Clinically Proven Cosmetics: What Evidence a Cosmetic Product Safety Report Files for Each Claim
Compartir
The short answer
Clinically proven means a study on human volunteers supports a specific stated effect, and the claim is only as strong as the study's design, endpoint and sample size. Dermatologist tested means a dermatologist was involved in testing the product, usually for tolerance rather than efficacy, and says nothing about the result unless the brand states it. The Cosmetic Product Safety Report does not generate these studies; it records that claims are consistent with the assessment, while the study reports themselves are filed in the Product Information File claims table.
Key takeaways
- Clinically proven is not a regulated phrase, so the question is always: proven what, measured how, on how many people, for how long.
- Dermatologist tested describes involvement, not outcome, and most such tests assess tolerance rather than performance.
- Instrumental studies measure a physical endpoint such as hydration or elasticity; perception studies record what volunteers felt, and the two must not be presented interchangeably.
- Sample size, duration, controls and blinding determine how much weight a study carries.
- The evidence lives in the Product Information File and must be producible on request, so a claim without a filed report is a claim without support.
- A supplier's study on an ingredient does not substantiate a claim for your finished product unless the level and formula match.
Clinically proven, dermatologist tested, independently tested, proven results: four phrases that suggest rigour and reveal almost nothing on their own. None is defined in Regulation (EC) No 1223/2009 or the UK regulation enforced through the Cosmetic Products Enforcement Regulations 2013, which means each is a claim judged by the six criteria in Regulation (EU) No 655/2013: truthful, evidenced, honest, fair and capable of informing a decision. The Advertising Standards Authority upholds complaints about them regularly, and almost always for the same reason, which is that the study behind the phrase did not measure what the phrase implied.
This article explains what each phrase can legitimately mean, the study types that support cosmetic claims and what each can and cannot prove, how sample size and design change the weight of evidence, where the reports are filed and who can ask for them, the difference between an ingredient study and a finished product study, how consumer perception results should be worded, a worked example of one serum's claims matched to its evidence, and a checklist for buying testing that will actually support the claim you want to make. It is written for brands commissioning studies and useful to anyone trying to judge one.
What can each phrase legitimately mean?
| Phrase | What it should mean | What it often means | What to state alongside |
|---|---|---|---|
| Clinically proven | A clinical study on human volunteers demonstrated a specific measured effect | A small in use study with self reported outcomes | The effect, the measure, the duration and the number of participants |
| Clinically tested | A study was carried out; the result is not asserted | The same as above with less commitment | What was tested and what was found |
| Dermatologist tested | A dermatologist supervised or conducted a test, usually of tolerance | Nothing about efficacy or the result | What was tested and the outcome |
| Dermatologically approved | A dermatologist reviewed and endorsed | An opinion rather than a study | The basis of the approval |
| Independently tested | An external laboratory conducted the study | In house testing described loosely | The laboratory and the protocol |
| Proven results | Anything or nothing | Usually a perception survey | The study type and endpoint |
| 9 out of 10 users agreed | A consumer perception study | Correctly used, this is honest | The number of participants and the exact question asked |
In practice
For any of these phrases, ask the three question test: proven what, measured how, on how many? A brand that can answer all three in one sentence has a study. A brand that cannot has a phrase.
The study types, and what each can prove
• Human repeat insult patch test. Tolerance and sensitisation potential, typically on a panel of 50 to 200. Supports dermatologically tested and claims about tolerance, not efficacy.
• In use tolerance study. Real world tolerance over weeks, often with dermatological assessment. Supports suitable for sensitive skin where results allow.
• Instrumental efficacy study. A measured physical endpoint on volunteers: corneometry for hydration, transepidermal water loss for barrier function, cutometry for elasticity, image analysis for wrinkle depth or pigmentation. Supports quantified performance claims.
• Controlled clinical study. Comparative, ideally with a vehicle control and blinded assessment. Supports the strongest claims and comparatives.
• Consumer perception study. What volunteers reported feeling or observing. Supports agreement style claims when clearly worded as perception.
• Ex vivo and in vitro study. Mechanism on skin models or cells. Supports carefully worded mechanistic claims, never a claim of an effect on a person.
• Ingredient literature. Published data on an ingredient at a level. Supports function claims where your level matches.
How do design and sample size change the weight?
