regulation defines a ‘pre-
may lead consumers to overestimate the healthiness of packaged foodstuff’ as ‘any single item for presentation as
products(11). In order to reduce this possibility and also the such to the ultimate consumer and to mass caterers, con-
possibility of producers making claims about beneficial sisting of a foodstuff and the packaging into which it was
aspects of products while ‘hiding’ their non-beneficial put before being offered for sale, whether such packaging
properties (e.g. when making a ‘low fat’ claim for a pro- encloses the foodstuff completely or only partially, but in
duct that is high in salt), the EU regulation proposes that any case in such a way that the contents cannot be altered
products making health or nutrition claims should meet without opening or changing the packaging’(16). In our
minimal nutritional criteria. These criteria were to be study, we were concerned with pre-packaged foods and
defined using a nutrient profile model but to date the non-alcoholic beverages that are consumed by adults or
European Commission has not succeeded in developing children (but not babies, for whom health and nutrition
such a model. Nutrient profiling has been defined by the concerns are different). We were not concerned with
WHO as, ‘the science of classifying or ranking foods alcoholic beverages as the claims permitted on these
according to their nutritional composition for reasons types of products are regulated differently. We included,
related to preventing disease and promoting health’(12). separately, product items that were available in different
Australia and New Zealand have recently agreed a sized packages, on the basis that the packaging for the
nutrient profile model, the Nutrient Profiling Scoring same product in different sized packages may carry
Criterion (NPSC), to use for determining the eligibility of different health and nutrition claims.
products to make a health claim(13). The USA has long used The products were sampled from the home-shopping
a nutrient profile model for such a purpose. In the USA, website of Tesco, the retailer with the largest market share
health claims are permitted only if they do not exceed set in the UK at the time (November 2011). The following
thresholds for fat, saturated fat, cholesterol and sodium(14). types of products were excluded prior to sampling:
The proportion of food and beverage products in the
∙ unpackaged foods;
UK that carry health and nutrition claims has never been
∙ products that could not be identified as a pre-packaged
systematically assessed. An audit – simply involving visit-
food or beverage (e.g. there was no description
ing retail outlets and identifying pre-packaged products
identifying the item as a food, there was no picture,
with health claims – was carried out in the UK in 2003(15).
no nutritional information, etc.);
The audit examined 372 claims in relation to 182 products
∙ products that were marked as ‘product currently
and found that most health claims were ‘nutrient and other
unavailable’;
function claims’. Our study sampled packaged products
∙ alcoholic beverages; and
available through the home-shopping website of the lar-
∙ baby and infant foods and beverages.
gest retailer in the UK (Tesco) to obtain an accurate esti-
mate of the prevalence of health and nutrition claims This left a total of 13 700 packaged products, from which
within this defined population of foods. To date, it has also 400 were randomly selected.
 Health and nutrition claims in the UK 3
The primary research question was what is the pre- process), or a prominent web address promising nutri-
valence of health and nutrition claims for packaged food tional advice (because we thought that the presence of
and non-alcoholic beverage products sold by Tesco in the such a web address implies that the product is generally
UK? A power calculation was conducted to estimate the healthy). In line with the INFORMAS taxonomy the fol-
precision of the prevalence estimates for different num- lowing were not considered to be health or nutrition
bers of sampled products. After adjustment for a finite claims:
population and assuming a prevalence rate for health and
nutrition claims of 50 %, 400 products would produce a
∙ the terms ‘natural’, ‘organic’ and ‘Halal’;
confidence interval of ±5 %, which was deemed to be
∙ information on the absence of additives, preservatives,
colourings and flavourings;
precise enough for the purpose of the present study.
Three different Tesco branches were visited – two in
∙ allergy advice (e.g. ‘contains nuts’);
Oxford and one in Abingdon (Oxfordshire, UK) – between
∙ statements in relation to specific diets (e.g. dairy and/or
lactose free; wheat and/or gluten free; vegetarian or
November 2011 and March 2012 in order to purchase the
vegan);
products. The packaging was removed for data extraction.
Nutritional information for the content of energy, protein,
∙ storage advice (e.g. ‘stays fresh for longer’);
carbohydrate, sugars, fat, saturated fat, fibre and sodium
∙ reference to the presence of a ‘food or food group’ in
the product that does not state, suggest or imply a health
(g/mg/ml per 100 g/ml) was recorded from the Tesco
benefit (e.g. ‘contains chocolate’);
website, along with information on product name, brand
name, manufacturer, package size, price, etc.
∙ advertising in relation to sport (e.g. ‘official product of
the Olympics’) or to health concerns unrelated, or only
loosely related, to a healthier diet (e.g. ‘supporting
Product categorisation breast cancer research’); and
Public Health Nutrition

