Digital Health Is Splitting Into Two Markets

A lot of digital health products look modern. Fewer are built for the market they actually want to win. That is the problem.

Many teams still talk about digital health as if it is one category. It is not. The market is splitting into two very different businesses. One is built around consumer wellness, engagement and low-friction behaviour change. The other is built around regulated care, measurable outcomes and evidence that can stand up to scrutiny. The FDA now describes digital health technologies as spanning “from applications in general wellness to applications as a medical device”. That is not semantic nuance. It is the fault line shaping the next phase of the sector.

The old category has stopped being useful

For years, digital health was a catch-all label. Sleep apps, meditation tools, wearables, symptom trackers, remote monitoring platforms and clinical software were all discussed as if they belonged to the same market.

They do not. The FDA’s own framework makes that clear. Some digital health technologies are general wellness tools. Others are medical devices. Some may support care without crossing that line. Others sit directly inside diagnosis, treatment, monitoring or clinical management. The category is no longer one market with different features. It is two markets with different rules.

The real dividing line is not the interface. It is the claim

Here is the counter-intuitive part. The biggest strategic decision in digital health is often not the technology. It is the sentence on your homepage.

If your product helps people build healthier habits, reflect on sleep, move more or improve general wellbeing, you are likely in the wellness market. The FDA’s January 2026 guidance is explicit that software intended for maintaining or encouraging a healthy lifestyle, when unrelated to the diagnosis, cure, mitigation, prevention or treatment of a disease or condition, is not a device under that framework.

The moment your language shifts towards disease management, condition monitoring, treatment support or medical decision-making, the ground changes. FDA device classification depends on intended use and indications for use, and those can be conveyed in labelling and even oral or promotional claims. In other words, many products do not cross into a new market because of what they do. They cross because of what they say they do.

A simple framework: the 3-question market test

Most teams classify themselves by feature list. That is too shallow. A better test is this:

1. What are you promising?

Are you promising healthier behaviour, or are you promising a clinically meaningful result?

2. What happens if your output is wrong?

If an inaccurate output is mildly annoying, that is one market. If it influences patient care, clinician judgement or treatment pathways, that is another.

3. What kind of proof will your customers eventually need?

Engagement screenshots and testimonials work in one category. Structured evidence, traceable data and performance validation matter in the other.

This is where many teams get stuck. They design like consumer software, then realise later they want to sell into care delivery, reimbursement pathways or regulated environments. By then, the product architecture is often pointing in the wrong direction.

The second market is getting more serious, not less

This is not theory. The policy signals are getting sharper.

The FDA’s Digital Health Center of Excellence now sits at the centre of a broader regulatory effort covering software as a medical device, device software functions, sensor-based digital health technology and real-world evaluation. On 3 February 2026, the FDA said it had launched the TEMPO pilot, linked to the CMS ACCESS model, to promote access to certain digital health devices while safeguarding patient safety. The agency’s stated objectives include advancing science and evidence for digital health technologies and creating a more tailored regulatory paradigm for them.

That matters because it signals something bigger than oversight. It signals market formation. Regulators do not build pilots, guidance frameworks and evidence programmes around categories they consider trivial.

Evidence is becoming part of the product

This is the shift many founders still underestimate.

In the first wave of digital health, data was mostly presented as a feature. Dashboards. trends. alerts. engagement loops.

In the next wave, data is becoming part of the product’s credibility.

On 2 April 2026, the FDA published a new set of 73 examples of marketing authorisations using real-world evidence from fiscal years 2020 to 2025. The agency said appropriately validated RWE can support a broad range of regulatory activities, including serving as the primary clinical evidence in premarket submissions. It also pointed to a wider mix of evidence sources, including electronic health records, device-generated data, and digital health platforms. That is a material change.

It means the question is no longer whether a product collects data. Plenty do. The more important question is whether those data are relevant, reliable and structured well enough to support a decision that matters. That is a much higher bar.

Advanced tip: stop asking whether your product is “AI-enabled”

That is the wrong question. The better question is whether your outputs can survive an evidence review.

The FDA has already warned that advanced algorithms in digital health technologies can be susceptible to error or bias, leading to malfunction or misinterpretation of health data. It specifically points to the need for regulatory science tools and methods that protect data integrity, promote health equity and improve reliability. That means the future advantage is not having AI in the workflow. It is having outputs that can be trusted in the workflow.

Why this matters for strategy right now

The biggest mistake in digital health today is category confusion.

Some products are being built like wellness tools while quietly hoping to become part of clinical care later. Others make quasi-clinical claims without the evidence systems to support them. Some collect high volumes of user data without any structure that would make those data useful in regulated or reimbursement settings.

That middle ground is getting harder to defend. The cleaner strategy is to choose your side early.

If you are building for wellness, stay disciplined. Build for habit change, retention and user experience. Keep claims narrow. Do it well.

If you are building for regulated care, accept the trade-offs. Stronger claims require stronger proof. Better data architecture matters. Intended use matters. Clinical workflow matters. Longitudinal evidence matters.

That is where infrastructure-first thinking starts to separate itself. Not in the interface, but in the system underneath it. It is also why platforms working closer to evidence generation and structured care pathways, including OnTracka’s broader philosophy around longitudinal data and defensible monitoring, sit in a different part of the market than general health apps.

The conclusion most people will resist

Digital health is still growing. But it is no longer growing as one market.

One side will continue to reward engagement, convenience and consumer trust. The other will reward evidence, accountability and clinical fit. Both can be valuable. But they are not interchangeable, and they should not be built with the same assumptions.

The teams that win the next phase will not be the ones asking, “How do we look more like a health app?”

They will be the ones asking the harder question.

Are we building for attention, or are we building for evidence?