The best question to ask about any aesthetic procedure is not "does it work?" but "how would I know if it did?" A photograph cannot answer that. Neither can a five-star review or a phrase like "clinically proven." What answers it is a short chain of questions about the study behind the claim: what the product legally is, who it was tested against, how it was measured, and how long anyone kept watching for both the benefit and the harm. This is an evidence-literacy piece, not a recommendation of any procedure, device, or brand.
Key points#
- A before-and-after image is a testimonial photographed at peak effect, not a controlled result.
- What a product legally is (device versus cosmetic versus drug) sets the minimum evidence that had to exist before it reached the market.
- Trustworthy aesthetic studies compare against a control and score outcomes with validated scales rated by someone blinded to the treatment.
- The follow-up window should be at least as long as the promised benefit, and long enough to catch delayed harms.
- Absence of reported harm often just means the study was too short and too small to find it.
Why a photograph is the weakest form of evidence#
Aesthetic results are unusually easy to make look convincing, which is exactly why the picture deserves the most suspicion. Lighting, angle, makeup, and the moment of capture (peak swelling can read as fullness) all move the image before any biology does. More importantly, a photo is a single person at a single instant with no comparison group. You cannot see what that face would have looked like untreated, and you cannot see the people for whom nothing changed, because they rarely appear in the gallery. Treat every before-and-after as a testimonial, then go looking for the study it is standing in front of.
First, pin down what the product legally is#
Before you weigh any study, establish the product's legal category, because that single fact sets the evidence floor. The marketing almost never makes the distinction clear, and the categories are not interchangeable.
Injectable soft-tissue fillers are regulated by the FDA as medical devices. The ones cleared for the face and hands went through Premarket Approval, the agency's most demanding device pathway. The FDA states that its approval of these fillers rests on data from controlled clinical studies of safe and effective use when the product is injected into specified areas of facial tissue and the hands. That is a real bar, and it is also a narrow one: approval attaches to a specific material, a specific site, and a specific indication, not to the general idea of a "filler."
Topical products live under different rules. Under the Federal Food, Drug, and Cosmetic Act, whether something counts as a cosmetic or a drug turns on its intended use, which the FDA reads from the claims made about it. A product meant only to beautify or change appearance is a cosmetic, and the law does not require FDA approval before a cosmetic reaches shelves. The moment a product claims to affect the structure or function of the body, including the skin, or to treat or prevent disease, it meets the legal definition of a drug and must clear drug requirements. The FDA is blunt that "cosmeceutical" carries no legal meaning and lowers no bar. So a serum promising to "rebuild collagen" is, on its face, making a drug-level claim while usually carrying only cosmetic-level evidence. That gap between what is promised and what was proven is the first thing worth noticing.
Follow the comparator and the money#
Once you know the category, look at how the underlying study was actually built. Aesthetic outcomes are vulnerable to bias because the endpoint is largely subjective and because expectation shapes what people see in the mirror.
Three questions do most of the work. Who was compared to whom? A single-arm series where everyone gets the treatment and reports feeling better tells you little; a randomized comparison against a control, with assessors blinded to who received what, tells you far more. How was the result measured? A validated wrinkle-severity or aesthetic-improvement scale scored by a blinded evaluator is worth more than an untrained "satisfaction" count. And who funded and designed the study, and were the full results published rather than shown only as a favorable summary slide? None of this needs a statistics degree. It needs reading one layer past the headline number to the sentence describing the comparison.
Match the follow-up window to the promise#
Durability is where the marketing and the evidence part ways most often, because a result photographed at its peak says nothing about the slope afterward.
The FDA's own material shows how much staying power depends on the material. Most approved fillers are absorbable and temporary. By the agency's account, hyaluronic-acid products last roughly six to twelve months, calcium hydroxylapatite roughly eighteen months, and poly-L-lactic acid, given as a series of injections over several months, may last up to about two years. The FDA has approved only one filler made from a material that stays in the body and is not absorbed, a product using polymethylmethacrylate beads. That fact reframes the whole question, because a longer-lasting material also means any problem lasts longer. So when a claim promises a certain duration, check that the study followed people at least that long. A twelve-month promise backed by three months of data is an extrapolation dressed as a finding.
Read the harms the way you read the benefits#
Harms are the thinnest part of most aesthetic evidence and, for anyone weighing a rare but serious risk, the part that matters most.
The FDA notes that most filler side effects, such as swelling and bruising, appear soon after injection and often settle within days to weeks. But it also warns that the gravest risk, accidental injection into a blood vessel, can cut off blood supply and lead to complications that include tissue death, vision problems including blindness, and stroke; the odds are low, but the damage can be permanent. The agency adds that some side effects can surface weeks, months, or years later. That timeline is the whole lesson: a study that follows patients for twelve weeks simply cannot describe a harm that emerges in year two. When you read a harms table, check how events were collected (actively asked about, or passively reported), how long surveillance ran, and whether rare serious events were counted at all or were merely absent because the sample was too small to detect them. Absence of reported harm is not proof of safety; it is usually proof of a short, small study. The FDA's separate advice against using fillers for body contouring is a reminder that using an approved product outside its studied indication steps outside the evidence entirely.
A four-question habit#
Put together, appraising an aesthetic claim comes down to a repeatable habit. Name the legal category and the evidence floor it implies. Confirm the outcome was measured against a comparator on a validated scale. Check that the follow-up window covers the promised durability. And read the harms reporting for both severity and how long anyone was watching. The photograph is where the pitch starts; these four questions are where the evidence either holds up or falls apart.
Sources and further reading
Questions and answers
Does "FDA-approved" mean a treatment is proven safe for me?
It means a specific material cleared a specific bar for a specific site and indication, based on controlled studies. It does not extend to other uses, other body areas, or off-label combinations, and it does not remove individual risk. Approval narrows the question; it does not close it.
Is a "clinically proven" cosmetic the same as a tested drug?
No. "Clinically proven" is a marketing phrase with no fixed legal definition, and "cosmeceutical" has none at all. A cosmetic can carry that language while meeting only cosmetic-level evidence expectations, which are far lower than the standard a drug must satisfy.
If I cannot find any reports of harm, does that mean it is safe?
Not necessarily. Rare and delayed harms only show up in studies that are large enough and long enough to catch them. Silence in a short, small study is more likely to reflect the study's limits than the treatment's safety.