Clinical Grade or Consumer Toy? The Line That Decides What You Can Build

Last updated: September 2026
By: Kevin Yamazaki, Partner and CEO at Sidebench

The line is drawn by your claims. Two wearable medical devices with the same sensors can sit on opposite sides of regulation depending on their intended use and the words you publish. If you decide the product first and discover the claim second, you’ve already paid the tax. Decide the claim first.

Health systems, founders and product executives are racing to ship wearables that matter. The difference between a wellness gadget and a regulated medical device is not the chipset. It is the intended use you declare and the clinical reliance you invite. That distinction decides your path, your budget and your timeline.

---

Where the line actually sits

The FDA draws the line on intended use, not on hardware. Two wearables with identical sensors can sit on opposite sides of it, because what decides the category is the claim you publish and the reliance your software invites. Your words and your workflows set your regulatory posture from day one.

The FDA finalized its General Wellness: Policy for Low Risk Devices on January 6, 2026. Low-risk general wellness products stay outside device regulation, and the test is what the product is intended to do.

So a heart rate feature built for general fitness sits in wellness, while the same sensor marketed to detect atrial fibrillation triggers premarket requirements. That one decision sets your scope, your validation plan, your fundraising and your sales model.

Language does most of the work. Words like diagnose, treat, detect, predict, prescribe and medical grade all signal regulated intent. Workflows do it too. Route a reading to a clinician inbox for action and you have implied clinical reliance, whatever the copy says.

There is a carve-out worth knowing. A wellness app can tell someone a visit to a clinician might be worthwhile, as long as it names no disease and no diagnostic threshold. What turns that into a device is routing the reading to a clinician to act on.

The fastest way to waste capital here is to ship a feature that quietly reframes your intended use. Roadmaps drift into regulated territory on a single headline, a chart label or a push notification.

---

The WHOOP case, from warning letter to closeout

In July 2025 the FDA told WHOOP its Blood Pressure Insights feature needed clearance, citing phrases like daily blood pressure estimations and medical grade insights. In June 2026 the agency closed the letter. WHOOP had changed how the feature presented its numbers. The sensor was the same on both dates.

The July 14, 2025 warning letter quoted WHOOP’s own copy back at it: daily systolic and diastolic estimations, and a product delivering medical-grade insights. The FDA’s position was that estimating blood pressure is inherently tied to diagnosing hypertension, so the feature was an unapproved device. WHOOP’s position was that the feature was for wellness.

Then the ground moved. The January 2026 general wellness guidance said non-invasive products may estimate physiologic parameters, blood pressure included, for wellness purposes, provided they avoid disease references and diagnostic thresholds. WHOOP adjusted the feature’s dial boundaries so it no longer read as clinical classification, the company told MedTech Dive, and on June 17, 2026 the FDA closed the letter.

Read the whole arc rather than the headline, because the arc is the lesson. The same sensor was an unapproved medical device in July 2025 and an acceptable wellness feature in June 2026. Nothing physical changed. What changed was the words on the dial and the guidance those words were read against.

That cuts both ways, and I would not let a client read the closeout as an all-clear. WHOOP spent 11 months in enforcement, and its own marketing copy was the evidence the FDA quoted. Lock claim language before sprint 1. I still see draft decks that would turn a general wellness build into a device overnight.

---

The two markets: who builds for trials, and who builds for consumers

The wearable market has already split along the line the FDA draws. One set of vendors builds for trials and monitored care, carries device clearances and quality systems, and sells to researchers and health systems. The other builds for consumers and stays inside general wellness. Which market you sit in decides your hardware partner as much as your claims.

On the clinical side, Corsano Health’s CardioWatch 287-2 was cleared by the FDA under 510(k) and carries CE marking under EU MDR, manufactured under ISO 13485. The company reports involvement in more than 100 clinical trials across cardiovascular, oncology and stress research. Ametris, formerly ActiGraph, has built medical-grade activity and sleep sensors for clinical research since 2004, and was acquired by Signant Health in May 2026 to pair sensor data with patient-reported outcomes. Both companies are organized around evidence generation.

On the consumer side, WHOOP and Oura are the reference points, and they are very good at what they are built for: continuous, comfortable measurement that people actually keep wearing. Adherence is the hard problem in any longitudinal program, and consumer devices solve it better than clinical hardware does.