Considerably, and this is where most weak claims hide. A study on twenty volunteers over two weeks with no control and self reported outcomes is a very different thing from a study on eighty volunteers over eight weeks with instrumental measurement and a vehicle control, yet both can be described as clinically tested. The elements that matter are the number of participants, the duration relative to the biological effect claimed, whether there was a control or comparison, whether assessors were blinded, whether the endpoint was measured or reported, and whether the statistics support the difference claimed. None of this requires a brand to run pharmaceutical grade trials, and cosmetic studies are rightly smaller and simpler than drug trials. It does require that the claim matches the study: a small perception study supports a perception claim, and only an instrumental study supports a percentage.
Where does the evidence live?
In the Product Information File. The regulation requires the file to contain proof of the effect claimed where justified by the nature of the claim, which in practice means the study report itself, not a summary, indexed in a claims table with a row for the claim, the study, its date and where it is held. Trading Standards can request it, marketplaces increasingly ask for it, and the ASA will expect it within days if a complaint is made. The Cosmetic Product Safety Report sits alongside: it does not generate efficacy evidence, but the assessor reviews draft claims against the assessment and will flag any claim that contradicts it, such as a performance claim relying on an ingredient at a level the formula does not contain. Our article on how a CPSR exposes greenwashing sets out the claims table in detail.
Why an ingredient study is not a product study
This is the commonest substantiation error in the category. A raw material supplier provides a study showing that its active, at five percent, increased a measured endpoint over eight weeks. A brand uses that active at one percent in a different base and claims the supplier's result. The claim fails, because the evidence does not describe the product: the level differs, the formula differs, penetration and stability differ, and the study was not conducted on what the customer will buy. Supplier studies are useful, and they legitimately support function claims where the level matches and the claim is worded at ingredient level, such as contains niacinamide at a level shown to support an even tone. They do not support a percentage claim for a finished product. Where a brand wants that, it has to test the finished product, and that is a real cost that belongs in the launch budget rather than a surprise at artwork stage.
Ask yourself
For your strongest performance claim, could you send the study report to a regulator this afternoon? Not the summary, not the supplier's brochure, the report.
Wording perception results honestly
Consumer perception studies are legitimate and useful, and they become misleading only in the wording. The honest form names the proportion, the number of participants and the question: 87 percent of 62 participants agreed their skin felt smoother after four weeks. The misleading forms drop the denominator, convert a feeling into a measurement, or present perception as proof: proven to smooth skin, when nothing was measured. A useful discipline is to write the claim and the study description side by side and check that a reader would understand what was actually done. Brands often find the honest version more persuasive anyway, because a specific number with a real sample sounds like evidence while a vague superlative sounds like advertising. This distinction matters most for proven results and visible results claims, since both imply a clinical test stands behind them even when the underlying work was only a perception survey.
Who should run the study?
An independent contract laboratory or clinical research organisation specialising in cosmetic testing, with the protocol agreed in advance and the full report delivered to the brand. Independence matters for credibility rather than for law: nothing prohibits in house testing, but a claim of independently tested requires it to be true, and an external report carries far more weight with a retailer, a marketplace or the ASA. A claim described as clinically tested should be backed by an actual clinical trial, not an internal perception panel dressed up in clinical language. Ask three things before commissioning. Which standard or protocol will be followed, and is it recognised. Who assesses the endpoint, and are they blinded to treatment. And what exactly will the report contain, including raw data, statistics and a clear statement of the result. It is also worth agreeing at the outset what happens if the result is negative, because a study that does not support the hoped for claim is a useful outcome that saves a brand from printing something it cannot defend, and a laboratory relationship that treats that as a failure is the wrong relationship.
A worked example: one serum's claims matched to its evidence
A brand submits a serum with four claims. Clinically proven to reduce the appearance of fine lines: the evidence is an instrumental study on 42 volunteers over eight weeks measuring wrinkle depth by image analysis, with a statistically significant reduction, so the claim is supported and the brand is advised to state the numbers. Dermatologist tested: the evidence is a human repeat insult patch test on 55 volunteers under dermatological supervision with no reactions, so the claim is supported and is best stated as dermatologically tested for tolerance. Boosts collagen: the evidence is a supplier in vitro study on fibroblasts, which supports neither the claim nor its implication and, in the UK and EU, edges towards a medicinal claim about physiological function, so it is struck. 95 percent said skin felt firmer: the evidence is a perception study on 20 participants, so the claim is supported only if the number of participants is stated, and the brand chooses to run a larger study rather than print 95 percent of 20. Two claims stand as written, one is reworded, one goes, and the pack ends up more specific than it started. Running proper clinical testing on a reasonable sample size, rather than relying on a single small clinical trial, is usually what separates a defensible claim from one an assessor has to strike.