The products were categorised into fifteen product groups ∙ nutrition labelling, either back-of-pack or front-of-pack
using the FoodEx2 food classification system(17). FoodEx2 (e.g. traffic-light labelling for specific nutrient levels).
categorises products into twenty broad product groups,
Claims were included if they could be observed on any
such as grains and grain-based products, vegetable and
surface of the packaging which is observable to the pur-
vegetable products, milk and dairy products and sugar,
chaser. Claims were not included if they could only be
confectionery and water-based sweets. There are eight
observed once the packaging had been opened.
levels of sub-categorisation within these broad products
Health and nutrition claims were further categorised
groups, each with increasing specificity. The products
according to the nutrient or other food component to which
were also re-categorised at a later stage into five larger
they referred. Nutrient and other function claims were also
product groups for a regression analysis of the nutrient
categorised according to their health-related function or
content of products with or without claims. The new
structure as classified by the International Classification of
categories were: (i) beverages; (ii) fruit, vegetable and
Functioning, Disability and Health (ICF)(18) and reduction of
grain-based products; (iii) fish, meat and ready meals; (iv)
disease risk claims were categorised according to the
milk and dairy products (including dairy imitates); and (v)
International Classification of Diseases (ICD)(19).
miscellaneous products. The first four product groups
mapped readily onto FoodEx2 product categories; how-
ever, the miscellaneous category contained a variety of Inter-rater reliability for the detection and
products such as confectionery and seasonings which categorisation of claims
could not be re-categorised into the previous four cate- Inter-rater reliability was assessed for: (i) level of agree-
gories. Further details of the re-categorisation can be ment on whether the product packaging contained a
found in Table 3 below. health or nutrition claim or not; and (ii) level of agreement
on how claims were categorised using the INFORMAS
taxonomy. For both of these levels, inter-rater reliability
Claim detection and categorisation
was assessed using kappa scores generated with the sta-
Two researchers applied the INFORMAS taxonomy inde-
tistical software package Stata version 11. All disagree-
pendently and their decisions were compared (see below).
ments were then discussed between the study co-authors
Use of this ‘expert’ taxonomy in the study does not imply
to reach agreement as to whether the text or graphic
that consumers would identify and classify health and
should be considered as a health or nutrition claim, and
nutrition claims in the same way. Little is known about the
the claim category in which it should be included.
way consumers would classify claims into health and
nutrition claims and their different types if asked to do so.
Health or nutrition claims (as defined above) may take Comparison of the healthiness of products with or
the form of text (e.g. single words, phrases or sentences), without claims
brand names (e.g. ‘Healthy Choices’), a symbol, logo or First, the healthiness of products with or without claims
picture (e.g. representing a party of the body or a bodily was examined by comparing the difference in mean levels
 4 A Kaur et al.
of energy, protein, carbohydrates, sugars, fat, saturated fat, Research question 1: are health and nutrition
fibre and sodium (per 100 g, generally as sold) between claims easily identified and categorised?
products that carried claims and those that did not, for all Table 1 shows that there was good agreement between
products and for different food categories, using t tests. the two researchers over whether or not a food carried a
The nutrients (plus energy) were selected because their health claim or nutrition claim. Table 1 also shows that for
values were available through the Tesco website and are some health and nutrition claims there was more agree-
commonly found in nutrient declarations under EU law(16). ment over their type than for others. For example, there
Differences in nutrient levels shown by these t tests was very good agreement over the classification of
could be due to confounding by product group because nutrient comparative claims but less good agreement over
different product groups were found to have a different the classification of health-related ingredient claims.
prevalence of claims and have different nutritional Nutrient comparative claims and reduction of disease risk
properties. For this reason a regression analysis was claims had the highest percentage agreement (>99 %).
conducted where the results were adjusted for the five
broad product groups shown in Table 3. For each nutrient,
two regression models were run: (i) model 1 which did not Research question 2: what is the prevalence of
adjust for product group (and therefore produces results health and nutrition claims for packaged foods
that are comparable to the results of the t tests); and and non-alcoholic beverages sold by Tesco
(ii) model 2 which did adjust for product group. in the UK?
Second, the healthiness of foods with or without claims Table 2 shows that nutrition content claims were the most
was examined using a nutrient profile model: the NPSC. frequent type of claim (sixty-three products with a total of
eighty-seven claims), while reduction of disease risks
Foods that do not meet the model’s criteria would not be
claims were the least common claim type (three products
Public Health Nutrition