What they are not built to do is carry a clinical endpoint. Clients running trials and longevity programs tell us the same thing: the output is too noisy for that job. The reasons are structural. Sampling rates change with firmware. Algorithms update without notice or version history. Gaps appear whenever a device is charging. None of that matters when the output is a nudge about last night’s sleep. All of it matters when a clinician is about to act on the number.

	Built for	Regulatory posture	What the data can carry
Clinical-grade (Corsano, Ametris)	Trials, research, monitored care	FDA clearance, CE marking, quality system	Endpoints, monitored care within labeling
Consumer (WHOOP, Oura)	Individuals, engagement, adherence	General wellness, outside device rules	Trends, habits, screening, engagement

The expensive mistake is picking the device for adherence and then writing a claim the device cannot support. Decide the claim first, then choose hardware that can stand behind it.

---

General wellness vs regulated device: what changes

If you frame a wearable as general wellness, you avoid medical device requirements but must avoid clinical reliance. If you frame it as diagnosing, treating or guiding care for a specific condition, you face medical device controls, validation and premarket review. The table below shows how the categories differ.

Dimension	General Wellness Wearable	Regulated Medical Device Wearable
Intended use	Supports general health or wellness goals without diagnosing or treating a specific condition	Intended for diagnosis, cure, mitigation, treatment or prevention of disease or for guiding clinical care
Example claims	Track daily heart rate trends. Support mindfulness. Promote better sleep habits.	Detect AFib. Estimate blood pressure for hypertension management. Provide dosing guidance.
Validation expectations	Values that mimic clinical ones must be validated, though the guidance does not define the standard; no claims of clinical accuracy, clinical equivalence or medical grade	Clinical and analytical validation appropriate to claim; verification, validation and human factors expected
Premarket pathway	Generally not regulated as a device under the General Wellness policy finalized Jan 6, 2026	Subject to device classification and, where applicable, 510(k), De Novo or PMA premarket requirements
Marketing language	Avoids terms that imply diagnosis or treatment, and cannot claim clinical accuracy, clinical equivalence or medical grade	May use clinical terms aligned with cleared or approved intended use and labeling
Software features	No features that encourage clinical reliance, alarms for conditions, or clinician routing for action	Features can support clinical decisions consistent with cleared labeling; often include clinician workflows
Risk profile	Low risk by design and by claim	Managed under quality system and risk controls proportionate to the device classification
Data use in care	Not to be used to make clinical decisions	May be used in clinical care consistent with labeling and intended use

The 2026 policy keeps the direction set in earlier guidance and tightens one part of it. Where a wellness product shows a value that mimics a clinical one, that value has to be validated, and the product still cannot claim clinical accuracy or medical grade.

The smartest wearable teams write the intended use statement first, then design the product around it.

---

What changed in the 2026 General Wellness policy

The January 6, 2026 guidance replaced the 2019 version and widened the lane. Non-invasive products may now show physiologic values, blood pressure included, for wellness purposes. The catch sits in the same guidance: a value that mimics a clinical one has to be validated, and you still cannot claim clinical accuracy or medical grade.

Three things are genuinely new:

• Wellness products may estimate and display physiologic parameters, provided they name no disease and no diagnostic threshold.
• Values that mimic clinical ones must be validated. The guidance does not say what that validation has to consist of, which leaves the standard to you and your regulatory adviser.
• Claiming to be a substitute for an FDA-cleared device is out, and that expressly covers claims of clinical equivalence, clinical accuracy or medical grade.

Covington & Burling’s summary names the tension neatly: you are expected to validate, and you are not allowed to say so in the words a buyer would find persuasive.

If your roadmap includes alarms to clinicians, condition detection, decision prompts or the words medical grade, you are building a device. Budget the quality system, the validation work, the labeling and the premarket strategy.

---

The data engineering nobody budgets for when consumer devices feed clinical decisions

When consumer wearables feed care decisions, the hardest work sits below the app. It is cleaning the data you plan to act on, reconciling it across sources, and being able to prove any of it later. That work is engineering, procurement and compliance at once, and it should be budgeted before screens are.

