Substantiation of cosmetic claims

Date: 31st January 1996

A POSITION PAPER PREPARED ON BEHALF OF THE AUSTRALIAN SOCIETY OF COSMETIC CHEMISTS
FIRST EDITION PREPARED BY:
Dr Keith Hall
Proctor & Gamble
DATE: April 1989

SECOND EDITION PREPARED BY:
Dr Frances Scriven
S.M.A.R.T Research
(Assistance with clinical references from Dr Sandra Webb, Quintiles)
DATE: January 1996

SUMMARY


There is a number of reasons for ensuring that cosmetic claims can be substantiated. Obviously, there is a need to protect the consumer from untrue advertising. Excessive claims however are not only unfair to the consumer, but they ultimately lead to consumer disbelief and skepticism concerning cosmetic claims in general. This is unfair to other companies that are making claims which can be substantiated.

Within the cosmetic industry there is general agreement that the industry should show that it operates in a fair and ethical fashion.

There are four main groups of methods for evaluating product efficacy:

  • objective technical methods, which are generally instrumental in nature
  • clinical testing, in which a trained judge (usually medical expert) assesses product effects
  • sensory analysis by trained panels or by consumers consumer/market research i.e. beliefs and opinions
  • Each of these different groups gives useful data. As we go down the list above there is a trend from methods which are objective to those which are subjective. In other words, the methods towards the top of the list objectively measure the nature and magnitude of the product effect, whilst those towards the bottom quantify human perception and consumer preference. A combination of different methods provides the most convincing demonstration of product efficacy.
  • In all types of product testing/ substantiation, the trial should be designed to ensure that the results will occur in the real world and are representative and free from bias.
  • There is also the method of:
  • literature research which is often conducted before a formulation is prepared. For example, the literature may show many examples where fruit acids impart exfoliating properties and if used in your product should do the same.
  • A literature search, although providing evidence, does not guarantee that an ingredient will perform in exactly the same way in a new/different formulation.
  • Advertisements for cosmetic products vary enormously in the types of claims being made. In general, the wording of a claim should reflect the level of support data. If this is not the case, the claim is misleading.
  • Legislation exists at both Federal and State level to protect consumers against false or misleading product claims and there is an Advertising Code of Ethics used by the whole advertising industry. All advertising should be checked against these criteria.
  • The ASCC believes that all claims for cosmetic products serve a useful purpose in informing the consumer of a product's merits. Claims should, however, always be capable of substantiation. If this were not the case, claims would be unfair to the consumer and ultimately damaging to the cosmetic industry. Current legislative and self-regulatory controls are designed to ensure the truthfulness of claims and the industry should comply with both the requirements and spirit of these controls.

INTRODUCTION

The substantiation of claims for cosmetic products is an area of continuing and current interest within the cosmetic industry, and there is a number of reasons for ensuring that cosmetic claims can be substantiated. Obviously, there is a need to protect the consumer from untrue advertising. However excessive claims are not only unfair to the consumer; they ultimately lead to consumer disbelief and skepticism concerning cosmetic claims in general. This is unfair to other companies that are making claims which can be substantiated.

Within the cosmetic industry there is general agreement that the industry should show that it operates in a fair and ethical fashion. When the industry is not considered to operate in this fashion, government agencies may need to enforce legislation to maintain standards (as in the US). The use of excessive claims can thus force a choice between pre-emptive self policing and external regulation.

This paper discusses the various methods that may be used to substantiate cosmetic claims, and presents the ASCC position concerning such claims. Only cosmetics/toiletries are covered (eg moisturisers, skin cleansing and perfuming products and shampoos). Although many of the general principles will be similar, the question of substantiating claims for pharmaceuticals/therapeutics is addressed, only briefly.