How does this differ for medicinal sounding claims?
Evidence cannot rescue a claim that is in the wrong category. If a claim asserts that a product treats, prevents or cures a condition, or modifies a physiological function, the product is being presented as a medicine, and no amount of study data makes that lawful for a cosmetic under the cosmetic regulation; it would need a marketing authorisation from the MHRA instead. This is why boosts collagen, repairs the skin barrier and reduces inflammation are treated differently from improves the appearance of fine lines, helps maintain the skin barrier and soothes the look of redness. The first group claims a physiological action and the second describes appearance and comfort. Brands commissioning studies sometimes discover this late, having paid for research into a mechanism they cannot then mention, which is an expensive way to learn the distinction. Settle the category before the protocol is written, not after the results arrive.
Buying testing that will support your claim
• Write the claim first, then commission the study that proves it. The reverse order is how brands end up with a study that measures something they do not want to say.
• Test the finished product in its final formula.
• Match duration to the effect. Hydration in hours, barrier over weeks, wrinkle appearance over months.
• Size the panel for the claim. A percentage needs enough participants for the number to mean something.
• Use a control where the claim is comparative or strong.
• Get the full report, not a summary slide.
• File it in the claims table with a review date.
• Have your assessor review the claim wording against the report before print.
• Diarise a review date for every study. Evidence supporting a claim should be re examined when the formula changes, when a new market is added, or every two years, whichever comes first.
Match every claim to its evidence before the pack is printed
Spectra reviews draft claims against your evidence and the six common criteria as part of every Cosmetic Product Safety Report, and builds the Product Information File claims table where the study reports are indexed. Certified assessors, 3 to 5 day standard turnaround, Complete Compliance Pack at £295.
Myths we hear
• Dermatologist tested means it works. It usually means tolerance was assessed, and it says nothing about the result unless stated.
• Clinically proven is a protected phrase. It is not; the study behind it is what matters.
• A supplier's study covers my product. Only at ingredient level with a matching use level; a finished product claim needs a finished product study.
• Perception studies are not real evidence. They are, for perception claims, when the sample and question are stated.
• We can substantiate later if challenged. The evidence must exist when the claim is made, and the ASA gives days, not months.
What to do this week
• List every efficacy claim and the study behind it.
• Flag any claim resting on a supplier study rather than a finished product study.
• Add participant numbers to every percentage claim.
• Reword any in vitro result being presented as a human effect.
• File the full reports in the claims table with review dates.
Evidence is the most persuasive thing a beauty brand can publish, and the least used. State what you measured, how and on how many, and you will be both more lawful and more convincing than the pack beside you.
Related reading on the Spectra blog
• How a Cosmetic Product Safety Report exposes greenwashing
• Clean beauty and the Cosmetic Product Safety Report
• Natural, organic and chemical free claims in a Cosmetic Product Safety Report
• All Spectra articles
Spectra services mentioned in this article
• The Complete Compliance Pack: CPSR, PIF, SDS and SCPN or CPNP (£295)
• CPSR services by product type
• PIF services
• UK and EU Responsible Person
• Contact Spectra
Official sources and legislation
• Regulation (EU) No 655/2013 on cosmetic claims
• Advertising Standards Authority
• Regulation (EC) No 1223/2009 on cosmetic products, Annex I
• Medicines and Healthcare products Regulatory Agency
• OPSS guidance on cosmetic products regulations
About the author
Spectra Cosmetic Compliance is a UK based cosmetic regulatory consultancy (Spectra Private Limited, Chadwell Heath, Essex) whose certified safety assessors and toxicologists prepare Cosmetic Product Safety Reports, Product Information Files, Safety Data Sheets and SCPN and CPNP notifications, and provide UK and EU Responsible Person cover, for brands selling in the UK, EU, USA, Middle East and Asia. Standard turnaround is 3 to 5 days with a Regulator-Ready Guarantee. Talk to the team on WhatsApp at +44 7468 895676 or at info@spectracomply.com.
Disclaimer: This article is general information, not legal advice. Regulations change and product decisions depend on your exact formulation, claims and market. Have your product assessed by a qualified professional before sale.