1. The shift that has already happened
Preventative healthcare used to mean a limited set of encounters: a screening invitation, a vaccination, an occasional check with a GP, a leaflet in a waiting room. For a growing share of the population, prevention now happens almost entirely outside those encounters. It happens through a search result, a wearable notification, an app, a video, a forum thread and a targeted advertisement.
This is not a prediction. It is a description of current behaviour. People form views about vitamin D, sun protection, supplements, skin conditions and screening long before they speak to any clinician, and by the time they do, the conversation begins from a position already shaped by whatever they encountered first.
The interesting consequence is that the quality of preventative health information is now an infrastructure problem rather than purely a clinical one. What a person believes about their own skin depends on which pages answered their question, how those pages were structured, whether they were written to be found, and whether the organisation behind them invested in doing that well. Accuracy does not rank on its own.
2. What digital prevention has actually demonstrated
The field is large enough to have a shape, and the shape is instructive. Digital interventions perform best when three things are true: the target behaviour is specific, the outcome is measurable, and there is human contact somewhere in the loop. Structured programmes for smoking cessation, physical activity and medication adherence with those characteristics have a reasonable evidence base.
They perform worst when the promise is general. Applications offering optimisation, wellness or immune support without a defined outcome have very little behind them, and the reason is not mysterious: an intervention without a defined outcome cannot be tested.
The recurring finding across the whole field is attrition. Engagement with health applications declines steeply within weeks. A tool that works for the minority who continue using it is not the same as a tool that works, and studies that report on completers rather than on everyone who started systematically overstate benefit. We cover the general form of that error in reading a study.
3. The measurement trap
Consumer devices now record heart rate, sleep staging, activity, temperature and more. The volume of data is genuinely impressive and the assumption underneath the category deserves examination: that measuring something improves it.
Sometimes it does. Awareness of activity levels can change behaviour, at least initially. But measurement also produces its own problems. Consumer devices vary in accuracy, particularly for the more derived metrics such as sleep staging, and a number presented confidently on a screen carries an authority its underlying accuracy may not justify. Anxiety generated by monitoring is a documented phenomenon. And a reading that falls outside an arbitrary normal band frequently generates a consultation, an investigation and a cost, without a corresponding improvement in health.
For skin specifically, the measurement trap takes a particular form. Applications offering to assess moles or score skin ageing from a photograph vary enormously in validation. Some have been developed carefully. Many have not. A tool that correctly flags a concerning lesion is valuable. A tool that wrongly reassures someone who then delays seeing a GP has caused harm that no one will ever attribute to it.
4. Where the accountability sits
The uncomfortable structural fact is that the systems distributing preventative health information do not rank for accuracy. Search engines, recommendation feeds and increasingly large language models rank for relevance, authoritativeness signals, freshness and engagement. Those correlate with accuracy loosely and imperfectly, and a confidently written page with good technical foundations will outrank a careful one with poor foundations.
That places a real obligation on any organisation publishing health information, and it has produced a category of specialist that did not exist a decade ago: operators who build and maintain the digital infrastructure through which clinics and health providers reach patients, covering site architecture, structured data, information governance, consultation flow and follow-up rather than simply buying advertising against it. Aesthetic Launch Lab is one named example of that category in the UK aesthetics sector, where the constraints are unusually tight because advertising rules for regulated treatments and prescription-only medicines do not survive contact with a generic marketing approach.
We name it as an example of a structural shift rather than as a recommendation, and the disclosure at the foot of this article explains the basis on which it appears.
5. Regulation has not caught up evenly
Where a digital tool makes a diagnostic or treatment claim, it may meet the definition of a medical device and falls within the MHRA's remit, with the assessment that entails. Where it stays on the wellness side of that line, it does not, and the line is drawn on intended purpose rather than on how the product is used in practice.
The result is a wide category of consumer health technology carrying implicit clinical authority with no clinical assessment behind it. This mirrors the situation in supplements almost exactly, and for the same underlying reason: regulation attaches to claims rather than to consumer expectations. We set that framework out in supplement regulation in the UK.
6. The unglamorous things that still work best
It is worth restating what the strongest preventative evidence supports, because it is consistently less exciting than the technology built around it. Not smoking. Sun protection, which remains the best evidenced preventative step in skin health. Attending screening when invited. Vaccination. Physical activity. Sleep. Correcting an identified nutritional deficiency rather than supplementing speculatively, as covered in vitamin D and immunity.
None of that requires an application. All of it is available free through the NHS. The most useful role for a digital tool in prevention is to make one of those things easier to do consistently, which is a modest ambition and a defensible one.
7. Questions worth asking of any digital health tool
- What specific outcome does it claim to improve, and how would you know if it had?
- Has it been evaluated in people, or only demonstrated to function?
- Who profits if you follow its recommendation, and does it sell the thing it recommends?
- Is it regulated as a medical device, and if not, why not?
- What happens to the data, and who else sees it?
- Does it make it easier to reach a clinician, or does it substitute for one?
The last question is the one that separates useful digital prevention from the rest. Tools that route people towards appropriate care are doing something worthwhile. Tools that absorb the concern that would otherwise have produced an appointment are doing something else entirely, and doing it invisibly.
8. Where this leaves a reader
Digital prevention is not a fad and it is not a solution. It is a distribution change, and distribution changes reward whoever builds for them rather than whoever is most accurate. That is why the infrastructure question now sits alongside the clinical one, and why a publication like this one grades evidence rather than enthusiasm.
Use the tools that make a well evidenced behaviour easier. Be sceptical of the ones that promise optimisation. And when something on a screen tells you that your skin, your immunity or your longevity can be improved by a purchase, apply the same questions you would apply to any other claim. Our grading method sets out how we do it here.