The phrase 'there are studies'

This page is written for general readers. It does not recommend any procedure, product or clinic, and it carries no efficacy, duration or price information. Decisions about treatment belong with a qualified clinician who has examined you.

Question

What does it mean when something is said to be proven by studies?

That studies exist tells you almost nothing on its own. What tells you something is what kind of studies they are. Work done in a culture dish, work done in rats, and work randomising people into compared groups do not carry the same weight. Papers are also retracted, and after retraction they remain high in search results and continue to be cited.

The type is the weight

In the sentence “proven by studies”, what needs checking is not the existence of studies but their type. The weight differs roughly in this order.

There is work where a substance is added to cells in a dish and observed. This shows what the substance can do to cells, but not that the same happens inside a body.

There is work applying it to animals and examining tissue. This is considerably stronger. Rat skin and a human face are still not the same thing.

There is work in people that looks only at before and after. Here, change caused by time and change caused by the treatment are hard to separate.

There is work randomising people into groups and treating only one. This is the strongest form.

“Several papers” is not a quantity of evidence

The same group can publish several papers using the same approach, and a review paper organises existing work without adding new data. Paper counts map poorly onto the thickness of evidence.

This is why this site limits each material to between three and five citations and records the type of each. Stating what something is turns out to be more useful than stacking more of it.

Retractions are more common than expected

This is recorded here because it happened while building the site. Searching for particular materials returned several retracted papers among the top results, including a large trial in people.

Retraction happens quietly. The paper remains where it was and citation continues. So when a paper number is offered, a step that checks the current status of that number is needed. This site puts that check in front of citation.

Conflicts of interest are part of the evidence

Papers carry a section stating the relationship between authors and manufacturers. A relationship does not make the work wrong, but it can shape which literature was selected and how it was interpreted. Some authors state exactly that themselves.

One mechanism page on this site carries the conflict-of-interest disclosure from a review it cites, reproduced as written, because that information is part of the answer.

What to check first

Was it done in people, in animals, or in cultured cells?

That order is the order of weight. Something observed in cultured cells does not mean the same happens in a body, and something observed in animals does not mean the same happens in people. It is why this site marks a study grade against every citation.
Was there a comparator?

Without a comparison group, an observation that things improved cannot separate improvement over time from improvement due to the treatment. What it was compared against comes before any figure in the result.
How many people?

With few participants, differences arise by chance. A small study is not a wrong study, but a result from few people is hard to apply to everyone.
Is that paper still standing?

Retracted papers stay high in search results and continue to be cited. Several were filtered out while building this site, including a large trial in people. Where a paper number is given, its record status can be checked.
Did that study use this product?

It may have used a different product carrying the same material name. Why the same material name does not mean the same product is covered on a separate page.

What this page does not say

This page does not evaluate the evidence behind any particular paper or product. It covers only how to tell study types apart. The distinctions here are general, and whether an individual study was well designed is not settled by its type. A randomised trial can be poorly designed, and a small observational study can show something important.