Sidebench has delivered 60+ healthcare implementations over 14 years, and in wearable-heavy builds the real lift almost always sits in the data layer, well below the interface. We saw it again working with one of the largest outpatient medical imaging and radiology centers in the US on its move into longevity, where the hard part was never the screens. It was making imaging history, lab results and continuously streaming device data agree with each other well enough for a clinician to act on any of it. Device ingestion, provenance tagging, time alignment, and rules that stop consumer values from masquerading as clinical data: that is where the engineering budget goes.

Here’s what to scope if wearable data may influence care or research:

• Know where every number came from. Source, device, firmware version, and whether it was measured or worked out by an algorithm.
• Get the clocks to agree. Devices batch and backfill, so timestamps rarely line up on their own.
• Decide what happens when data goes missing. Charging, off-wrist and failed syncs are normal, not edge cases.
• Keep consumer readings out of the clinical record unless your labeling lets them in.
• Plan for the device changing under you. Firmware and app updates ship without asking, and they move your numbers.

If your partner is talking screens before source-of-truth, they’re aiming past the real risk.

For more on the practicalities of wearable integration in care, we’ve published a dedicated piece: https://sidebench.com/wearables-integration-healthcare-2026/

---

Consumer signal vs clinical-grade signal: what you can do with each

Consumer wearable signals are good for engagement, habit change and early research screening. Clinical-grade signals can guide diagnosis and treatment inside approved labeling. What separates them is accuracy, and behind accuracy sits provenance, and behind that the intended use you declared.

Dimension	Consumer Signal	Clinical-Grade Signal
Accuracy	Fit for trends, behavior feedback and population insight; varies by device and context	Validated for a stated intended use and conditions of use
Provenance	Metadata is often thin, and firmware or algorithm updates can move outputs without notice	Full version control under a quality system, with changes documented and validated
Data rights	Consumer terms and consents; revocation and portability differ by vendor	Clinical consents and HIPAA-aligned handling, with labeling for clinician use

Teams often try to split the difference, but the safest builds make a clear choice and communicate it relentlessly to users, clinicians and sales.

Many programs gain more by nailing a consumer signal with proof of engagement before climbing the validation hill.

---

If you already have a spec, how to judge a build partner

Pick a partner that can write an intended use statement, defend it, and either hold the line on general wellness or take you through SaMD controls. Ask for proof of clearance experience, data provenance design and device change-management. If they haven’t shipped a regulated product, you’ll fund their learning curve.

Here’s the checklist we run:

• Intended use discipline
• Can they draft and pressure-test your intended use in plain language that marketing, legal and clinical stakeholders will sign?
• Do they run claim reviews across product, UI, marketing and sales collateral?
• Regulatory pathway
• What have they taken through an FDA pathway, in what role, and what did it change about how they work?
• Can they align your roadmap with whether you need 510(k), De Novo, or whether you can live in general wellness?
• Data engineering
• How will they tag provenance, handle firmware churn and normalize device data?
• What is their plan for time sync, gap handling and transform governance?
• Quality and change
• Do they run design controls, verification and validation where needed?
• Can they manage device SDK and firmware updates with a holdout testing process?
• Clinical boundaries
• Can they keep consumer values out of clinical workflows if you’re staying wellness-only?
• Can they design clinician experiences that match the labeling if you are regulated?
• Evidence and claims
• Can they define what evidence you need to support what your values reflect under the 2026 policy without implying clinical accuracy?
• Do they know how to write UI copy that avoids clinical reliance language while preserving user value?
• Commercial considerations
• Can they help you pick channels that match your claims, such as wellness programs vs RPM billing models?
• Do they know payer and provider sensitivities around consumer data in clinical charts?

Sidebench has built on both sides of this line. On the regulated side we were the engineering and product partner on AppliedVR’s RelieVRx, which the FDA authorized through the De Novo pathway. On the consumer side we built a tactical-fitness app, anonymized here, running on WHOOP and Apple Health integrations.

Regulatory work will slow your roadmap more than a feature rewrite. Plan for it. And ask any partner what they have taken through an FDA pathway, in what role, and what it changed about the way they build.

---

FAQ: wearable medical devices and the claims that define them

Does the FDA regulate all wearable medical devices?

