The Category Nobody Designed

Somebody wrote those four words. Not a committee of philosophers deciding how science should relate to commerce — an actual drafter, at an actual desk, working on an actual regulation about the labelling of diagnostic test kits, in the mid-1970s. The problem in front of them was small and specific. A company developing an assay needs to send early versions to laboratories for evaluation before the assay is finished. Those shipments are commerce. Commerce in a medical device attracts the full apparatus of pre-market review. But the thing being shipped is not yet a product; it is a work in progress. What text on a carton makes that distinction visible?

The answer they arrived at was a required statement, and it is still there, unchanged in substance, half a century later: For Research Use Only. Not for use in diagnostic procedures. It was never a philosophy. It was a labelling solution to a distribution problem, and it has since become one of the most consequential phrases in a market its author had no way to foresee.

Written for test kits, in a hurry

The context was the American medical device amendments of 1976, which brought in vitro diagnostic products — the reagents, kits and instruments used to test specimens taken from the body — under device regulation for the first time. That created an immediate practical difficulty. Diagnostics are developed iteratively, in collaboration with the laboratories that will eventually use them, and material has to move between companies and laboratories throughout that process. A regime that treated every such shipment as the marketing of an unapproved device would have made development impossible.

So the labelling regulation carved out two staging categories. A product in the laboratory research phase of development, not yet ready for any evaluation of its diagnostic performance, must bear the research-use statement. A product that has moved on to being evaluated for diagnostic performance must instead bear an investigational-use statement together with a declaration that its performance characteristics have not been established 1. The two phrases are consecutive rungs on a development ladder, and both exist to say the same underlying thing: this is not a finished diagnostic and nobody is offering it as one.

That is the entire origin. There is no larger idea in it. The phrase does not descend from research ethics, from any principle about scientific freedom, or from a considered decision that research should occupy a lighter-touch legal space. It is a piece of administrative plumbing for the diagnostics industry, and it acquired its wider life purely because it turned out to be a convenient form of words.

One structural feature of the original design deserves attention, because its absence elsewhere explains a great deal. The two statements were rungs on a ladder with a top. A product bore the research-use statement, then the investigational-use statement, and then — if it was going anywhere — it was submitted, reviewed and cleared, at which point neither statement applied any longer. The categories were temporary by construction. They described where a product was on a journey towards a decision, and every product carrying one was understood to be in transit.

In the market the phrase now inhabits there is no such journey and no such decision. A compound can carry the research-use designation indefinitely, for years, for decades, with no submission pending, no review contemplated and no destination in view. The label was written to describe a stage, and it is being used to describe a permanent condition. Nothing in the drafting prevents that, because it never occurred to anyone that a thing might sit on the first rung of a ladder forever.

The regulator's own view of what the label is worth

By the 2000s the convenience had become a problem. Products bearing the research-use statement were being distributed to clinical laboratories and used, in practice, to generate results that informed the treatment of patients — a use for which nobody had ever established that they worked. The agency's concern was not abstract: unvalidated assays were producing numbers that doctors were acting on. In 2013 it published a final guidance on the point, announced in the federal gazette that November 3.

The guidance's central holding is the sentence anybody using the phrase should know. The label statement is not determinative. Where the totality of circumstances shows that a product is intended for clinical diagnostic use, the presence of a research-use disclaimer does not change that, and the product is subject to the requirements applicable to what it actually is. The guidance then lists the kinds of circumstance it will weigh: written and oral statements by the manufacturer and its representatives, promotional materials and the venues in which they appear, instructions that explain clinical interpretation, technical support offered to clinical laboratories, certifications sought, and the knowledge a distributor has about what its customers are doing with the product 2.

That is the same doctrine of intended use that runs through everything else in this cluster, applied here in its purest form. What a product is, in law, is determined by what it is offered for; and what it is offered for is established from conduct, not from a disclaimer. A label saying research use only is evidence about intention. It is not proof of intention, and it can be comprehensively outweighed by everything else a supplier does.

The same idea, arrived at in Brussels

European law came to a comparable position independently. The Union's in vitro diagnostic regulation, which replaced a much older directive and became fully applicable at the end of the 2010s, defines its own scope by reference to what the manufacturer intends. Products for general laboratory use and research-use products fall outside the regulation unless, in view of their characteristics, they are specifically intended by their manufacturer for in vitro diagnostic examination 5. The escape is real, but it is again the manufacturer's own stated intention that determines it, and again the characteristics of the product can contradict what the manufacturer says.

The wider backdrop in both jurisdictions is that an enormous amount of real clinical testing runs on assays that were never submitted to anybody for approval. Laboratories design and validate their own tests, a practice with its own decades-long regulatory argument, and the components those tests are built from very often carry research-use labelling 4. So the phrase is not marginal. It sits underneath a substantial part of working diagnostic medicine, which is precisely why regulators keep returning to it and why the boundary keeps being litigated.

How a stock-room convention became a consumer phrase

For most of its life the phrase was invisible to the public, because the people reading it were laboratory buyers. It appears on antibodies, primers, enzymes, cell lines, standards, reference materials — the ordinary contents of a research supply catalogue. In that setting its meaning is completely uncontroversial and everyone understands it identically. It means: we have characterised this for research applications, we have not validated it for any clinical purpose, and if you use it in a clinical workflow the validation burden is yours. Nobody in a laboratory reads it as a hint.

What changed was the audience. Once material carrying that labelling convention became visible to, and purchasable by, people who are not laboratories, the phrase acquired a second reading that its drafters never contemplated. To a general reader it does not look like a disclaimer of claims at all. It looks like a category — a nudge, a knowing formula, the sort of thing written on something that is really one thing while formally being another. The words are identical and the meanings are almost opposite.

It is worth being exact about what the label does commit a supplier to, because it is not nothing. Any specific claim actually made alongside it — an identity, a purity figure, a certificate of analysis, a storage specification — is a claim, and remains fully subject to the ordinary law on misdescription and consumer protection wherever the supplier operates. The research-use statement disclaims a diagnostic or therapeutic purpose. It does not disclaim the accuracy of anything else on the page, and treating it as a general-purpose escape from responsibility misunderstands its scope as badly as reading it as a wink does.

And what it does not commit anybody to is equally exact. It is not an approval, because no approval process was involved. It is not a safety finding, because nobody assessed safety. It is not a quality grade, because quality is not what the phrase is about. It is not a purity guarantee, and it is not an endorsement of anything at all. It is a sentence declining to make claims, which is the least assertive thing a piece of labelling can do, and which is very close to the opposite of how the phrase now travels.

That gap — between a plumbing fix for the 1970s diagnostics industry and a phrase carrying implied meaning in a consumer market fifty years later — is the last accident in this cluster, and the most characteristic. Nobody designed this category. It was a side effect of solving a different problem, it was never revisited, and it has been asked to bear far more weight than any four words drafted for that purpose could reasonably carry.