METHODS


There are four main groups of methods for collecting data on product efficacy:

  • objective technical methods
  • clinical testing
  • sensory analysis by trained panels or by consumers
  • consumer/market research

Each of these groups gives useful data in its own way. They are not however, identical and interchangeable. Generally speaking, as we go down the list above there is a trend from methods which are objective instrumentally measured to those which are subjective and more consumer relevant. In other words, the methods towards the top of the list measure, objectively, the nature and magnitude of the product effect, whilst those towards the bottom quantify human perception and consumer preference. All are clearly useful pieces of information - both for developing superior products and for substantiating their efficacy. Since the different methods give different types of information, a combination of different methods provides the most convincing demonstration of product efficacy. A further consequence of the existence of these different methods is that we have to consider what types of claim can be validly substantiated by each type of method. This question will be addressed below, after considering the various groups of test methods in more detail.

General considerations


In all types of product testing/substantiation the trial should be designed to ensure that the results are representative and free from bias. Ideally, testing should be performed double blind i.e. neither the investigator nor the panellists should know the identity of the products being tested.

Data collected should be, and is usually assumed to be:

  • valid,
  • reliable, and
  • representative.

These terms are, however, very often misunderstood. Valid means that the result is likely to occur in the real world conditions of product use. For example, it may be that a moisturiser cream works very well in a laboratory test where temperature and humidity etc are well controlled, but the cream may not perform at all well in the real world, where the wind blows and weather conditions change rapidly, for example it may be 20%RH in an aeroplane and 90% RH when the person steps off in Bangkok. One way of improving the validity of any test is to incorporate real world conditions into the test as much as possible.

Reliable means that you are likely to get the same answer again if you were to do the test again. This does not mean the same answer from exactly the same human assessors, for example. Reliable means obtaining the same answer from a completely different set of human assessors. The best way of increasing reliability is to use random sampling procedures to select products or human assessors and to make the sample sufficiently large to be able to reasonably calculate how much the data varies. Sufficiently large is a deliberately vague term because experience with the particular type of data usually dictates sample size. Many researches feel that large differences will be seen in studies with between 50 and 100 humans, but small differences that are significant require more than 100.

Representative means that the sample has the same characteristics as the total population. Of course, you usually don't know the characteristics of the population and this is why you are doing the research. Statistical theory tells us that the only way to avoid this `catch 22' situation is to use random sampling procedures.

It should be clear that valid, reliable and representative are all interrelated and rely on random sampling.

There are two basic principles which govern random sampling. Firstly all products or humans in the population should have the same equal chance of being involved in testing. Secondly there should nothing systematic about the way the sample is selected. For this reason, random number tables are commonly used to identify subjects to be tested.

Some references on sampling are provided in the list of references.

Statistical methods of data analysis should be appropriate for the type of data collected. Statistics should be interpreted in the conventional manner, eg “significant” usually means “significant at the 95% confidence level”. It is always better to state the actual significance. This enables a decision to be made about whether 95% is really appropriate for the decision to be made, or whether, in fact, a lower level is appropriate. Misleading use of statistics is clearly unacceptable. Care should also be taken not to confuse “significant” with “substantial”. Significance indicates the risk associated with a decision, but substantiveness is a judgment about whether the difference is big enough to have meaning in the real world, even if it is significant.

Careful thought needs to go into the question of what to test a product against. For example, application of any moisturiser may be significantly better than using none, but is this very useful information? In any case, the choice of product for comparison may well limit the claims which can be made. Claims should never be based merely on the presence of an ingredient with a reported benefit, but should be based on data obtained using the actual formulation. This is because in the real world the actual formulation is used not the single ingredient.

1. Instrumental/technical methods

These tend to be the most controlled and objective means of product testing. They usually have the advantage of being precise, sensitive and well characterised. They generally only measure however, part of the total product effect ie the property which can be measured instrumentally. For this reason, it is common to use several such methods and to confirm results by a different type of study, eg market research. Another approach is to establish a significant correlation between the instrumental method and human perception/preference. This is particularly desirable since instrumental methods may be sensitive enough to detect changes which are too small for humans to notice, and vice versa, humans may be able to detect differences which cannot be quantified with instrumental measures.

