Do health-related claims on food labels increase purchasing and/or consumption behaviour and intentions?
How much do health-related claims on food labels increase purchasing behaviour and intentions?
How does this effect differ according to the type of health health-related claim? - E.g. do health claims have a greater impact than nutrition claims?
We will search MEDLINE, PsycINFO, EMBASE, CAB abstracts, Business Source Complete, and Web of Science/Science Citation Index & Social Science Citation Index. The search strategy will only include terms relating to or describing the intervention. To be eligible for inclusion articles must be written in English and published in a peer review journal. No date restrictions will be placed on the search terms.
Types of study to be included
An article will be included if it:
- Is a controlled experiment that examines the effect of health-related claims on purchasing or consumption behaviour.
- Collects longitudinal individual level data - specifically panel data investigating whether the presence of a health-related claim will increase the likelihood of the participant purchasing or consuming that product.
- Collects population level data – specifically studies that examine population-level sales data before and after the implementation of a health-related claim will also be included.
An article will be excluded if it:
- Measures the effect of health-related claims on children’s behaviour
- A control group or an appropriate comparator cannot be identified within the study.
- Examines the effect of: negative health-related claims or health warnings e.g. ‘high in fat’, nutrient specific systems such as %GDAs and traffic light labelling of nutrients, or summary indicator systems such as 1-100 ratings of healthiness.
- The health-related claim is presented in a food service setting, such as a restaurant or in a menu, rather than presented in a retail setting or scenario like a supermarket.
- The health-related claim is not delivered in a written format; studies where the health-related claim was delivered orally or via moving image will not be included.
- If a symbolic claim was tested then its criteria for use must be publicly available. If a new, proposed, or relatively unknown symbol is being tested then an explanation of the symbol should be included in the study instructions (i.e. prior to the exposure to the symbol).
- Measures the understanding, preference and/or acceptability of health-related claims without assessing the overall impact on purchasing or consumption behaviour.
- It does not present data on the effect of health-related claims on purchase or consumption behaviour, reviews, editorials etc. will not be included.
Condition or domain being studied
Adult food purchasing and consumption choices.
Adults. Purchases of baby foods and foods for specific nutritional uses will be excluded.
Definitions of claims are detailed below. Only these explicitly health-related claims shall be considered in this review. Implicit claims are claims that convey factors, feelings, or actions associated with health, for example a picture of a person running, a heart shaped logo (without underlying criteria). These types of Implicit claims will be considered as contextual factors and thus not within the remit of this review. A health claim will be defined as ‘any claim which states, suggests or implies that a relationship exists between a food category, a food or one of its constituents and health’ (1). For example, ‘good for your heart’. A nutrition claim will be defined as ‘any claim that states, suggests or implies that a food has particular beneficial nutritional properties due to the energy, nutrients or other substances it contains, contains in reduced or increased proportions or does not contain’ (1). The definitions and categorisations of health health-related claims are those proposed by the International Network for Food and Obesity/non-communicable disease Research, Monitoring and Action Support (INFORMAS) (2) which is based on the definitions for the different types of health-related claim proposed by the Codex Alimentarius Commission (Codex) (3). Health-related claims may be presented as text, a symbol or a combination of both.
Included studies must measure the impact of at least one of the following:
- Health symbols - A symbolic claim is defined as a pictorial, or combined pictorial and text, health-related claim for which there are criteria underpinning its use. For example, foods must undergo pH-telemetry testing in order to be eligible to carry the “Toothfriendly” logo.
- Health claims: General health claim – claims concerning the general beneficial effects of the consumption of foods or their constituents on health. Nutrient function claims – claims that describe the physiological role of a nutrient or food in growth, development and functions of the body. Disease risk reduction claims - claims relating the consumption of a food or food constituent, in the context of the total diet, to the reduced risk of developing a disease or health-related condition. Nutrition claims: Nutrient content claims - claims that refer to the amount of a nutrient in a product (e.g. ‘low in fat’). Nutrient comparative claims - claims that compare the nutrient levels between two or more products (e.g. ‘lower in fat than …’). Health-related ingredient claims - claims that refer to substances other than nutrients or energy (e.g. ‘low calorie’).
(1) European Union (2006). Regulation No 1924/2006 on nutrition and health claims made on foods. Official Journal of the European Union L404/9.
(2) Rayner M, Wood A, Lawrence M, et al. (2013). Monitoring the health-related labelling of 458 foods and non-alcoholic beverages in retail settings. Obesity Reviews, 14(S1), 70–81.
(3) CODEX Alimentarious: Guidelines for use of nutrition and health claims CAC/GL 23-1997.
An article will be included if a comparator can be identified. A comparator may include:
(a) the same product with and without a health-related claim, or
(b) a control participant group – where participants can be randomised to an arm of a study where purchasing/consumption decisions are measured on products carrying health-related claims, or measured involving products where no such health-related claims are present.
