What a clinical study really is (and why it matters for skin care)

“Clinically tested.” “Proven in a clinical study.” These phrases do a lot of work on skin care packaging, and most of us nod along without knowing what sits behind them. A clinical study is simply a test on real people. Whether it tells you anything useful depends on how it was built. Here’s the anatomy, in plain words.

What counts as a clinical study?

At its simplest, a clinical study is research done on people rather than in a test tube or on animals. Health Canada describes clinical trials as studies “to find out whether a drug is safe and effective for people”, and the NHS puts it even more plainly: a clinical trial “compares the effects of 1 treatment with another.”

That word “compares” is the whole game. A study that just hands a cream to 30 people and asks how they liked it is a clinical study in the loosest sense, but it isn’t comparing anything. The value of a study comes from the comparison, not from the word “clinical”.

For prescription drugs, Canada requires trials to be reviewed by Health Canada and by a research ethics board, with informed consent from every participant. Cosmetic studies don’t go through that gate. A company can run whatever test it likes and describe it however it likes, as long as the claim is truthful. That’s why it helps to know the parts.

Participants: who was tested, and how many?

Start with the people. Were they the kind of people the product is for? A hand cream tested on 20-year-olds with healthy skin tells you little about your dad’s cracked winter hands. A study of adults says nothing about babies.

Then, how many. A 2024 paper in The Oncologist walks through what happens with tiny groups: the authors drew random samples of ten from two known populations, over and over, and found that the same populations sometimes looked significantly different and sometimes looked the same, purely by chance. Their conclusion: results “are unreliable with small samples,” and larger groups reduce that random noise.

There’s no fixed threshold that makes a study “big enough”; it depends on how large the effect is and how precisely it can be measured. But when a label rests on 12 or 15 people, treat it as an early hint, not a settled fact.

The control group: what would have happened anyway?

Skin changes on its own. Dryness improves when the furnace goes off in spring. Redness from a flare settles over a few weeks. And simply joining a study, being watched and paying attention to your skin, changes how you rate it.

A control group is a second set of participants who get something else: a plain, unscented base cream with no active ingredient (a placebo, meaning an inactive look-alike), the product people already use, or nothing at all. Whatever the product group gains beyond the control group is the part you can credit to the product.

The 2012 paper on randomized double-blind placebo-controlled studies makes a point that matters for skin care: “no treatment is not the same as placebo treatment.” Rubbing anything on your skin, and being part of a study, has effects. A good control isolates them.

Without a control, you get the classic “87% saw improvement” number. Improvement compared with what? If the control group also improved 80%, the product did very little.

Randomization and blinding: keeping the deck unstacked

Two more pieces protect the comparison.

Randomization means a coin flip (usually a computer) decides who gets the product and who gets the control. As the 2012 paper explains, this avoids the bias that creeps in if a researcher or participant picks their own group, and it balances out factors nobody thought to measure. If the people with the mildest dryness happen to land in the product group, the product looks better than it is. Random assignment stops that.

Blinding means the people involved don’t know who got what. The NHS describes it simply: many trials are set up “so nobody knows who’s been allocated to receive which treatment,” which “helps reduce the effects of bias when comparing the outcomes.” In a single-blind study, participants don’t know. In a double-blind study, neither do the people grading the skin. That matters a lot in skin care, where outcomes are often judged by eye: an assessor who knows which arm got the “active” cream will, without meaning to, see more improvement there.

A study that is randomized, controlled and double-blind is what the 2012 paper calls the gold standard, because it’s the design that can show cause and effect rather than coincidence.

What was measured, and for how long?

A study can have every safeguard above and still measure the wrong thing. Ask what the outcome actually was:

• Did an instrument measure skin hydration or water loss, or did people fill in a questionnaire?
• Was the change something you’d notice, or a small shift in a lab reading?
• How long did it last? Two weeks tells you about two weeks. Many skin conditions cycle over months.
• Was the outcome decided before the study started, or picked afterwards because it looked good?

Neither instrument readings nor questionnaires are wrong; they just answer different questions. A reading tells you the product did something physical. A questionnaire tells you people liked it. A claim should match the kind of measurement behind it.

Was it published?

The NHS says that at the end of a trial, researchers “should publish the results.” Publication in a peer-reviewed journal (one where outside experts check the work before it appears) means someone independent looked at the methods, the numbers and the conclusions.

Most cosmetic studies never get that far. They live in a company’s files, summarized in a footnote on the box. That doesn’t make them fake. It does mean nobody outside the company has checked whether the study supports the claim, which is exactly the gap an independent review tries to fill.

Are company-funded studies bad?

Almost every skin care study is paid for by the company selling the product, because nobody else has a reason to run it. So dismissing all manufacturer-funded research would mean dismissing nearly everything.

The honest position sits in the middle. A Cochrane review updated in 2017, covering 75 papers, found that industry-sponsored drug and device studies reported efficacy results favourable to the sponsor more often (a risk ratio of 1.27) and reached favourable conclusions more often (1.34) than studies funded elsewhere. Notably, the authors found this lean “cannot be explained by standard ‘Risk of bias’ assessments,” meaning the studies often looked well designed on paper. The tilt shows up in subtler choices: which comparison to make, which outcome to headline, whether a disappointing result gets published at all.

So the funder isn’t a verdict; it’s a reason to look harder. A manufacturer-funded study that is randomized, controlled, blinded, adequately sized, pre-registered and published in full is good evidence. A manufacturer-funded study that is none of those is a marketing document with a lab coat on.

If you’re dealing with a real skin problem, eczema (atopic dermatitis, the itchy inflamed rash), persistent acne, or a rash that won’t settle, product studies are not a substitute for care. If it isn’t better after two weeks, see your doctor or pharmacist.

What to remember: a clinical study is only as good as its comparison. Look for a control group, random assignment, blinding, a reasonable number of the right people, an outcome that matches the claim, and publication where someone outside could check it. Company money doesn’t sink a study, but it does mean the design has to carry the weight.