Some technical methods are applicable in vivo, whilst others may only be used in vitro. In general it is preferable to use in vivo methods when possible since they give a more real world indication of product effect. Model systems (eg animal models or in vitro models) are useful for developing/screening products, but should not normally be used as the sole basis of claim substantiation. There are also ethical issues associated with the use of animals.

Some journals which regularly report articles demonstrating instrumental methods of testing are provided in the Appendix.

2. Clinical testing

While instrumental methods are objective in terms of their evaluation of one or more product attributes, clinical testing may be more subjective. The conditions under which the testing is carried out are, as in technical evaluation, carefully controlled but the assessment of the outcome may be made subjectively. The clinical assessment is made by a judge who is an expert in the field eg dermatologist, and who has been briefed beforehand as to which attribute(s) of the product/substrate is/are to be assessed. Of course, it is possible to include instrumental measurements in the clinical study as well. eg instrumental measures of metabolites appearing in the blood or urine.

Clinical testing is always in vivo by its very nature and may involve the use of either animals or humans as subjects. Because the assessments are made by allocating a score on an arbitrary scale or alternatively by ranking, the data may need to be analysed by non-parametric statistical methods. If clinical testing on humans involves the use of unregistered substances, then this is strictly controlled by the Therapeutic Goods Act.

Examples of clinical testing include skin irritation studies, using rabbits, where attributes such as erythema and oedema are rated on separate scales. Another example, this time involving humans as subjects, is the half-head salon study for evaluating shampoos, conditioners and hair treatments. Generally in this type of study products are assessed with respect to a number of attributes which may include for a shampoo, for example, consistency, texture and volume of lather, ability to clean hair, ability to rinse out of hair, wet/dry comb properties of hair, holding hair in place after drying, static generation in hair and so on.

A reference on clinical trials is given in the appendix.

3. Sensory analysis


There are essentially two types of sensory analysis: trained panels and consumer panels. All sensory analysis collects perceptions of product attributes, but in addition to this consumer panels are used to collect preferences and other opinions. A trained panel is a group of 10 to 20 people who act as highly sensitive experts, who measure product attributes in a machine like way. Trained panels are human instruments which need to be calibrated with references and make measurements of random representative samples of product to determine the differences between product. Consumer panels comprise random representative samples of target consumer populations and well defined products are used to determine both differences in perception and differences in acceptability.

Trained panels answer questions of the type `Is this product different to that one, with respect to a given attribute?' Consumer panels provide answers to questions of the type `what proportion of the population can detect a difference?' and `Is any difference perceived acceptable?'

There are numerous methods of data collection and these are described in the references provided. The reference section attached also shows several text books and a number of standards which provide background and protocols for trained panel work.

4. Consumer/market research

Consumer/market research for the purpose of making a claim usually aims at collecting data either to make a competitive claim, for example `four out of every 5 consumers prefer to use product A instead of product B', or to claim that consumers will perceive a specified benefit, for example `consumers who buy product A find it to be softer than product B'. In this sense the market research is the same as consumer sensory testing The Standards Association of Australia has established sensory analysis standards for paired comparison testing and these are given in the references. Market research is different in the sense that it often deals with behaviour and attitude data, such as the following hypothetical claims- 8 out of every 10 men take a shower in the morning, or - 9 out of every 10 women believe that animals should not be used for testing cosmetic products. It is important that consumers selected for such studies are representative of the target market and should have been selected according to some sort of sampling criteria. Similarly the products to be tested should be purchased to be representative of those in the market place and not supplied by the company wishing to make the claim. Because the field of market research is complex, most companies wishing to make a claim employ an external agency to increase the credibility of their claim. An American Society of Testing and Materials standard has been listed in the references, and covers some of the issues.