permitted to carry health claims in Australia and New

Zealand. The model was applied using Stata version 11 and a total of four claims). Table 2 also shows that in total,
and a logistic regression was conducted to determine the 32 % (95 % CI 28, 37 %) of products carried either a health
likelihood that a food carrying a claim would meet the or nutrition claim, with 15 % (95 % CI 11, 18 %) of products
model’s criteria. carrying at least one health claim and 29 % (95 % CI 25,
34 %) carrying at least one nutrition claim. On the 123
products that carried claims in our sample, we found a
total of 263 claims, an average of 2·1 claims per product
Results carrying a claim. Dairy products and beverages were the
product categories most likely to carry both health and
Missing data nutrition claims (Table 3).
Eighteen out the 400 products sampled were unavailable for
purchase; therefore 382 products were collected and ana-
lysed. Twenty-eight products did not have any nutritional Nutrients and ingredients referred to in health and
information. These products were included in the claim nutrition claims
extraction phase but were not included when assessing the Table 4 shows that of the 172 different nutrition claims the
healthiness of products carrying claims. An additional most common nutrient or ingredient referred to was fat
twenty-three products had incomplete nutritional informa- (e.g. ‘less than 2 % fat’) followed by fruit and vegetables
tion (one nutrient or more missing) and were included only (e.g. ‘contains one of your five-a-day’). Nutrition claims
in the analyses for which data were available. that referred to nutrients that, at high levels of intake, can

Table 1 Inter-rater reliability across the presence and categorisation of claims on a random sample of food and non-alcoholic beverage
products available through the website of a national UK food retailer (Tesco), November 2011–March 2012

Agreement type Agreement (%) Expected agreement (%) κ Prob > Z

Presence of claims on the 382 products
Health or nutrition claim 92 59 0·80 < 0·001
Nutrition claim 91 59 0·78 < 0·001
Health claim 91 75 0·65 < 0·001
Claim categorisation of the 274 claims identified by at least one researcher
Nutrition claim 82 51 0·65 < 0·001
Nutrient content claim 89 60 0·77 < 0·001
Nutrient comparative claim 99 89 0·87 < 0·001
Health-related ingredient claim 85 65 0·58 < 0·001
Health claim 84 60 0·60 < 0·001
General health claim 86 65 0·60 < 0·001
Nutrient or other function claim 97 93 0·66 < 0·001
Reduction of disease risk claim 99 97 0·66 < 0·001
 Health and nutrition claims in the UK 5

Table 2 Prevalence of health and nutrition claims on a random sample of food and non-alcoholic beverage products (n 382) available
through the website of a national UK food retailer (Tesco), November 2011–March 2012

No. of products % of products with 95 % CI for % of products

Claim type with claims No. of claims claims with claims
Nutrition claim 111 172 29·1 24·5, 33·6
Nutrient content claim 63 87 16·5 12·8, 20·2
Nutrient comparative claim 17 18 4·5 2·4, 6·5
Health-related ingredient claim 59 68 15·4 11·8, 19·1
Health claim 56 91 14·7 11·1, 18·2
General health claim 46 75 12·0 8·8, 15·3
Nutrient or other function claim 10 12 2·6 1·0, 4·2
Reduction of disease risk claim 3 4 0·8 0·0, 1·7
Health or nutrition claim 123 263 32·2 27·5, 36·9

have a damaging effect on health (e.g. fat, sugar or Of the products that did not carry any health or nutrition
sodium) were more frequent (55 %) than nutrition claims claims, 51 % (95 % CI 46, 56 %) met the NPSC model
that referred to nutrients with a beneficial impact such as criteria, while 62 % (95 % CI 49, 76 %) of the products
fibre and protein (38 %). Health claims frequently did not carrying health claims and 66 % (95 % CI 57, 75 %) of
refer to a specific nutrient or ingredient. In fact, 70 % of products carrying nutrition claims did so.
health claims did not do so (e.g. ‘healthy and delicious’). Table 7 shows that the main categories responsible for
the differences in fat and saturated fat between foods with
Public Health Nutrition

