Under the Dietary Supplement Health and Education Act, a supplement may describe how it affects the normal structure or function of the body. It may not claim to diagnose, treat, cure, mitigate or prevent a disease. That sentence is short and the boundary it draws is where most claim problems actually live.
The boundary in practice
The difference is rarely about the ingredient and almost always about the verb and the noun beside it. Saying a product supports a normal function is structure/function territory. Naming a condition, or describing the relief of one, moves the same product into drug-claim territory regardless of what the data shows.
Context counts too. A compliant sentence can be pulled across the line by what sits around it — imagery of a medical setting, a customer review quoted on the page, a paid search term, a testimonial in an email. Regulators read the whole impression, not the sentence in isolation.
Substantiation is a separate test
Staying inside structure/function makes a claim permissible. It does not make it supported. Those are two gates and a claim has to pass both.
Support means competent and reliable evidence for the specific claim, in a relevant population, for the actual formulation and dose being sold. Evidence for a single ingredient at a research dose does not automatically support a finished product containing less of it in a different matrix — an assumption that quietly underpins a great deal of category copy.
Match the evidence to the sentence
Different sentences need different work, and it is worth being blunt about which is which.
A composition statement — what is in the product, how much, verified — needs analytical evidence. A statement about how people experienced or accepted the product needs a prospective consumer study with a prespecified endpoint. A statement about a functional or physiological outcome needs a clinical study designed for that endpoint. A discovery-scale sensory test supports none of the three as an external claim, however encouraging its numbers.
Write the sentence first
The most reliable way to end up with an unsupportable claim is to run studies and then look for something quotable in the results. The most reliable way to end up with a supportable one is to write the sentence you want, then design the evidence chain backwards from it.
Working in that order also surfaces the cheap answer early. Quite often the sentence a brand actually needs is a composition claim that analytical work already supports, and the expensive study was never required.
What review will ask for
Expect to show the claim as it appears in context, the study or certificate behind it, why the population and dose are relevant to the product as sold, and who reviewed it. A claim that can be traced through one documented chain — from certificate to protocol to report — clears review faster than one assembled from several sources that were never designed to sit together.
This note describes curí’s own laboratory practice and is written for brand and regulatory teams. It is not regulatory advice on a specific product. All programs are designed to support structure/function claims permissible under DSHEA; curí does not design studies to support disease claims.