Wearable health tech needs proof, not promises

There has never been more health information gathered from the human body than there is today and rarely more confusion about what to do with it. More than 150 million smartwatches were sold worldwide last year, a great many of them now marketed on a seductive promise – that the device on your wrist will catch disease before it catches you. As someone who has spent a career researching and creating wearable medical devices, I find this moment genuinely exciting and quietly worrying in equal measure. The promise is real. It is also being oversold, and the gap between the two is where patients could get hurt.

Let me start with the uncomfortable part, because most writing on this subject skips it. The evidence that consumer wearables actually make people healthier is far thinner than the marketing implies. A 2025 scoping review in PLOS Digital Health examined eighty studies of wearables used for health monitoring outside hospital settings and found that only 8% were randomised controlled trials, the rest were observational or feasibility work. Its conclusion was blunt: evidence for the clinical effectiveness of these devices remains scant. That is the crucial distinction. We have abundant proof that wearables can collect data and far less that acting on it improves the health of an otherwise healthy person. A device that measures something accurately has cleared a much lower bar than a device that measurably makes you better off and those two things are routinely conflated.

The harms, meanwhile, are not hypothetical. Continuous monitoring generates a steady stream of readings and even a technically excellent algorithm processing many times a day will produce false alarms that send people who are well to A&E for tests they never needed. There is a quieter harm too, the anxiety of being handed a constant feed of data about your own body and reading dangers into normal fluctuation. Atrial fibrillation is the sharpest example. A common heart-rhythm disorder that wearables are increasingly good at detecting, it is exactly the kind of condition where the technology works as intended and still does harm. As a 2025 review in Current Cardiology Reports describes, giving patients on-demand access to their own heart-rhythm data can, in susceptible people, provoke intense cardiac fear and obsessive symptom monitoring. The device detects the very thing it was built to detect, but what it was never validated against is the psychological cost of placing that stream of data in an anxious person’s hand. This is a real harm and it falls hardest on the anxious and the well.

The evidence that consumer wearables actually make people healthier is far thinner than the marketing implies.

None of which means wearables are worthless – far from it. So which is it, then: the technology that catches disease early or the technology that manufactures anxiety? The honest answer is that the question is badly posed, because “wearable” is not one thing. A continuous glucose monitor prescribed after a diabetes diagnosis and a smartwatch that offers to read your blood pressure fall under the same word, yet what stands behind them could hardly be more different. Faced with that variety, the natural instinct is to look for a badge of trust, and the one everyone reaches for is whether a device is “validated.” Unfortunately, that word, on its own, now tells you almost nothing. Nearly every device on the market has some form of validation, but what matters is what that validation consists of. What was it validated against, for what purpose, in which population and what claims does it support? A wearable can be honestly validated to track blood-pressure trends – a real piece of engineering – without ever having been validated to diagnose high blood pressure, and the careful wording on the box is chosen precisely to blur the two. On a screen, the outputs look alike. Clinically, they mean entirely different things.

Why this matters is easy to see. Picture a person at elevated stroke risk whose smartwatch tells them their blood pressure looks “fine,” and who is reassured by a device that was never validated as a diagnostic tool and was marketed, quite deliberately, around that limitation. A falsely reassuring “normal” can delay a diagnosis as surely as a false alarm can manufacture one. When the validation behind a number does not match the use someone is putting it to, that number is worse than no number at all, because it carries an authority it has not earned. And an ordinary person has almost no way of telling the two apart, because the box is written precisely so they can’t. This is why clear guidance matters – the kind provided by bodies like NICE, the UK’s official adviser on which health technologies are fit for clinical use. Manufacturers are extremely sophisticated at navigating regulatory language, so without an external referee saying which devices, for which purposes, actually clear the bar, that judgement falls to the one person least equipped to make it, the user holding the wrist.

Once you apply that lens – what is the validation precisely – the genuine promise of this technology comes into focus and it is not where the hype points. It is not in nudging the worried well toward marginally better habits. It lies in the harder clinical work – screening people who would otherwise be missed, reaching a diagnosis without a wait that runs through scarce specialists and equipment and helping people manage a known condition in ways that were never feasible when they depended on a clinician watching over them. The continuous glucose monitor – a small wearable sensor that tracks blood sugar in real time – is the obvious success on that last front. It lets people with diabetes see the consequences of every meal and adjust immediately, something no periodic clinic visit could ever deliver. But the gain can be just as large earlier on, in reaching people with a serious, treatable condition, the health system has simply never had the capacity to catch. The example I know best, because I have spent years working on it, is obstructive sleep apnoea – a disorder in which breathing repeatedly stops and starts during sleep. It affects nearly a billion adults worldwide, the majority of them undiagnosed – not because the disease is poorly understood or the treatments unavailable, but because the diagnostic pathway physically cannot scale. Conventional diagnosis requires multi-sensor equipment applied by trained staff, specialist scoring of the results, and waiting times that in many systems stretch to months or years. Untreated, the condition carries a substantially raised risk of hypertension, stroke and dementia and a several-fold increase in road-traffic accidents from daytime sleepiness. The bottleneck is structural, not scientific.

