The regulator followed the footnote home

Two ads, two brands, one upheld ruling each#

On 26 August 2026 the Advertising Standards Authority published a ruling against Vir Health, trading as Numan, over two paid Facebook ads seen in January 2026. A companion ruling against BioIdentical Hormone Therapy, trading as BioID Health, landed the same day. Both were upheld. Neither turned on a fabricated citation.

The first Numan ad carried the line "40% OF MEN WITH OBESITY HAVE LOW TESTOSTERONE", the caption "Testosterone deficiency is a battle, you don't have to fight it alone", and a clickable button reading "Blood tests from GBP 74.50". The second claimed "95% OF MEN WITH LOW TESTOSTERONE ARE UNDIAGNOSED" with the same button. Both carried an academic citation in small print, which is more than most medical advertising bothers with.

That is what makes this worth reading. The brand did the thing every medical advertising guide tells you to do: it found peer-reviewed evidence and put the reference in the ad. It still lost, because the regulator did something advertisers rarely expect. It opened the paper.

Worth noting how these were caught. The ruling states this forms part of wider work on low-testosterone ads, identified through complaints and intelligence from the ASA's Active Ad Monitoring system, which uses artificial intelligence to search proactively for online ads that might break the rules. One complaint reached the Numan case. The scent was picked up by a machine.

The number that shrank at every step#

Numan's 40% rested on a 2014 review in the Asian Journal of Andrology, which in turn drew on an analysis of men from the Hypogonadism In Males study. The review says, in one sentence, that in a cross-sectional study of 1,849 obese US men, 40% had low testosterone.

Open the underlying analysis itself and that sentence does not survive. The 40% figure belongs to obese non-diabetic men aged 45 and over, a subgroup of 489. Of the 1,849 analysed, 962 were non-diabetic men who were lean or overweight, and 398 were diabetic.

The ASA put it plainly: the review "wrongly stated that this was based on the total cohort of 1,849, rather than the true figure of 489, and did not make clear it was exclusively non-diabetic men". It also found the ad failed to make clear the finding did not apply to men aged between 18 and 44.

Then the detail that should stop any performance marketer mid-scroll. The same analysis reports six prevalence figures, not one, and the diabetic men score higher at every weight. The claim picked the number that suited the audience it wanted, and dropped the five that did not.

Numan's fallback evidence fared no better. It offered a systematic review of 18 studies covering 4,546 obese men, pooled prevalence 42.8%. The ASA noted that figure measured total testosterone, while the 40% claim rested on free testosterone, for which the same review found 32.7%. Prevalence across the constituent studies ranged from 0 to 51.5%, and all but one carried a higher risk of bias for taking a single measurement.

The 95%% claim collapsed faster. Numan accepted the survey cohort were men with symptoms of low testosterone, not men with a confirmed diagnosis, and withdrew the ad.

Two further details from the ruling deserve a place in any medical advertising post-mortem. The ASA noted the cross-sectional study was published in 2006 and the analysis in 2010, making the evidence twenty and sixteen years old respectively. It also noted the study involved a non-UK population and had been conducted on behalf of a pharmaceutical company.

Numan argued the US finding travelled to a UK audience because both countries classify obesity as a body mass index of thirty or above, in line with the World Health Organization. That part of the argument was reasonable. It simply did not rescue a claim whose problem was the cohort, not the country, which is a distinction worth holding on to in medical advertising generally: fixing the wrong objection does not answer the right one.

The second ruling is about positioning, not numbers#

BioID Health's ads took a different route to the same place. One opened "Your GP said your testosterone is 'fine'. But you still feel exhausted, foggy, and flat. We test deeper than your doctor did". Another ran "Still Exhausted after 8 hours Sleep? Your GP Says You're Normal, We Say You Deserve Better".

The ASA investigated two issues: whether the superiority claims over GPs could be substantiated, and whether the ads were irresponsible for implying GPs were insufficiently thorough. Both were upheld.

The second ad went further, listing symptoms under the heading "The symptoms they dismiss": waking tired despite adequate sleep, constant brain fog and forgetfulness, zero motivation or drive, and unexplained weight gain around the middle. It then set out a specific technical argument, that NHS reference ranges span men aged 18 to 80, and that at 300 nanograms per decilitre a man can be inside the range and still exhausted.

Read coldly, that is a sophisticated piece of medical advertising. It names a measurable threshold, explains why a normal result might not feel normal, and offers a route. What it does not do is establish that the advertiser's testing genuinely finds what a GP misses, which is the claim the copy actually makes.

This is the sharper lesson for anyone running direct-to-consumer telehealth creative, because the offending material is not a statistic at all. It is a comparison. Positioning a private service against the NHS is a claim about another provider's competence, and it carries the same substantiation burden as a percentage.

Fairness demands one note in BioID's favour. Their underlying clinical point, that laboratory reference ranges and clinical action thresholds are different things, is not absurd. The NHS's own page on the subject describes the contested territory around testosterone and ageing in plain terms. The ruling went against how the point was sold, not against the existence of the debate.

Four lessons for medical advertising that has to hold up#

The pattern across both rulings is the same, and it is not dishonesty. It is compression. Somebody took a finding with four qualifiers attached and wrote it as a headline with none, because headlines with qualifiers do not perform. That pressure exists in every medical advertising brief ever written, which is why the check has to sit outside the person feeling it.

First, the citation is not the substantiation. A reference in small print proves a paper exists; it does not prove the paper supports the sentence above it. Every claim in dtc healthcare creative needs someone to have read the source and checked the population, not just the abstract.

Second, the population is the claim. "40% of men with obesity" and "40% of non-diabetic obese men aged 45 and over" are different sentences with different truth values. Where a study restricts by age, comorbidity or geography, so must your copy, or you have quietly widened the finding.

Third, check the chain, not the link. The review Numan cited was peer-reviewed and it still misdescribed its own source. The World Health Organization defines obesity consistently enough that the cross-border argument held, but the cohort argument did not. One honest intermediary is all it takes to inherit somebody else's error.

Fourth, comparisons are claims. Any line that positions your service as more thorough than another provider needs evidence for that specific comparison. Section 3 of the CAP Code, on misleading advertising and Section 12, on medicines and health-related products both apply, and the MHRA's Blue Guide sets out where promotional material crosses into regulated territory.

Neither company was fined. The sanction in UK telehealth advertising is the ruling itself, published on the ASA's public rulings database, carrying the brand name and the finding, indexed and quotable forever. For a category that lives on trust, that is the expensive part.

The check that would have caught both#

None of this needs a compliance department. It needs one person who will open the paper before the ad goes live, and who has the standing to say no.

The economics favour doing it. Numan's second ad ran from 17 December 2025 to 21 April 2026, roughly four months of spend behind a claim that was later withdrawn and published against the brand. The first ran twenty-two days. Nobody costs a medical advertising review at four months of media, but that is the number it is competing against.

The patient prowl beats the fast pounce here. Trace each number to its own source, write down the population it describes, and keep that note beside the creative. It is the same discipline any healthcare marketing programme should already apply to a client claim, and the same trail the regulator will follow if a single complaint arrives.

Both brands were selling something legitimate to people with a real complaint. That is precisely why the copy mattered. In direct-to-consumer healthcare, the gap between what the evidence says and what the headline says is where the regulator lives, and where the trust goes to die.

Frequently asked questions#