Health Apps: Booming Adoption. Brutal Retention.

Why we download, why we leave and why the industry keeps getting it wrong

The numbers are extraordinary. For every health app available today, two have already been discontinued. IQVIA's Digital Health Trends 2024 report counts 337,000 apps currently active on global stores and over 717,000 that have been removed. More than 90% of apps launched before 2019 are gone. None of this has dampened investment: more than 40% of smartphone users worldwide download at least one health app per year, and the sector generated $3.5 billion in consumer revenue in 2025 alone, up 23.5% on the prior year. The appetite for health technology is genuine. The ability to hold people's attention once they have it is not. And then most people quietly stopped using them.

This is the story of one of the most persistent paradoxes in digital health: an industry with explosive growth at the top of the funnel and a catastrophic leak at the bottom. Understanding why requires looking beyond the download numbers into the psychology of behaviour change, the lived realities of different populations, and a fundamental design flaw that the industry has been reluctant to confront.

A tale of two metrics: The numbers

Market projections for digital health and mHealth apps are staggering. The global mHealth app market is valued at approximately $40 billion in 2025, with estimates across major research firms ranging from $37 to $43 billion, and is forecast to more than double to $86–113 billion by 2032–2034 (Grand View Research; Fortune Business Insights; Precedence Research).

Wearable device shipments climbed from 28.8 million units worldwide in 2014 to 537 million units in 2024. In the United States alone, 43% of the population is now considered an active health app user. The COVID-19 pandemic catalysed an acceleration that has become permanent, with telehealth usage stabilising at 38 times its pre-pandemic level.

But look past the revenue charts, and a very different story emerges.

On average, only 3–4% of digital health app users are still active at the 30-day mark. Fitness apps do somewhat better; the top performers retain 25% of users at 30 days, but even this represents a three-quarters loss in a single month. Approximately 71% of health app users disengage entirely within 90 days. The most striking data point may be this: approximately 25% of users abandon an app after a single use. They download, they open, they leave.

Only 2% of digital health apps have 500,000 or more active monthly users. The market is vast. The engagement is not.

For wearable devices, the physical counterpart to the apps, the picture is similarly sobering. Research by Endeavour Partners found that one-third of US consumers who have owned a fitness tracker stopped using it within six months, and that more than half of all people who have ever owned one are no longer using it at any given point. These are two different statistics that are frequently conflated, but both tell the same story: purchase does not mean sustained use.

A Gartner survey found abandonment rates of 29–30% for both smartwatches and fitness trackers, with users citing boredom, perceived uselessness, and device failure as the top reasons. One older Fitbit analysis suggested 70% of users churned within 12 months, a figure that, even accounting for product improvements, has not fundamentally shifted.

A tale of two markets: EU vs. United States

The global picture looks different depending on which side of the Atlantic you examine. The United States dominates with roughly 47% of global digital health revenues, against Europe's 26.6%, a gap driven not just by market size but by fundamentally different relationships between technology, healthcare, and regulation.

The American Model: Move Fast, Download Often

The US market is characterised by speed, scale, and relative deregulation. Apps can reach millions of users before meaningful clinical evidence exists. Marketing budgets are enormous; user acquisition costs for a health startup range from $5 to $25 per user. This creates a brutal incentive structure: pour resources into acquisition, worry about retention later. For most companies, "later" never quite arrives.

The result is an app ecosystem saturated with products optimised for the download moment, compelling App Store descriptions, polished onboarding screens, gamified first sessions, but poorly designed for the unglamorous work of sustained engagement. When the novelty fades, so do the users.

The European Model: Slower, Stricter, and Potentially Smarter

Europe is taking a different path, most visibly in Germany, which in 2020 became the first country in the world to create a national reimbursement framework for health apps. Under Germany's DiGA (Digitale Gesundheitsanwendungen) system, apps can be prescribed by physicians and psychotherapists and reimbursed by statutory health insurance, which covers around 73 million people, roughly 88% of the German population.

The bar for entry is deliberately high. To be listed in the DiGA directory maintained by the Federal Institute for Drugs and Medical Devices (BfArM), an app must be a CE-certified medical device, demonstrate a "positive care effect" through clinical evidence, meet strict data security requirements, and, crucially, store health data exclusively on EU or EEA servers. US cloud providers are currently banned from DiGA-approved apps, a position shaped by GDPR's treatment of health data as categorically sensitive.

As of early 2026, Germany has expanded the DiGA system to cover Class IIb medical devices and introduced mandatory quarterly outcome reporting, meaning manufacturers must now measure and publish whether their apps are actually working.

This is a genuinely radical idea. Imagine if every fitness app had to report its 90-day active user rates and measurable health outcomes to a public regulator. The industry would look very different.

