Why the “wellness vs device” line matters
For companies building wearables and health apps, the distinction between general wellness products and medical devices is not academic. It determines whether a product can launch quickly with minimal regulatory friction, or whether it triggers years of validation, documentation, and FDA interaction. In 2026, this line matters more than ever because wearables increasingly sit close to clinical use, often without intending to cross that boundary.
From a business perspective, misclassification carries asymmetric risk. A product positioned as general wellness can reach the market rapidly and iterate frequently. The same product, if deemed a medical device, may require premarket submissions, quality system compliance, clinical evidence, and post-market surveillance. Reclassification late in development or after launch can lead to costly remediation, forced claim changes, or enforcement action. Still, cost and timing are only part of the equation. Regulatory status also shapes liability exposure, investor expectations, and partnership options with healthcare systems. Hospitals, payers, and regulated digital health platforms increasingly ask vendors to justify why their product is not a medical device. At the same time, app stores and advertising platforms scrutinize medical claims more aggressively, amplifying downstream consequences of FDA misalignment.
In 2026, the risk profile is heightened by AI-enabled personalization. Even products originally designed for lifestyle use may drift toward medical inference through individualized insights, alerts, or predictive features. The FDA has made clear that intent is inferred not just from what developers say, but from what the product reasonably appears to do in practice. Understanding and respecting the wellness/device boundary is therefore a core go-to-market discipline, not a legal afterthought.
The FDA concept of “general wellness”
Low-risk vs disease-related claims
FDA defines general wellness products using a two-part framework: intended use and risk. First, the product must be intended only to support general health or a healthy lifestyle, and not to diagnose, treat, mitigate, cure, or prevent disease. Second, even if framed broadly, the product must present low risk to users’ safety if it performs inaccurately. In practice, this means FDA looks beyond slogans to the reasonable interpretation of claims. Statements tied to maintaining or encouraging healthy habits (such as activity, sleep regularity, or stress awareness) typically remain within wellness scope. By contrast, claims that reference diseases, clinical conditions, or abnormal physiological states, even indirectly, push the product toward device classification.
Importantly, disclaimers do not neutralize intent. A wearable that says “not for medical use” but then discusses disease risk, clinical thresholds, or symptom detection is unlikely to qualify as general wellness. In 2026, FDA messaging continues to emphasize that context and user expectation outweigh fine-print language.
Examples that typically stay “wellness”
Products that track or display data without medical interpretation are most likely to remain wellness. Common examples include step counting, general activity tracking, sleep duration and consistency, non-diagnostic heart rate trends, and stress indicators framed as self-awareness tools, not clinical measures.
Wellness positioning is strengthened when outputs are descriptive rather than evaluative—showing patterns over time instead of labeling values as “normal” or “abnormal.” Population-level insights, educational content, and user-controlled goal setting further support non-device status. In short, tools that inform without concluding and encourage without directing care are the safest candidates for general wellness treatment under FDA policy.
Claims that push you into medical device territory
Diagnosis / treatment / mitigation language
Certain words and phrases almost automatically move a wearable or app out of general wellness and into medical device territory. FDA consistently treats language related to diagnosis, treatment, mitigation, or management of disease as a strong signal of medical intent, regardless of whether the product actually delivers care.
Obvious triggers include verbs such as “diagnose,” “treat,” “manage,” “mitigate,” or “cure.” Less obvious, but equally risky, are phrases like “clinical-grade,” “medical accuracy,” “improves outcomes,” or “supports clinical decisions.” Even when paired with lifestyle framing, these terms imply a role in healthcare delivery rather than self-directed wellness.
The 2026 policy emphasis is that intent is inferred from user expectation. If a reasonable user could interpret the product as helping them make medical decisions or manage a condition, FDA is likely to view it as a device. This is especially true when claims are repeated across marketing channels, onboarding flows, and investor materials, creating a cumulative impression of medical purpose.
“Detects”, “predicts”, “prevents” and biomarker claims
Detection and prediction language presents a particularly common failure mode for wearables. Claims that a product “detects” a condition, “predicts” risk, or “prevents” disease, even probabilistically, imply clinical assessment. FDA does not require certainty; implied medical inference is sufficient. Biomarker-related claims raise similar concerns. Referencing thresholds, abnormalities, or risk ranges for metrics such as heart rhythm, oxygen saturation, glucose trends, or inflammatory markers suggests clinical interpretation. Even if the sensor is consumer-grade, framing outputs as indicators of disease or future illness pushes the product toward regulation.
In 2026, FDA scrutiny is heightened when such claims are individualized. Population-level education may remain wellness, but personalized detection or prediction is far more likely to be viewed as medical device functionality.
Wearables + AI: additional risk factors
Personalized risk scoring and clinical recommendations
AI-driven personalization is one of the fastest ways a wellness wearable can drift into medical device territory. The regulatory risk does not stem from the use of AI itself, but from the nature of the output. When algorithms generate individualized risk scores, rank users by likelihood of a condition, or suggest next steps tied to health outcomes, FDA is more likely to view the product as performing a clinical function.
Personalized risk scoring is particularly sensitive because it resembles diagnostic stratification, even when framed probabilistically. A score that implies higher or lower risk invites interpretation and action, especially if paired with language about prevention or early detection. Similarly, AI-generated recommendations, such as advising rest, medical follow-up, or changes to therapy, can be construed as clinical guidance rather than lifestyle coaching.