and without claims were ‘Fish, meat and ready meals’ and
Diseases and health-related functions and ‘Milk and dairy products’. It is products in the ‘Fruit,
structures referred to in health claims vegetable and grain-based products’ category which were
Throughout the entire project, there were four disease risk responsible for the differences in fibre content.
reduction claims identified. All four of these related to the
reduction of risk of CVD (Chapter I51·6 of ICD-10)(19). The
twelve nutrient and other function claims referred to a Discussion
broad range of health-related functions and structures
(Table 5). The present study found that health and nutrition claims
can be relatively easily identified and categorised. It also
found that nutrition claims are almost twice as common as
Research question 3: do products that carry health health claims (29 % of products carried at least one
or nutrition claims have a better nutritional profile nutrition claim compared with 15 % of products carrying at
than products that do not carry such claims? least one health claim). In addition it was found that the
Table 6 shows the difference in the amount of selected nutrient profile of products carrying claims tends to be
nutrients between products that carry claims and those healthier, in some respects (e.g. in their fat and saturated
that do not using regression analysis. The results of fat content), than that of products not carrying claims.
model 1 indicate that, across the board, products We have not yet done a systematic review of previous
with health claims were significantly lower in energy studies of the prevalence of health and nutrition claims, so
density (232·6 kJ/100 g), fat (6·7 g/100 g), saturated fat here we compare our results with three selected recent
(3·1 g/100 g) and sodium (152·8 mg/100 g). When adjusted studies(20–22). Our study found a higher prevalence of
for product group in model 2, the differences in fat and health claims than the EU-funded project ‘Food labelling to
saturated fat were smaller (5·7 g/100 g and 3·0 g/100 g, advance better education for life’ (FLABEL)(21) but a lower
respectively) but remained significant. The differences for prevalence than the surveys carried out in Australia(20) and
energy and sodium disappeared and a significant differ- Ireland(22).
ence in fibre content appeared. A similar pattern of The UK arm of the FLABEL study found that health
differences was observed for nutrition claims. claims and health logos are found on only a small
Products carrying health claims were 61 % more likely percentage of products (e.g. it found that only 4–6 % of
to meet the NPSC model criteria than products that did not products carried health claims, including symbolic health
carry such claims; a difference that was reduced after claims). However, the FLABEL study did not randomly
adjusting for product group (53 %). However these sample across all product categories and instead focused
differences were not statistically significant. Products only on five (breakfast cereals, soft drinks, biscuits,
carrying nutrition claims were significantly more (18 %) yoghurts and pre-packed fresh ready meals). Furthermore,
likely to pass the NPSC model criteria. This difference was in the FLABEL study the researchers were not required to
increased after adjusting for product group (22 %). record the wording of any claims for further analysis.
 Public Health Nutrition

6
Table 3 Prevalence of claims, by product category, on a random sample of food and non-alcoholic beverage products (n 382) available through the website of a national UK food retailer (Tesco),
November 2011–March 2012

No. of % of 95 % CI for % of No. of products % of products 95 % CI for % of No. of No. of

No. of products with products with products with with nutrition with nutrition products with health nutrition
Description FoodEx2 categories products health claims health claims health claims claims claims nutrition claims claims claims
Beverages Coffee, cocoa, tea and infusions 36 10 27·8 13·1, 42·4 19 52·8 36·5, 69·1 16 26
Fruit and vegetable juices and
nectars
Water and water-based beverages
Fruit, vegetable Fruit and fruit products 129 16 12·4 6·7, 18·1 36 27·9 20·2, 35·6 24 57
and grain- Vegetables and vegetable products
based products Starchy roots or tubers and products
thereof, sugar plants
Grains and grain-based product
Fish, meat and Fish, seafood, amphibians, reptiles 105 10 9·5 3·9, 15·1 17 16·2 9·1, 23·2 17 22
ready-meals and invertebrates
Meat and meat products
Composite dishes
Milk and dairy Milk and dairy products 39 12 30·8 16·3, 45·3 22 56·4 40·8, 72·0 20 38
products (incl. Products for non-standard diets, food
dairy imitates) imitates and food supplements or
fortifying agents
Miscellaneous Animal and vegetable fats and oils 73 8 11·0 3·8, 18·1 17 23·3 13·6, 33·0 14 30
foods Sugar, confectionery and water-
based sweet desserts
Seasoning, sauces and condiments
TOTAL 382 56 14·7 11·1, 18·3 111 29·1 24·6, 33·7 91 173

A Kaur et al.
 Health and nutrition claims in the UK 7
Table 4 Nutrients and ingredients referred to in health and nutrition claims on a random sample of food and non-alcoholic beverage products
(n 382) available through the website of a national UK food retailer (Tesco), November 2011–March 2012