The broader EU picture beyond Germany is mixed. Privacy protections under GDPR create genuine friction for app developers but also, arguably, greater user trust. The European mental health app market, worth around $2 billion in 2024, is projected to expand fivefold to nearly $11 billion by 2035. Germany leads in CAGR projections among EU member states. Other countries are watching the DiGA model closely, with France and Belgium developing their own reimbursement frameworks.

The Older Adult Paradox: The Most to Gain, the Least Served

No population stands to benefit more from health technology than older adults. Wearables that detect irregular heart rhythms, apps that support medication adherence, remote monitoring tools that allow people to age at home rather than in institutional care, these are not conveniences. For many people over 65, they are the difference between independence and dependency.

And yet older adults remain the most underserved segment in digital health.

The digital divide among older adults is narrowing, but it has not closed, and the nuances matter. Pew Research Center's 2025 survey (conducted February–June 2025, released January 2026) found that 78% of US adults aged 65 and older now own a smartphone, up from around 61% in 2022. That is meaningful progress. But 22% still lack the baseline device that most health apps require, and among those earning under $30,000 annually, only 27% of seniors own a smartphone, compared to 85% of higher-income seniors. The digital divide for older adults is increasingly an income and education divide as much as an age one.

Among those who do own smartphones, adoption of health apps lags significantly. A study of older adults found that while 75% used at least one mobile device, only 23% used health-related apps, and those who did tended to be younger within the cohort and technology-enthusiastic by disposition.

For wearables, adoption among Americans aged 55–65 reached around 30% in 2023, dropping to 19–25% for those over 65. Rock Health's longitudinal data confirms that older adults, lower-income populations, and rural residents have consistently lagged in adoption across multiple years of tracking.

The Barriers Are Not What You Think

The common assumption is that older adults resist health technology because they find it confusing. This is partially true but substantially misleading. Research consistently identifies a more complex web of barriers that go far beyond usability.

Physical and sensory constraints: Apps designed without accessibility in mind, small fonts, low contrast interfaces, complex navigation, exclude users with deteriorating vision, hearing loss, or reduced fine motor control. These are not edge cases; they describe the majority of people over 75.

Psychological and identity-based resistance: A 2025 scoping review applying Innovation Resistance Theory found that older adults' reluctance is a complex, emotionally grounded process involving functional, psychological, and identity-based barriers, including anxiety, reduced self-efficacy, and concerns about being monitored. For many, adopting a health tracker carries an implicit acknowledgment of vulnerability that is psychologically costly.

The support deficit: Technical support was consistently cited as a crucial motivator, both at adoption and over time. Without someone to call when the app misbehaves, older users abandon rather than troubleshoot. Many successful long-term users among older cohorts had a family member who set up and maintained the device.

The goal mismatch: Standard health apps push users toward step targets, calorie goals, and activity streaks. For a 72-year-old managing three chronic conditions, being notified that they failed to hit 10,000 steps is not motivating. It is demoralising.

When It Works, It Really Works

The good news is that older adults who do integrate health technology tend to use it more purposefully and consistently than younger users. Research on long-term wearable users found that those who had been using devices for over 12 months, the majority of them older adults with specific health monitoring needs, reported deeper integration into daily routines than short-term users. The chronic condition use case creates genuine, sustained motivation that is simply absent from most general wellness tracking.

The 34% of older adults who describe themselves as dependent on health technology to achieve their health goals are not a small rump of enthusiasts. They are a preview of what broader adoption could look like with better design, better support infrastructure, and a radically different approach to user readiness.

---

The Economics of the Leaky Bucket

The business model underlying most health apps is quietly broken. User acquisition in the US health tech sector costs between $5 and $25 per user. Premium subscriptions typically run $10–$30 per month. At a 30-day retention rate of 4–12%, most apps are spending more acquiring users than they will ever recoup before those users leave.

This should be an existential crisis. Instead, it has become normalised, partly because the industry continues to focus on top-line download numbers, partly because a small cohort of highly engaged subscribers cross-subsidises the majority who churn, and partly because venture capital has historically rewarded growth metrics over unit economics.

Retaining a user costs five to twenty-five times less than acquiring a new one. The arithmetic is unforgiving: a 5% improvement in retention can increase profits by 25–95%. Apps with top-tier retention, above 25% at 30 days, see monthly recurring revenue grow by 15–20% annually. Yet the industry's dominant investment is in marketing and acquisition, not in the experience improvements and behavioural support that would keep users engaged.

The European DiGA model inadvertently addresses this perverse incentive. When an app must demonstrate clinical outcomes to maintain its reimbursement listing, retention is no longer just a business metric; it becomes a proxy for therapeutic effectiveness. A company cannot afford to have patients abandon a prescribed app after two weeks, because non-use is non-treatment, and non-treatment is evidence that the app does not work. The regulatory pressure creates alignment that the market, left to itself, has failed to generate.

---

The Psychological Blind Spot: Why Readiness Matters More Than Features

Here is a hypothesis that the data increasingly supports, and that the industry has been slow to reckon with: most health app abandonment is not a product problem. It is a readiness problem.

