Why the burden sits with you
The principle underlying this instrument is not a matter of opinion. Under the UK advertising codes, the person making a claim is expected to hold the evidence for it before the claim is published, and to produce that evidence on request. The obligation is not triggered by a complaint; it precedes publication.
Most clinics do not disagree with this and still fail it, because claims arrive by inheritance. A distributor supplies copy for a device. A page is written from that copy. Two years later the device has a new version, the study the copy referred to has been superseded, and nobody at the clinic has ever seen either. The claim on the page is now an assertion with no owner.
A claim inventory breaks the inheritance. Reading a page sentence by sentence and asking what each sentence asserts is a mechanical exercise, and it consistently finds three categories: claims the clinic can support immediately, claims it can support with a document it has to go and find, and claims it cannot support at all. The third category is usually smaller than clinics fear and larger than they expect.
The non-broadcast advertising code is the reference document for what is required. Reading its provisions on health and beauty claims before scoring this instrument is time well spent.
How to build a claim inventory
Take the page as plain text. Number every sentence. For each sentence, write what it asserts, in your own words, as a proposition that could be true or false. Sentences asserting nothing are marked as such and set aside.
Classify each proposition: efficacy, safety, comparative, superlative, durability or experiential. The classification matters because different types are supported differently. An efficacy claim needs evidence about outcomes. A durability claim needs evidence over time, which is scarcer. A comparative claim needs evidence that made the comparison, which is scarcer still.
For each proposition, record the source. Product instructions for use, a named published study, the clinic's own audited records, or nothing. Record the date and version. "Nothing" is a legitimate entry and the whole point of the exercise.
Score the ten criteria against the completed inventory. Where a criterion asks whether evidence is held for every claim of a type, score 0 if a single claim of that type lacks it. Substantiation does not average.
Keep the inventory. It is the substantiation record the tenth criterion asks for, and rebuilding it later costs more than maintaining it.
Common scoring errors
Accepting supplier marketing as substantiation. A distributor's brochure is a claim, not evidence for a claim. Instructions for use are a usable source; a sales deck is not.
Scoring a claim as supported because it is widely repeated. Frequency is not evidence, and a claim common across an entire sector is still your claim once you publish it.
Missing implied claims. A before and after pair with no interval stated implies a timescale. A photograph of a patient who plainly had other treatment implies attribution. Implied claims count.
Treating testimonials as a way to make an unsupportable claim safely. Presenting a claim in a patient's voice does not shift the substantiation burden.
Scoring the record as present when only its author can use it. A folder of PDFs on one person's laptop is not a record. The criterion asks whether a colleague could use it.
What to do with a claim you cannot support
There are three defensible options and one that is not.
Remove it. Fastest, and usually costs less than clinics expect, because unsupportable claims are typically the vaguest sentences on the page and their removal is not noticed by readers.
Narrow it until it is supportable. "Results last up to eighteen months" often becomes "in the studies we rely on, effect was observed at twelve months in most participants, and our own patients typically return between nine and fifteen months". Longer, duller, defensible, and more useful to a reader deciding.
Reframe it as clinic practice rather than as a general claim. "We use this technique because of what we have observed in our own patients" asserts something about you, which you can support from your own records, rather than something about the world.
The option that is not defensible is leaving it and hoping. The burden does not lapse with time, and a page that has carried a claim for five years is not thereby safer.
When the inventory is clean, run the advertising compliance self-audit, which covers matters substantiation does not, and the before and after imagery governance scorecard if your pages carry photographs.