Nutrition claims % of all nutrition claims Health claims % of all health claims
Nutrient
Energy 8 5 7 8
Protein 3 2 2 2
Sugar/sugars 12 7 0 0
Fat 45 26 0 0
Saturated fatty acids 11 6 1 1
Omega 3 fatty acids 2 1 1 1
Fibre 17 10 3 3
Beta-glucan 0 0 1 1
Sodium/salt 6 3 0 0
Cholesterol 1 1 1 1
Folic acid 1 1 0 0
Vitamin C 2 1 0 0
Vitamin D 2 1 0 0
Phosphorus 1 1 0 0
Calcium 7 4 2 2
Magnesium 1 1 0 0
Nitrite 1 1 0 0
Multiple nutrients 11 6 3 3
Ingredient
Caffeine 3 2 0 0
Fruit and vegetables 23 13 0 0
Wholegrain 7 4 3 3
Public Health Nutrition

Other ingredients 4 2 3 3
Unspecified nutrient or ingredient 4 2 64 70
TOTAL 172 100 91 100

Table 5 Health-related functions and structures referred to in nutrient or other function claims on a random sample of food and non-alcoholic
beverage products (n 382) available through the website of a national UK food retailer (Tesco), November 2011–March 2012

Health-related functions Health-related structures Example No.

Mental functions (B1) Structures of the nervous system (S1) 0
Sensory functions and pain (B2) The eye, ear and related structures (S2) 0
Voice and speech functions (B3) Structures involved in voice and speech ‘So sip on soya and help build stronger 3
(S3) teeth’
Functions of the cardiovascular, haematological, Structures of the cardiovascular, ‘Oats contain soluble oat fibre which is 7
immunological and respiratory systems (B4) immunological and respiratory systems proven to be good for your heart’
(S4)
Functions of the digestive, metabolic and Structures related to the digestive, ‘Dairy free soya drink is naturally kind 2
endocrine systems (B5) metabolism and endocrine systems on tummies’
(S5)
Genitourinary and reproductive functions (B6) Structures related to genitourinary and 0
reproductive system (S6)
Neuromusculoskeletal and movement-related Structures related to movement (S7) 0
functions (B7)
Functions of the skin and related structures (B8) Skin and related structures (S8) 0

The Australian survey(20) reported a higher prevalence health-related ingredient claims and general health claims
of health and nutrition claims than our study but did not were more difficult to identify and categorise. These
randomly sample across all food categories and instead problems have not been identified or quantified in
concentrated on three product groups known to carry a previous studies.
high number of health claims. Similarly a survey of health Even using a predefined and agreed approach and a
claims in Ireland(22) found that 18 % of products carried a clear taxonomy, there were still disagreements on
health claim and that 47 % of products carried a nutrition approximately one in ten products. There were also types
claim; however, the study used a convenience sample that of text and graphics that do not seem to have been
covered only a small number of product categories. anticipated by Codex or EU legislators. For example: does
In the present study we found that some types of health a website address that offers healthy eating advice
and nutrition claims were relatively easy to identify – constitute a health claim? We have taken it that it does.
particularly nutrient content claims, nutrient comparative The present study benefited from using a previously
claims and reduction of disease risk claims. However, developed taxonomy based on international (Codex)
 8 A Kaur et al.

Table 6 Difference in nutritional quality for products carrying health or nutrition claims compared with those that do not carry health or
nutrition claims for a random sample of food and non-alcoholic beverage products (n 382) available through the website of a national UK food
retailer (Tesco), November 2011–March 2012

Model 1 Model 2

Nutrient Diff. P Diff. P

Health claims
Energy (kJ/100 g) −232·6 0·02 −118·1 0·20
Protein (g/100 g) + 1·1 0·39 + 2·1 0·06
Carbohydrates (g/100 g) +1·5 0·70 +5·3 0·09
Sugars (g/100 g) −0·6 0·79 −0·4 0·86
Fat (g/100 g) −6·7 0·00** −5·7 0·00**
Saturated fat (g/100 g) −3·1 0·00** −3·0 0·00**
Fibre (g/100 g) +0·3 0·44 +0·7 0·05**
Sodium (mg/100 g) −152·8 0·02** −97·5 0·14
NPSC odds ratio† 1·61 0·118 1·53 0·182
Nutrition claims
Energy (kJ/100 g) −310·6 0·00** −235·4 0·00**
Protein (g/100 g) −0·9 0·37 +0·5 0·58
Carbohydrates (g/100 g) −3·8 0·20 −2·3 0·36
Sugars (g/100 g) −2·3 0·23 −3·2 0·08
Fat (g/100 g) −6·1 0·00** −5·3 0·00**
Saturated fat (g/100 g) −3·2 0·00** −3·3 0·00**
Fibre (g/100 g) +0·2 0·55 +0·5 0·07
Sodium (mg/100 g) −115·7 0·03** −56·2 0·30
NPSC odds ratio† 2·18 0·001** 2·22 0·002**
Public Health Nutrition

Model 1, no adjustment for food groups; model 2, adjustment for food groups; Diff., difference; NPSC, Nutrient Profiling Scoring Criterion; + denotes an
increase; − denotes a reduction.
** P < 0·05.
†These results were calculated through a logistic regression analysis and report the odds that a product carrying a claim passes the NPSC model before
(model 1) and after (model 2) adjusting for food groups.