In 2026, FDA analysis focuses on reasonable reliance. If typical users are likely to act on AI outputs as medical advice, the product’s wellness framing becomes difficult to defend, regardless of disclaimers or internal intent.
Vulnerable populations and mental health
Wearables that address mental health or vulnerable populations face heightened regulatory sensitivity. Stress and mood tracking can remain within wellness scope when framed as self-awareness tools, but claims related to anxiety, depression, burnout, or suicide risk rapidly approach medical device classification.
FDA is particularly cautious when AI models assess emotional states or behavioral risk, as misclassification or false reassurance could lead to harm. Products aimed at children, older adults, or individuals with known health conditions also receive closer scrutiny. In these contexts, even subtle personalization or alerting features may be interpreted as medical intervention rather than general wellness support.
Practical safe-claims library
Safer alternatives by use case
FDA policy does not require wellness companies to avoid health-related language entirely, but it does require careful claim construction. In 2026, safer positioning focuses on support, awareness, and tracking, rather than interpretation, prediction, or action. Below are practical examples of language that generally stays within general wellness boundaries when used consistently and without contradictory context.
Sleep
- Safer: “Tracks sleep duration and consistency over time,” “Helps you understand your sleep patterns,” “Supports healthy sleep habits.”
- Risky: “Detects sleep disorders,” “Identifies insomnia,” “Improves sleep quality in patients.”
Stress
- Safer: “Provides insights into stress trends,” “Helps you recognize periods of higher stress,” “Supports stress awareness and relaxation routines.”
- Risky: “Detects anxiety,” “Predicts burnout,” “Monitors mental health conditions.”
Physical activity
- Safer: “Tracks daily movement,” “Encourages regular physical activity,” “Helps set and monitor fitness goals.”
- Risky: “Prescribes exercise,” “Optimizes rehabilitation,” “Prevents cardiovascular disease.”
Heart rate and related metrics
- Safer: “Displays heart rate trends,” “Helps users understand changes during activity or rest,” “Supports general cardiovascular awareness.”
- Risky: “Detects arrhythmias,” “Identifies abnormal heart rhythms,” “Predicts cardiac events.”
Across all use cases, wellness framing is strengthened when outputs are descriptive, and not evaluative; language avoids thresholds, abnormalities, or clinical significance; users are encouraged to interpret data in context, not to act medically.
Consistency matters. Even a single medical-sounding claim can undermine an otherwise wellness-positioned product if it alters how users reasonably understand the tool’s purpose.
Go-to-market checklist
Labeling review
Before launch, all outward-facing materials should be reviewed as a single regulatory artifact, not as disconnected pieces. FDA evaluates intended use based on cumulative messaging across product pages, app store descriptions, onboarding screens, screenshots, FAQs, press releases, and sales decks. Teams should verify that claims are consistent, non-clinical, and free of implied diagnosis or treatment language. Particular attention should be paid to verbs (“detects,” “prevents,” “optimizes”), comparative claims (“clinical-grade,” “medical accuracy”), and examples that might imply disease relevance. In 2026, inconsistent labeling is one of the most common triggers for regulatory reclassification.
Evidence expectations
General wellness products are not expected to produce clinical trial data, but they are expected to have credible support for accuracy and reliability. User testing, bench testing, and internal validation may be sufficient if claims remain descriptive and low risk. Evidence should match the claim level: the more personalized or interpretive the output, the stronger the support FDA may expect. Overstating evidence, even informally, can push a product into device territory.
Post-market complaints & incident handling
FDA increasingly views post-market behavior as a signal of true product intent. Companies should maintain a documented process for reviewing user complaints, adverse feedback, and misuse patterns. Recurrent reports of medical reliance, false reassurance, or harm—even if unintended—can prompt FDA interest. Having a clear internal escalation and correction pathway helps demonstrate ongoing risk management.
FAQs
Can we mention disease risk at all?
Mentioning disease risk is one of the fastest ways to cross from wellness into medical device territory. In general, FDA expects general wellness products to avoid disease-specific framing altogether. High-level educational content about health conditions may be acceptable if it is clearly separated from product outputs and does not suggest that the wearable assesses, predicts, or manages risk for an individual user. Once risk is personalized or linked to sensor data, regulatory exposure increases sharply.
What about heart rate variability (HRV) or SpO₂?
Metrics such as HRV or oxygen saturation are not prohibited in wellness products, but how they are framed matters more than their presence. Displaying trends or raw values for self-awareness is typically safer. Interpreting those values as abnormal, clinically meaningful, or predictive of disease, be it explicitly or implicitly, pushes the product toward medical device classification.
Does adding AI automatically make it a device?
No. FDA does not regulate AI per se. However, AI often enables personalization, prediction, and recommendation, all of which increase regulatory risk. If AI outputs materially influence health decisions or invite medical reliance, the product is more likely to be treated as a medical device, regardless of disclaimers or internal intent.
References
- U.S. Food and Drug Administration. (2016, updated policy context through 2024). General Wellness: Policy for Low Risk Devices. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/general-wellness-policy-low-risk-devices
- U.S. Food and Drug Administration. (2022). Clinical Decision Support Software: Guidance for Industry and Food and Drug Administration Staff. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-decision-support-software
- U.S. Food and Drug Administration. (2024). Digital Health Policy Navigator. https://www.fda.gov/medical-devices/digital-health-center-excellence/digital-health-policy-navigator