Poor diet is a leading cause of ill health and non-communicable diseases. Food labelling may serve as a ‘nudge’ to prompt consumers to make healthier food purchasing decisions. Research has found that consumers who read the nutritional information on food labelling have a healthier diet, however for many this information is difficult to understand and/or interpret. Nutrition claims are statements relating to the nutritional quality of a food for example; ‘reduced sugar’, ‘zero calories’ etc. Health claims are statements that indicate that a relationship exists between a food, ingredient or constituent, and a health outcome. They may be textual or pictorial. Health and nutrition claims (henceforth referred to as ‘health-related claims’) may help consumers identify healthier products; however, the effectiveness of health-related claims as a tool to improve diet has been debated. Whilst some research found that health-related claims can improve diet other research has found little or no effect.
The primary outcome is the change in dietary or purchasing choices due to the presence of a health-related claim. This may be presented as individual level data or population level data.
The outcome measures are: - product selection: the number/percentage of participants that choose a product, amount consumed: the amount consumed, this may be expressed in grams, energy/calories or other nutrients, probability: often presented as odds ratios, refers to the odds of selecting a product.
Secondary outcomes are:
- willingness to buy and/or intention to purchase, often presented in 7-point Likert scales where 1 refers to little or no intention to buy the product and 7 indicating a high likelihood of buying the product.
Data extraction, (selection and coding)
References will be imported into Endnote. A single researcher will complete the first title screen to remove any duplicate references. Titles that are clearly unrelated to the primary or secondary research questions shall also be excluded at this stage.
The abstracts of the remaining references shall then be screened. Another researcher will check 10% of the included titles and abstracts of the resultant database in order to check for disagreements. If there are a significant proportion of disagreements then the inclusion and exclusion criteria will be reviewed and the abstract screening shall begin again. Any disagreements will be resolved via discussion with the other researcher with input from another member of the review team if necessary.
The data extraction will be assembled in Excel and the following data will be extracted:
- Paper details including the country and year that the study was conducted in.
- Claim details including the type of claim(s) tested and the product categories used in the study.
- Study population including how the population were recruited
- Study type and setting.
- Methods including description of the study procedure and analyses
- Outcome measure and unit of measure
- Results reported in the study and including the following headings: 'Does the study answer research question 1? (Y/N)', 'Does it support the hypothesis that health-related claims increase purchases? (Y/N)'', 'Does the study answer research question 2? (Y/N)', and 'How much do health-related claims effect purchases?'.
Risk of bias (quality) assessment
The Cochrane Risk of bias tool will be used to assess the study quality. The following questions will be used to assess the quality of the included studies and the findings will be summarised in the review:
- Selection bias: Were participants/products randomised to the health-related claim condition?
- Performance bias: Were participants blinded to the aims of the study? (e.g. health-related claim impact)
- Detection bias: Were participants aware of the study outcomes?
- Attrition bias: How much completed outcome data was obtained?
- Other bias: How were participants recruited? Were participants/products representative of the target population? How was the study funded? Were there any conflicts of interest reported?
Strategy for data synthesis
A two-step strategy will be used. First a sign test that indicates if the study answers the primary research question ‘do health-related claims increase purchasing intentions’? The second step will be to quantify these effects. From the initial scoping review we expect there to be much heterogeneity in the study design (for example, different food categories, and different outcome measures). Therefore a random effects meta-analysis will be performed which generates a weighted average effect size from the included studies.
Analysis of subgroups or subsets
If data are sufficient, there will be an analysis by sub-claim type (e.g. do reduction of disease risk claims have a greater effect than general health claims?) and/or by participant characters such as gender or socioeconomic status.
We aim to present the findings from this review in an article to be submitted to a peer-reviewed publication and a PhD/DPhil thesis.
Contact details for further information
British Heart Foundation Centre on Population Approaches for Non-Communicable Disease Prevention
Nuffield Department of Population Health
University of Oxford
Old Road Campus
Organisational affiliation of the review
Miss Asha Kaur, University of Oxford Dr Peter Scarborough, University of Oxford Professor Mike Rayner, University of Oxford
Anticipated or actual start date
02 November 2015
Anticipated completion date
02 September 2016
A Kaur is funded by the CLYMBOL project (European Union’s Seventh Framework Programme for research, technological development and demonstration, Contract n°311963).
P Scarborough and M Rayner are funded by the British Heart Foundation (grant no. 021/P&C/Core/2010/HPRG)
This review received no additional funding.
Conflicts of interest
Subject index terms status
Subject indexing assigned by CRD
Subject index terms
Choice Behavior; Diet; Humans
Any other information
We acknowledge that food choices are not a direct health measure. However, we feel that this systematic review does fit within the scope of PROSPERO as food labelling may help consumers make healthier food choices.
Stage of review
Date of registration in PROSPERO
03 August 2016
Date of publication of this revision
03 August 2016
Stage of review at time of this submission
Piloting of the study selection process
Formal screening of search results against eligibility criteria
Risk of bias (quality) assessment
PROSPERO This information has been provided by the named contact for this review. CRD has accepted this information in good faith and registered the review in PROSPERO. CRD bears no responsibility or liability for the content of this registration record, any associated files or external websites.