LEVELS OF CLAIM


Even a cursory glance through any women's magazine will show that advertisements for cosmetic products vary enormously in the types of claims being made. For example, some moisturisers claim to make you look better, some claim to make skin feel smoother and some claim to rejuvenate cells. Clearly, there is an a distinct difference in the level of claim support data that each of these claims implies. For example, “looking better” implies subjective data, whereas “smoother skin” implies the existence of instrumental clinical or sensory data. In general, the wording of a claim should reflect the level of support data. If this is not the case, the claim is misleading. The claim “rejuvenating cells” could not be used for a cosmetic as this type of claim relates to “modifying a physiological process” and can therefore only be used for products registered in the Australian Register of Therapeutic Goods.

LEGISLATIVE CONTROLS


The Therapeutic Goods Act 1989 ensures that any claim that a product “modifies a physiological process” renders it a therapeutic rather than a cosmetic (or indeed if it may be “taken to be” a therapeutic - e.g by being presented in a capsule or ampoule). That Act therefore defines a cosmetic in terms of what it does not do.

Part 2 of the Therapeutic Goods Regulations (under the TGA Act 1989) also proscribes certain representations for therapeutic goods (both drugs and medical “devices”) being made in advertisements and sets severe penalties for infringement.

The Trade Practices Act gives a definition of what a cosmetic product does do.

Trade Practices Act definition of a cosmetic product:

“intended for placement in contact with any external part of the human body, including:
a) the mucous membranes of the oral cavity; and
b) the teeth
with a view to:
c) altering the odours of the body
d) changing its appearance
e) cleansing it
f) maintaining it in good condition
g) perfuming it; or
h) protecting it”
The legislation is administered by the Trade Practices Commission and applies to cosmetic products manufactured or imported since 31 October 1993. Regulations for Consumer Products Information provide certain labelling requirements under the TP Act (e.g. full disclosure of names of ingredients in the labelling.

Guidelines for claims for cosmetics have also been developed by the National Coordinating Committee on Therapeutic Goods (TGA, Canberra). These are based on the corresponding Canadian Guidelines.

SELF REGULATORY CONTROLS


All advertising comes under the voluntary Advertising Code of Ethics, which includes the statement that:

“Advertisements must be truthful and shall not be misleading or deceptive”.

Advertisements which fail to comply will be prohibited by the media.

Even more stringent checks are placed on therapeutic claims, eg those for antiseptic skin cleansers and sunscreens. These are regulated by the Media Council of Australia's self-regulatory Therapeutic Advertising Code. More information may be obtained from the Proprietary Medicines Association of Australia.

THE ASCC POSITION


The ASCC believes that all claims for cosmetic products serve a useful purpose in informing the consumer of a product's merits. Claims should, however, always be capable of substantiation. If this were not the case, claims would be unfair to the consumer and ultimately damaging to the cosmetic industry. Current legislative and self-regulatory controls are designed to ensure the truthfulness of claims and the industry should comply with both the requirements and spirit of these controls.

APPENDIX


1 - JOURNALS - INSTRUMENTAL METHODS

The following journals routinely publish articles in which instrumental methods are reported.

Bioeng. Skin Cosmet. Toiletries
J. Acad. Dermatology
J. Soc. Cosmet. Chem
Int. J. Cosmet. Sci
Also see the book edited by A. Jarrett, The Physiology and Pathololphysiology of the Skin. 1980 Academic Press.

2 - REFERENCE ON CLINICAL TESTING

Spilker, B. 1991 Guide to Clinical Trials. Raven Press
3 - REFERENCES ON SENSORY TESTING AND MARKET RESEARCH METHODS

Cochran, W.G. 1977 Sampling Techniques. Third Edition Wiley
Green, T.E., Tull, D.S., Albaum, G. 1988 research for Marketing Decisions. Prentice Hall
O'Mahony, M. 1986 Sensory Evaluation of Food - Statistical Methods and Procedures. Marcel Dekker.
Piggott, J.R. 1986 Statistical Procedures in Food Research. Elsevier
Schaefer, E.E. 1979 ASTM Manual on Consumer Sensory Testing ASTM STP 682, American Society for Testing and Materials.
Standards Association of Australia, 1982 AS 2542.2.1 Paired Comparison Test (This standard has been recently revised.)