There are approximately 350,000 health and fitness apps available across major app stores as of 2026. The vast majority make health claims with no published clinical evidence. A small fraction — perhaps 5% — have undergone rigorous testing in randomized controlled trials. An even smaller fraction have received FDA clearance as regulated medical devices. Understanding the difference between these categories matters enormously for anyone trying to use digital tools to genuinely improve their health.
The FDA's Digital Health Framework
The FDA categorizes health-related software into distinct regulatory buckets, and understanding these helps interpret what app makers claim:
Software as a Medical Device (SaMD): Software intended to perform medical device functions — diagnosing conditions, guiding treatment, predicting disease risk — that requires FDA review before marketing. The regulatory pathway depends on risk level: Class I (lowest risk, general controls), Class II (moderate risk, 510(k) premarket notification), or Class III (highest risk, Premarket Approval). Most cleared health apps use the 510(k) pathway.
General Wellness Apps: The FDA has stated it will exercise enforcement discretion (i.e., will not regulate) for software intended to maintain or encourage general health, provided it does not make specific clinical claims. This is where the vast majority of consumer health apps live — period trackers, meditation apps, nutrition trackers, step counters — unless they make specific diagnostic or therapeutic claims.
Clinical Decision Support (CDS): A category with complex regulations under the 21st Century Cures Act. CDS tools that display information for clinicians to review and act on are less strictly regulated than those that automate clinical decisions without requiring clinician judgment.
The practical implication: most consumer health apps exist in a regulatory gray zone. They can market to health-conscious consumers without clinical evidence, as long as they avoid specific diagnostic or therapeutic claims. "Track your sleep" is general wellness; "diagnoses sleep apnea" is a medical device claim requiring FDA clearance.
Apps with Genuine RCT Evidence
Noom (Weight Management): Noom's CBT-based weight loss program is the most clinically studied major consumer weight loss app. Multiple published trials support its effectiveness, with the most rigorous showing 5.4 kg average weight loss at 16 weeks versus 0.8 kg in control. Noom's approach is grounded in behavioral psychology — color-coded food systems, psychological coaching, habit formation — rather than rigid calorie counting. The behavioral mechanism is well-supported by independent evidence even if the specific Noom delivery vehicle continues to be studied.
Omada Health (Diabetes Prevention): Omada runs a fully digital version of the CDC-recognized Diabetes Prevention Program (DPP), which has one of the strongest evidence bases in preventive medicine. The original DPP trial showed 58% reduction in Type 2 diabetes incidence in high-risk individuals. Omada achieved CDC recognition as a supplier of the digital DPP and has published peer-reviewed data showing comparable outcomes to in-person DPP delivery. Insurance coverage for CDC-recognized DPP programs makes Omada reimbursable in many plans — a meaningful marker of real-world evidence acceptance.
Woebot (CBT-based Mental Health): Woebot is a conversational AI chatbot that delivers Cognitive Behavioral Therapy (CBT) techniques for depression and anxiety. A published RCT in JMIR Mental Health showed statistically significant reductions in depression symptoms (PHQ-9 score) and anxiety symptoms (GAD-7) compared to a control condition, with effect sizes (d ≈ 0.44) in the small-to-medium range. Multiple subsequent studies have replicated these findings. Woebot is not a substitute for clinical therapy, but evidence supports its efficacy for subclinical depression and anxiety in non-clinical populations.
Headspace (Mindfulness and Stress): Headspace has funded multiple published studies — a practice that appropriately raises questions about independence — but results have also been replicated by independent researchers. Meta-analyses of mindfulness app interventions (including but not limited to Headspace) consistently show small-to-medium effect sizes for stress, anxiety, and sleep outcomes, comparable to self-help book interventions and considerably smaller than in-person MBSR programs.
Digital Therapeutics: Prescription-Grade Apps
Digital therapeutics (DTx) represent the most rigorously regulated category of health apps. These are prescription-only software programs with FDA clearance for specific clinical indications:
Somryst (formerly Sleep Healthy Using the Internet — SHUTi): FDA-cleared for chronic insomnia, prescribable by physicians, and backed by multiple RCTs showing outcomes comparable to face-to-face CBT for insomnia (CBT-I). CBT-I is the first-line evidence-based treatment for insomnia per the American College of Physicians, and Somryst delivers it digitally. This is one of the clearest examples of a health app with genuine clinical-grade evidence.