Table 7 Difference in nutritional quality for products carrying health or nutrition claims compared with those that do not carry health or
nutrition claims, by product category, for a random sample of food and non-alcoholic beverage products (n 382) available through the website
of a national UK food retailer (Tesco), November 2011–March 2012

Category

Fruit, vegetable and

Beverages grain-based products Fish, meat and ready Milk and dairy products Miscellaneous
(n 36) (n 129) meals (n 105) (including imitates) (n 39) foods (n 73)

Diff. P Diff. P Diff. P Diff. P Diff. P

Health claims
Energy (kJ/100 g) +214·7 0·237 +239·2 0·165 −294·9 0·035** −519·7 0·005 −398·3 0·215
Protein (g/100 g) +2·0 0·244 +2·8 0·024** +0·6 0·801 −3·8 0·245 +10·0 0·025**
Carbohydrates (g/100 g) +7·6 0·284 +16·3 0·010** −1·9 0·591 −4·5 0·140 +2·7 0·817
Sugars (g/100 g) +5·4 0·192 +5·6 0·221 −0·3 0·636 −2·7 0·208 −14·3 0·115
Fat (g/100 g) +1·4 0·258 −2·8 0·332 −7·4 0·030** −10·2 0·014** −10·0 0·177
Saturated fat (g/100 g) +1·3 0·124 −1·2 0·320 −2·4 0·124 −6·3 0·018** −7·0 0·118
Fibre (g/100 g) +0·9 0·090 +2·4 0·001** −0·2 0·582 +0·3 0·112 −1·0 0·481
Sodium (mg/100 g) +120·9 0·121 −33·8 0·538 −215·0 0·059 −104·3 0·265 −252·8 0·441
Nutrition claims
Energy (kJ/100 g) +77·2 0·658 −68·7 0·592 −211·9 0·059 −692·0 0·000** −402·0 0·092
Protein (g/100 g) +0·9 0·604 +0·4 0·682 +2·2 0·219 −5·8 0·053 +3·5 0·303
Carbohydrates (g/100 g) +3·1 0·654 +2·7 0·567 −3·7 0·199 −3·3 0·250 −12·6 0·139
Sugars (g/100 g) +4·0 0·332 −1·9 0·577 −0·3 0·551 −0·8 0·700 −13·9 0·041**
Fat (g/100 g) +0·5 0·658 −3·9 0·065 −5·1 0·063 −14·4 0·000** −4·5 0·414
Saturated fat (g/100 g) +0·8 0·332 −2·5 0·005** −2·3 0·065 −9·7 0·000** −3·4 0·318
Fibre (g/100 g) +0·8 0·121 +1·5 0·009** −0·2 0·571 −0·1 0·637 −0·1 0·940
Sodium (mg/100 g) +92·2 0·232 −5·3 0·900 +20·7 0·818 −169·4 0·054 −222·6 0·367

Diff., difference; + denotes an increase; − denotes a reduction.

** P < 0·05.

definitions. This taxonomy has been developed by structures (ICF)(18) were used for the analysis of the
the INFORMAS project(10). In addition, internationally results.
recognised methods for categorising foods (FoodEx2)(17), Unlike most previous studies of the prevalence of health
diseases (ICD-10)(19) and health-related functions and and/or nutrition claims(20–22) or food labelling elements in
 Health and nutrition claims in the UK 9
general, the present study used a random sample of collected), ‘Reference to general, non-specific benefits of
products across the majority of food categories, from a the nutrient or food for overall good health or health-
defined ‘population’ of foods. While sampling a broad related well-being may only be made if accompanied by a
range of products is more costly and time-consuming than specific health claim included in the lists provided for in
sampling specific food categories, it gives a more com- Article 13 or 14’. It will be interesting to see whether the
plete picture of the prevalence of food labelling elements. prevalence of general health claims changes in the light of
However, it should be noted that the sampling methods the implementation of the legislation.
used for our study did not generate a representative
sample of all products available for purchase in the UK as
products were sampled from just one supermarket (albeit Conclusions
the one with the largest market share). While it has been
estimated that up to 90 % of food purchases in the UK are The present study has also shown that, in general,
made in supermarkets(23,24), it might be expected that products that carry health and nutrition claims have a
foods sold in other types of stores would have a higher slightly more favourable nutritional profile than those that
prevalence of health claims (e.g. health food shops) or a do not. The main differences were in fat and saturated fat.
lower prevalence (e.g. discount stores)(15). However, it is difficult to say whether these differences
Another potential weakness of the study is that the were ‘clinically relevant’. For example, products carrying
nutrient profile model used to assess the healthiness of health and nutrition claims had a fat content that was 6 g
products with and without claims – the NPSC – was devel- per 100 g less than products without claims but it is this
oped for use in Australia and New Zealand and the present enough to justify such claims? At this point we are not clear
study investigated products sold in the UK. However it how best to judge whether our results are ‘clinically rele-
Public Health Nutrition

