How to Read Scientific Skincare Studies (Without Being a Scientist)

Table of contents

1. The problem with "scientifically proven"
2. The types of studies that exist (and which ones carry more weight)
3. The 5 questions you should ask yourself before any study
4. Red flags: when a study is not what it seems
5. Where to find real studies
6. Frequently asked questions

The problem with "scientifically proven"

You see a cream on Instagram. The product description says "clinically proven results" or "backed by scientific studies." You feel that should mean something. And you're right: it should. The problem is that many times it means nothing.

"Clinically proven" has no standard legal definition for cosmetics in the European Union. Any brand can write it if they have done even a basic satisfaction test with 20 people who evaluated themselves. It is not technical fraud. But it is also not what most people understand by that phrase.

The good news is that evaluating the real quality of a study does not require being a scientist. It requires knowing what to ask.

The types of studies that exist (and which ones carry more weight)

Not all studies carry the same weight. Science has a hierarchy of evidence, and understanding it changes how you interpret brand claims.

In vitro (in the laboratory, with cells)

These are studies with cells in Petri dishes or test tubes. They are useful for understanding basic mechanisms, but they do not prove that something works on real human skin. An isolated cell in a laboratory lives in completely different conditions than the cells on your face.

When a brand says "proven in a laboratory," it often means this. It is the lowest level of evidence.

In vivo in animals

Studies in mice or other animals. Useful as a preliminary step before testing on humans. Results do not automatically transfer to people. They are becoming less common in cosmetics today.

Observational studies in humans

A group of people who already use a product or treatment is observed, without randomly assigning who uses it and who does not. They can show associations, but cannot prove causality. Example: "people who use sunscreen age better" — it may be that they also take better care of their skin in general.

Controlled clinical trials

Participants are divided into groups (treatment and control) and are randomly assigned. The best are double-blind: neither the participant nor the evaluator knows who received what. This is the gold standard for proving that something works.

Systematic reviews and meta-analyses

These are studies of studies. They collect all clinical trials published on a topic and synthesize the results. When there are many studies pointing in the same direction, the evidence is considered robust. This is the highest level of evidence.

The 5 questions you should ask yourself before any study

1. How many people participated?

Small sample size = weak conclusions. A study with 15 people can show interesting trends, but cannot generalize. To draw conclusions applicable to a large population, hundreds or thousands of participants are needed, depending on the effect studied.

In skincare, many studies have 20-50 participants. They are not useless, but their conclusions should be read with caution. Reference studies usually have 100+ participants.

2. Was there a control group?

If everyone in the study received the active treatment, you cannot know if the results are due to the treatment or simply the passage of time, the placebo effect, or that the people paid more attention to their skin during the study.

A control group is the group that receives a placebo or follows the usual protocol without the evaluated treatment. Without it, the numbers can be misleading.

3. Was it blind or double-blind?

The placebo effect is real and powerful. If people know they are receiving an active treatment, they tend to report improvements even if the treatment does nothing. In a "blind" study, participants do not know if they are in the treatment group or the placebo group. In a "double-blind" study, neither the participants nor the evaluators know.

In device studies (such as LED masks), blinding is complicated because the person can see whether the light turns on or not. That is why many phototherapy studies cannot be completely blinded. This is a real limitation of the field, not dishonesty.

4. Who funded the study?

A study funded by the same company selling the product has an obvious conflict of interest. That does not automatically invalidate it, but it deserves more scrutiny. Independent studies or those funded by academic institutions have greater credibility.

Most scientific journals require declaring conflicts of interest. If they are not declared, it is a red flag.

5. Was it published in a peer-reviewed journal?

Peer review means that other scientists in the field evaluated the study before publishing it. It does not guarantee perfection — flawed studies pass peer review — but it is a minimum quality filter.

PubMed indexes peer-reviewed journals. If you can find the study on PubMed, at least you know it went through that process. If the study only exists on the brand's website, it has not passed any independent review.

Red flags: when a study is not what it seems

• "97% of users noticed improvement." Subjective self-evaluation, no blinding, no control. The person who bought the product with hope will tend to report improvement regardless of the actual effect.
• "Proven in 4 weeks." It may be true, but with what methodology? Who measured the improvement? How was it measured?
• "Dermatologically tested." This only means that a dermatologist evaluated the safety (that the product does not cause adverse reactions). It says nothing about efficacy.
• The sample is not described. If it is not specified how many people participated, what the inclusion/exclusion criteria were, and how the results were measured, the study is not evaluable.
• The study is for another product from the same manufacturer. Just because a serum from a brand has proven something in a study does not imply that its mask is effective.

Where to find real studies

If you want to go to the primary sources, these are the tools we use:

• PubMed (pubmed.ncbi.nlm.nih.gov): The biomedical literature database from the U.S. National Institutes of Health. Free access. You can search by terms such as "LED therapy skin," "photobiomodulation rosacea," or "red light collagen."
• Cochrane Library (cochranelibrary.com): High-quality systematic reviews on medical and health treatments. They filter the best studies available on each topic.
• Google Scholar: Broader than PubMed, includes more types of publications. Useful for an initial exploration.

A typical search on PubMed: write the terms in English (most scientific literature is in English), filter by "Review" or "Systematic Review" if you want the most solid summaries, and read the abstract, which is always accessible for free even if the full article is paid.

Frequently asked questions

What does "clinically proven" mean in a cosmetic?

In practice, it can mean almost anything. There is no standard legal definition for cosmetics in the EU. It can range from "we did a tolerance test on 10 people" to "we carried out a controlled clinical trial with 200 participants." Without access to the study data, the claim alone says nothing.

Does it mean anything if a product has CE certification?

Yes, although in a different sense than efficacy. The CE certification for aesthetic devices verifies that the device meets safety standards: it does not emit dangerous radiation, is electrically safe, and does not cause burns under normal conditions of use. It does not certify that the device is effective for the conditions it promises to treat. Safety and efficacy are two different things.

Can I trust a study done by the brand itself?

With caution. Internal studies do not go through independent peer review and there is a financial incentive to publish positive results. That does not automatically make them lies, but they do deserve more scrutiny than a study published in an academic journal by researchers with no conflict of interest with the manufacturer.

Is it normal for there to be contradictory studies on the same ingredient?

Very normal. Science works by accumulation and review. A first study generates a hypothesis. The following ones confirm, nuance, or refute it. When there are contradictory studies on something, the scientific consensus is formed by looking at the body of evidence, not an isolated study. That is why meta-analyses matter: they integrate all available studies.

Reading with a critical spirit does not mean mistrusting everything. It means asking the right questions. LED phototherapy has real scientific backing — and part of that backing comes from studies with solid methodology published in reference journals. Knowing how to distinguish that type of evidence from an Instagram testimonial is what allows you to make informed skincare decisions, not those driven by marketing.

Do you want to see which specific studies support LED phototherapy? We have the summary of the most important research.

Written by the Anelle Skin team