The clinical talent inside the NHS is world-class, but the machinery for buying and adopting new tools was never built for the speed at which this field moves.

This is precisely the kind of problem a properly validated wearable is built to solve and it is the work my own company has taken on. AcuPebble, the wearable that grew out of my research at Imperial College London, is worn on the neck overnight, in place of the tangle of equipment a conventional sleep test straps across the body. It was designed to be simple enough to send directly to patients, who set it up themselves at home, and small enough not to disturb their sleep. It records and analyses physiological signals to automatically extract biomarkers and produce an automated diagnosis of obstructive sleep apnoea, with no manual scoring and to the same standard as the multi-sensor test used in hospitals. Because the diagnosis is automatic, the clinician has the result immediately and treatment need not be delayed. The reason I offer it as an example is not that it is ours, but that it illustrates what the right kind of validation actually looks like. It was validated for that specific diagnostic task, in the relevant patient population and in December 2024, NICE incorporated it into a newly recommended national diagnostic pathway – a redefinition of how the condition is diagnosed, not merely an alternative within the old framework. That depth of validation is what lets diagnosis leave the hospital without the clinical standard leaving with it.

Get this right and the gains ripple outward. When a diagnosis no longer depends on a hospital appointment, the patient is spared the journey and the day off work it used to cost them and, more importantly, the wait itself shrinks, because the test is no longer trapped behind a queue for scarce equipment and staff. Specialists, freed from routine testing, can give their time to the complicated cases that actually need them and to treat patients who need it quicker. Where this new pathway has been adopted, most testing now happens at home, and waiting lists that once ran to months, or even years, have come down to weeks. That is not just an administrative tidy-up. Each of those patients is someone who might otherwise have gone undiagnosed for a very long time- their risk of a stroke or a road accident quietly climbing the whole time. Diagnosing them sooner is prevention in the truest sense.

When a diagnosis no longer depends on a hospital appointment, the patient is spared the journey and the day off work it used to cost them and, more importantly, the wait itself shrinks, because the test is no longer trapped behind a queue for scarce equipment and staff.

The same pathway happens to be lighter on the planet, too – and that is not a coincidence, but part of the same design. Testing at home strips out hospital visits and all the travel they entail across an entire population, and when the device itself is built to be reusable and efficient, the savings compound. As an engineer, this matters to me. Reusable, efficient design is not a virtuous extra bolted onto the clinical case, it is part of the same case. And the benefits are now starting to be measured. This year, the NHS Sustainable Healthcare Award went to Imperial College Healthcare NHS Trust for its wearable-enabled home-based sleep apnoea service, which reported cutting plastic waste by 99% and saving an estimated 12.6 tonnes of CO2 a year, while replacing older equipment shown to give less accurate readings on darker skin. That is one example; across the health system, this kind of evidence is only beginning to be gathered.

So here is a technology that widens access, shortens waits, eases pressure on specialists and treads more lightly on the planet – and still the system struggles to take it up. The clinical talent inside the NHS is world-class, but the machinery for buying and adopting new tools was never built for the speed at which this field moves. A device can be proven, useful and still take years to reach patients, by which time the next generation of it already exists. Hospitals are often tied into procurement frameworks that leave little room to consider anything new, and each trust tends to repeat evaluations and bureaucratic steps that a dozen others might have already done. That lag holds back innovation of every kind, but it bites hardest on fast-moving technology like this. The obstacle is no longer the science, and it is not a lack of clinical will – it is the system around adoption.

Fixing that is worth the effort, because none of this is really a question about wearables at all. It is a question about discernment, about our willingness to look past a glossy screen and a reassuring word on a box and ask what a device has actually been proven to do, for whom, and against what standard. Get that right and the reward is not a gadget that counts our steps a little more precisely, it is a person who would have spent years tired, untreated and slowly getting sicker, diagnosed instead in a week, from their own bed. That is the kind of thing wearable technology can open the door to. The task now is to build the judgement – and the systems – equal to it.