EndeavorRx: FDA-authorized for attention deficit/hyperactivity disorder (ADHD) in children ages 8–12. It's a video game that uses specific sensory stimuli and motor challenges designed to target attention systems. The pivotal trial showed significant improvements on attention function assessments compared to a control condition. It's designed as an adjunct to, not replacement for, conventional ADHD treatment including behavioral therapy and medication.
Quit Genius / Pelago Health: Developed prescription digital therapeutics for substance use disorders including smoking cessation, alcohol use, and opioid use. Published multiple RCTs. Despite clinical evidence, Pear Therapeutics — a leader in prescription DTx — declared bankruptcy in 2023 after failing to achieve reimbursement at scale, illustrating that clinical evidence doesn't automatically translate to commercial viability when payer systems haven't adapted.
Apps to Approach With Skepticism
The wellness app market is overwhelmingly populated by products with no clinical evidence for their specific claims:
Most "AI health coach" apps: Despite sophisticated interfaces, the vast majority have no published validation of their recommendations. Algorithms that claim to predict your "biological age," recommend supplements, or provide personalized nutrition plans based on genetic data typically have no peer-reviewed evidence that the specific output improves health outcomes.
Most bioresonance, frequency therapy, and energy medicine apps: These operate on mechanisms with no scientific basis and no clinical evidence. Red flags include claims based on quantum physics applied to biology, "frequencies that heal," or "cellular resonance." These categories are not legitimate medical technologies regardless of how sophisticated the app interface appears.
Most sleep coaching apps without hardware: Apps that claim to improve sleep quality without hardware sensors (EEG, PPG, or at minimum accelerometer) have no objective measure of whether they're working. Subjective improvement from sleep hygiene reminders is real — but no different from following a printed sleep hygiene checklist.
Mental Health Apps: A Complex Landscape
Mental health apps occupy a particularly complex position — the need is enormous (global mental health treatment gaps exceed 70% in most countries), the potential is clear, and the risks of ineffective or harmful apps are significant. Key findings:
Apps delivering structured CBT techniques (Woebot, Wysa, MoodKit) have the strongest evidence base. Apps primarily delivering mindfulness content (Headspace, Calm, Ten Percent Happier) have moderate evidence for stress and sleep but weaker evidence for clinical depression and anxiety. Apps that primarily offer community support without therapeutic structure (most peer support apps) have minimal clinical evidence despite high user satisfaction ratings.
The American Psychological Association's AppAdviser database and the M-Health Index and Navigation Database (MIND) at Massachusetts General Hospital provide clinician-oriented evaluations of mental health apps based on evidence quality, privacy, usability, and safety. These are better resources than app store ratings for identifying evidence-based options.
How to Evaluate Any Health App's Claims
A practical decision framework:
- Search PubMed: App name + "randomized controlled trial" or "clinical trial." No results after five minutes of searching is itself meaningful data.
- Check ClinicalTrials.gov: Legitimate RCTs are pre-registered. A company claiming trial evidence with no registered trials is a major red flag.
- Examine conflict of interest: Who funded the published studies? Company-funded trials report 30–40% larger effect sizes on average than independent replications.
- Look for FDA clearance: Not a guarantee of efficacy, but indicates the company submitted performance data to a regulatory body with medical expertise.
- Evaluate the mechanism: Does the app's proposed mechanism of action align with established science? "Improves sleep by training breathing" is plausible; "balances chakra frequencies to optimize cellular function" is not.
- Privacy assessment: Health apps collect sensitive data. Check whether the app's privacy policy allows selling health data to third parties — a concerning practice that several major period tracking apps engaged in.
The realistic conclusion: in a market of 350,000 health apps, genuine evidence-based options are rare but do exist. For the clinical indications where evidence-based digital health exists — chronic insomnia, diabetes prevention, obesity management, mild-to-moderate anxiety and depression, ADHD symptom management — digital tools can meaningfully extend access to effective interventions. Approaching the rest of the market with appropriate skepticism protects both your health and your data.