should be noted that that NPSC model is a modified version vant’ or not. We think modelling the health effects of
of a nutrient profile model first developed in the UK(25,26). consuming products with and without claims could
The statistical power of our study was set to estimate the provide some answers but currently we do not have
prevalence of health and nutrition claims in the total enough data to parameterize a model.
population of 13 700 products with an accuracy of ±5 %, As noted above, the EU regulation(9) proposes that there
but the study is under-powered to estimate differences in should be a nutrient profile model which sets minimal
prevalence between food categories and under-powered quality criteria for the nutrition content of products bearing
to detect small differences in nutrient content between health or nutrition claims. Had the results revealed that
products that carry and those that do not carry claims. products bearing health or nutrition claims were less
Our study found what to some might seem to be a sur- healthy than products not bearing health or nutrition
prisingly high prevalence of health and nutrition claims on claims then they would have demonstrated a clear need
the packaging of foods and non-alcoholic beverages sold for further regulation to ensure a minimum nutritional
though a major retailer in the UK. This high prevalence, in quality for products that carry health or nutrition claims.
and of itself, suggests that health and nutrition claims are Our observation that products bearing health or nutrition
important ways of marketing such products to consumers. If claims are slightly healthier than foods not bearing claims
this marketing is effective, there may be important public might imply, on the one hand, that there is no need for
health implications to justify the regulation of claims. such a nutrient profile model or, on the other, that the
The current study also confirms that nutrition claims difference is so slight that a model is urgently required.
appear much more frequently than health claims. How- There is a need for further research that addresses the
ever, much more attention has been paid to the regulation impact of potential nutrient profile models to regulate
of health claims – and in particular the comparatively rare health and nutrition claims. Also, more work is required to
forms of health claims, i.e. nutrient and other function understand consumer perception and reactions to health
claims and disease risk reduction claims – than to the and nutrition claims to ascertain whether these claims are
regulation of nutrition claims. informative or misleading.
One of the reasons why we identified a high prevalence
of health claims was due to our inclusion of general health
claims such as ‘healthy’, ‘good for you’, ‘full of goodness’ Acknowledgements
and ‘consume responsibly as part of a healthy diet’, within
the category of health claims. Over 80 % of health claims in Acknowledgements: The authors thank Sophie Hieke of
our study were classified as general health claims the CLYMBOL project and Amanda Wood for their help in
(seventy-five claims on forty-six products). In the future, developing some of the concepts in this paper. Financial
such claims will only be permitted in the UK if accom- support: This research received no specific grant from any
panied by a more specific claim. The EU regulation on funding agency in the public, commercial or not-for-profit
health and nutrition claims specifies that, from 24 January sectors. A.K., P.S., [Link]., [Link]. and M.R. are funded by the
2013 (i.e. after the data for the present study were British Heart Foundation; S.P. received no specific funding
 10 A Kaur et al.
for this project. Conflict of interest: None. Authorship: results from FDA experimental data. J Public Policy Mark
A.K.: development of methods, data extraction, data ana- 18, 89–115.
12. World Health Organization (2015) Guiding Principles and
lysis, writing paper. P.S.: development of methods, writing Framework Manual for the Development or Adaptation of
paper. [Link].: development of methods, data collection Nutrient Profile Models (First Edition). Geneva: WHO
and extraction, feedback on paper. S.P.: development of (In the Press).
methods, data extraction, feedback on paper. [Link].: data 13. Food Standards Agency Australia New Zealand (2014)
Australia New Zealand Food Standards Code – Standard
analysis, feedback on paper. M.R.: development of 1.2.7 – Nutrition, Health and Related Claims. [Link]
methods, feedback on paper. Each author has seen and [Link]/Series/F2013L00054 (accessed July 2013).
approved the contents of the submitted manuscript. Ethics 14. US Food and Drug Administration (2013) Code of Federal
of human subject participation: Ethical approval was not Regulations Title 21 (21CFR101.14). Title 21 – Food and drugs;
Chapter I – Food and Drug Administration; Department of
required. Health and Human Services; Subchapter B – Food for Human
Consumption. [Link]
cfdocs/cfcfr/[Link]?fr=101.14 (accessed July 2013).
15. Foods Standards Agency (2003) Audit of the Range of Health
Supplementary material Claims Made for Foods on the UK Market in September 2003.
To view supplementary material for this article, please visit London: FSA.
16. European Union (2011) Regulation No 1169/2011 on the
[Link] provision of food information to consumers. Official
Journal of the European Union L304, 18–63.
17. European Food Safety Authority (2011) The Food Classifi-
cation and Description System FoodEx 2 (draft-revision 1).
References Supporting Publications 2011:215. Parma: European Food
Safety Authority; available at [Link]
1. World Cancer Research Fund/American Institute for Cancer en/supporting/pub/[Link]
Public Health Nutrition