No. Under the General Wellness: Policy for Low Risk Devices finalized on January 6, 2026, the FDA preserves the policy of not regulating low-risk general wellness products as medical devices. Devices intended for diagnosis or treatment of specific conditions remain subject to medical device requirements.

What is the single biggest factor in whether my wearable is regulated?

Intended use. The FDA looks at what you say the product does and how users are expected to rely on it. Software features and marketing claims decide the category.

Can I display health-like metrics without being a device?

Yes, if the technology is low risk and non-invasive and you avoid encouraging clinical reliance. One condition comes with it: where the value mimics a clinical one, it has to be validated. The guidance does not define that standard, and you still cannot claim clinical accuracy or medical grade.

What kinds of claims push a product into device territory?

Claims that suggest diagnosis, cure, mitigation, treatment or prevention of a specific disease. Phrases like detect AFib, estimate blood pressure for hypertension management or medical grade performance are red flags when not cleared or approved.

What happened in the WHOOP case?

The FDA issued a warning letter on July 14, 2025 over the Blood Pressure Insights feature, quoting WHOOP’s descriptions of daily blood pressure estimations and medical-grade insights. After the January 2026 general wellness guidance, WHOOP adjusted how the feature displayed its readings and the FDA closed the letter on June 17, 2026. The hardware never changed.

If I avoid clinical words, can I still be a device?

Yes. Workflows can imply clinical reliance even without clinical words. For example, routing an alert to a clinician for action can suggest diagnosis or treatment intent. Both claims and features matter.

Do I need clinical validation for a general wellness wearable?

You are outside medical-device validation, but not outside validation. Under the 2026 policy, a value that mimics a clinical one has to be validated. The FDA does not say what that means in practice, so agree the standard with your regulatory adviser and document it before launch.

How should I think about remote patient monitoring devices?

If your product is intended for RPM billing or clinical decisions, it likely sits in the regulated category and must align with device requirements and labeling. If your product supports general wellness programs without clinical reliance, it may fit under the General Wellness policy.

What does a regulated software build involve?

A regulated build includes design controls, risk management, verification and validation, labeling, and a premarket strategy aligned with your intended use. Sidebench was the engineering and product partner on AppliedVR’s RelieVRx, authorized by the FDA through the De Novo pathway, and the proof work was as big a job as the features.

What is one practical way to avoid accidental clinical claims?

Write your intended use and explicit non-clinical statements early. Review every UI string, push notification and sales slide against them. The regulator reads your claims, not your circuit board. Train your growth team and PR to stay within those lines.

---

What separates the teams that ship from the teams that rebuild

Teams that ship wearable products write the intended use statement before they choose hardware, and hold every screen, alert and marketing line to it. Teams that rebuild pick the device that gets the best adherence, write the claim afterwards to match what the product turned out to do, and find out where the line was from a regulator.

The 2026 guidance made that ordering matter more, not less. Widening the general wellness lane gives consumer products more room, and it sharpens the edge of that lane for anyone whose language drifts toward clinical reliance. More room inside the line means a harder edge at it.

The sequencing advice is simple enough to act on today. Write the intended use first. Choose hardware that can stand behind it. Then design the screens, and hold the marketing copy to the same standard as the labeling, because the FDA reads both.

If you are deciding where your claim sits, we can review your intended use statement against your device choice and your roadmap, and tell you which of the three you need to change. See how Sidebench approaches product strategy and discovery, or send the spec and we will start there.

---

Cited sources

• U.S. Food and Drug Administration, General Wellness: Policy for Low Risk Devices, final guidance issued January 6, 2026, superseding the September 2019 version
• U.S. Food and Drug Administration, Warning Letter to WHOOP, Inc., July 14, 2025
• U.S. Food and Drug Administration, Close-out letter to WHOOP, Inc., June 17, 2026
• MedTech Dive, FDA drops Whoop warning letter over blood pressure feature, for WHOOP’s account of the changes it made
• Covington & Burling, FDA Issues Revised Guidance on General Wellness Products, January 2026
• Corsano Health, CardioWatch 287-2, FDA 510(k) cleared, CE marked under EU MDR, manufactured under ISO 13485
• Ametris (formerly ActiGraph), medical-grade wearables for clinical research, acquired by Signant Health, May 2026
• Sidebench. Wearables Integration in Healthcare 2026

---

---

About the author