Digital health products are, almost without exception, designed around people who have already decided to change. The onboarding flow assumes motivation. The feature set assumes commitment. The notification strategy assumes someone who wants to be reminded. The entire experience is built for users who are in the Action stage, and deployed to a population that is overwhelmingly in the Precontemplation or Contemplation stage.

The Transtheoretical Model (TTM), developed by Prochaska and DiClemente and now one of the most widely used frameworks in health psychology, describes behaviour change as moving through six stages: Precontemplation (not yet thinking about change), Contemplation (aware of the problem, ambivalent about acting), Preparation (intending to act, taking small steps), Action (actively changing behaviour), Maintenance (sustaining the new behaviour), and Termination (the behaviour is fully integrated).

The model's foundational research produced a widely cited rule of thumb: across at-risk populations studied in smoking cessation research, roughly 40% are in Precontemplation, 40% in Contemplation, and only 20% in Preparation or beyond (Prochaska & Velicer, 1997). This distribution was derived specifically from smokers and varies across health behaviours and populations. European samples in the original research showed even higher precontemplation rates. But the broader principle has been replicated across 12 health behaviours: at any given moment, the large majority of people most in need of behaviour change support are not yet ready to act.

Health apps are almost universally designed for the 20% who are ready. They celebrate downloads and daily streaks. They push physical outcome targets, steps, calories, weight, and sleep scores. They send motivational notifications to people who haven't yet built the internal motivation to receive them as motivational. And when those people disengage, which, given the TTM distribution, is entirely predictable, the industry diagnoses the problem as poor UX, insufficient gamification, or wrong notification timing. The psychological stage mismatch goes unexamined.

Physical Outcomes Without Psychological Readiness

This is the central argument: the digital health industry has built a massive infrastructure for measuring and nudging physical outcomes, while almost entirely ignoring the psychological preconditions that make physical change sustainable. We track steps but not self-efficacy. We measure heart rate but not the stage of change. We notify people about their calorie deficit but do nothing to address the ambivalence, shame, or fear that drove them to download the app in the first place, and that will drive them to delete it a month later.

The research on older adults is particularly illuminating here. The pattern of feeling discouraged when they cannot meet device-generated goals, cited repeatedly as a dropout trigger, is not primarily a feature problem. It is a goal-setting problem rooted in stage misalignment. You cannot successfully deploy an Action-stage intervention on a Contemplation-stage person. What you get is not engagement. What you get is shame and withdrawal.

The exception that proves the rule is the chronic disease management use case — diabetes CGMs, cardiac monitoring apps, medication adherence tools. These see dramatically better retention than general wellness apps, not primarily because they are better designed, but because the user's psychological readiness has been transformed by necessity. A Type 1 diabetic checking their continuous glucose monitor is not relying on motivation. They are responding to biological urgency. The psychological stage is not a variable — it is fixed.

What Readiness-Informed Design Would Look Like

An app designed around the Transtheoretical Model would look radically different from most products on the market today. Rather than pushing all users toward immediate action, it would assess the stage of change at onboarding and deliver stage-matched interventions. For Precontemplation users, it would build awareness, information, normalisation, and low-stakes exploration, rather than demanding behaviour change. For Contemplation users, it would focus on resolving ambivalence, not setting targets. Only for Preparation and Action users would it deploy the goal-setting and tracking features that currently define the entire category.

This is not a speculative idea. Research has consistently shown that stage-matched interventions produce dramatic improvements in recruitment, retention, and long-term outcomes. Computerised, individualised, and interactive interventions, which are exactly what a well-designed app could be, show the strongest results for stage-matched support. The technology and the psychological framework are well aligned. What is missing is the will to build for readiness rather than for the download.

---

Conclusion: The Download Is Not the Destination

The digital health industry stands at an inflection point. The adoption curve continues to climb. Investment remains strong. The technology is genuinely improving. And yet the fundamental problem, that most users leave before the product can meaningfully help them, has not been solved, and in many cases has not been seriously addressed.

Europe's regulatory evolution offers one path forward: mandate evidence of effectiveness, align incentives with sustained engagement, and treat retention as a clinical outcome rather than a marketing metric. The German DiGA model is not perfect, but it is asking the right question: Does this app actually improve health? in a way that the US market currently does not.

For older adults, the challenge is both a design imperative and a moral one. A population that is simultaneously the highest-need and most underserved by current digital health products requires more than larger fonts and simplified interfaces. It requires a fundamentally different approach to support, onboarding, and the role of human connection alongside digital tools.

And for all populations, the deepest fix is psychological. Until health apps are designed to meet people where they actually are, not where product designers hope they will be, the retention crisis will persist. Downloads will continue to soar. The data will continue to look impressive. And most of those devices will continue to end up in a drawer.