Research (2007) Food, Nutrition, Physical Activity, and the 18. World Health Organization (2001) International Classifi-
Prevention of Cancer: A Global Perspective. Washington, cation of Functioning, Disability and Health: ICF. Geneva:
DC: AICR. WHO; available at [Link]
2. World Health Organization/Food and Agriculture Organi- [Link]
zation of the United Nations (2003) Diet, Nutrition, and the 19. World Health Organization (2008) ICD-10: International
Prevention of Chronic Diseases. Report of a Joint WHO/FAO Statistical Classification of Diseases and Related Health
Expert Consultation. WHO Technical Report Series no. 916. Problems (10th Revised Edition). New York: WHO;
Geneva: WHO. available at [Link]
3. World Health Organization (2002) The World Health Report 20. Hughes C, Wellard L, Lin J et al. (2013) Regulating health
2002: Reducing Risks, Promoting Healthy Life. claims on food labels using nutrient profiling: what will the
Geneva: WHO. proposed standard mean in the Australian supermarket.
4. Murray CJL, Richards M, Newton J et al. (2013) UK health Public Health Nutr 16, 2154–2161.
performance: findings of the Global Burden of Disease 21. Storcksdieck genannt Bonsmann S, Celemín L, Larrañaga A
Study 2010. Lancet 381, 997–1020. et al. (2010) Penetration of nutrition information on food
5. van Trijp HC & van der Lans IA (2007) Consumer percep- labels across the EU-27 plus Turkey. Eur J Clin Nutr 64,
tions of nutrition and health claims. Appetite 48, 305–324. 1379–1385.
6. Sterken E (2012) The big issue is ultra-processing. 22. Lalor F, Kennedy J, Flynn MA et al. (2010) A study of
Troublesome claims and descriptions (Letter). World Nutr 3, nutrition and health claims – a snapshot of what’s on the
326–327. Irish market. Public Health Nutr 13, 704–711.
7. Marquart L, Weimer K & Jacob B (2001) Solid science and 23. Euromonitor (2000) Consumer Lifestyles – UK. Integrated
effective marketing of health claims. Nutr Today 36, 107–111. Market Information System. London: Euromonitor.
8. Calfee J & Pappalardo J (1991) Public policy issues in health 24. Caraher M, Dixon P, Lang T et al. (1998) Access to healthy
claims for foods. J Public Policy Mark 10, 33–53. foods: part I. Barriers to accessing healthy foods: differ-
9. European Union (2006) Regulation No 1924/2006 on entials by gender, social class, income and mode of trans-
nutrition and health claims made on foods. Official Journal port. Health Educ J 57, 191–201.
of the European Union L404, 9–25. 25. Ofcom (2007) Television Advertising of Food and Drink
10. Rayner M, Wood A, Lawrence M et al. (2013) Monitoring the Products to Children. [Link]
health‐related labelling of foods and non‐alcoholic consultations/foodads_new/statement (accessed July 2013).
beverages in retail settings. Obes Rev 14, Suppl. 1, 26. Department of Health (2011) The nutrient profile model.
70–81. [Link]
11. Roe B, Levy A & Derby B (1999) The impact of health claims PublicationsPolicyAndGuidance/DH_123494 (accessed
on consumer search and product evaluation outcomes: July 2013).

View publication stats