The Regulatory Path for a Wearable Medical Device

Two wearables can contain identical hardware — the same optical sensor, the same accelerometer, the same processor — and sit on opposite sides of a regulatory line. One ships next quarter from a contract manufacturer with no FDA involvement at all. The other needs a submission, clinical data, a quality system, and eighteen months. The difference is almost never the electronics. It is what the company says the device does — the most consequential decision in the program, and one that belongs in month one, not after the industrial design is locked.

The line the FDA actually draws

A product becomes a medical device when it is intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease, or to affect the structure or function of the body. "Intended use" is the operative phrase, and the agency reads it broadly: your marketing copy, your website, your app screens, your investor deck, and what your sales team says are all evidence of intent. The general test is walked through in is my product a medical device.

Alongside that, FDA has a general wellness policy that describes a category of low-risk products the agency does not intend to actively regulate. To sit inside it, a product generally has to satisfy two conditions at once.

• The claim is about general wellness — healthy lifestyle, fitness, sleep quality, relaxation, weight management — and not about a specific disease. Disease-related claims survive only in the narrow "may help reduce the risk of" or "help living well with" framing where the association is well accepted.
• The product is low risk. No invasiveness, no implantation, no laser or radiation exposure raising safety questions.

Both conditions must hold; a low-risk product with a diagnostic claim is still a device.

What flips a wearable into regulated territory

In practice, a small number of claim patterns account for most of the crossings. This table is the conversation to have with your regulatory advisor before you write any marketing copy.

Three triggers founders miss regularly. Alarms: once a device is expected to notify someone that a person is in trouble, its failure can cause harm. Absolute values: reporting a trend is very different from reporting a clinical number a doctor could act on. The clinician user: a product sold to hospitals for monitoring is a device even if a nearly identical consumer version is not.

Also note that the FDA's oversight follows the software, not just the hardware. If the algorithm running on the phone or in the cloud produces the clinical output, that software is the regulated part — see software as a medical device.

Which pathway a regulated wearable takes

Assuming you are in, the next question is classification, which is driven by risk. Most wearable monitors land in Class II, requiring a 510(k) with special controls; a few genuinely novel ones have no predicate. The class definitions are in FDA medical device classes.

510(k) when a predicate exists

If a legally marketed device already does substantially what yours does, you demonstrate substantial equivalence to it. This is the faster, cheaper route — several months of FDA review after preparation, with modest small-business user fees and consulting and testing costs that are not modest. The mechanics are in the FDA 510(k) submission process.

De Novo when nothing comparable exists

Novel wearables frequently have no predicate. De Novo is the route for a low-to-moderate-risk device without one, and a successful submission creates a new classification later entrants can use — a genuine competitive asset. It takes longer and demands more clinical evidence than a 510(k); the path is in the FDA De Novo pathway.

Use the Q-Submission program

Before committing, request a pre-submission meeting. You send a package describing the device, your proposed classification, your predicate, and your planned testing, and the agency gives written feedback. It costs nothing but preparation time, and it is the highest-return activity available to an early-stage device company. Ask the questions you are most afraid of.

Validating the algorithm, not just the sensor

A wearable's clinical claim usually rests on an algorithm converting a noisy signal into a number. Proving the sensor works is engineering verification. Proving the algorithm produces a clinically trustworthy result is validation, and it needs human data.

What reviewers look for:

• An appropriate reference standard. Compare against the accepted clinical measurement — polysomnography for sleep staging, a clinical ECG for rhythm, arterial blood gas for oxygen saturation. Another consumer wearable is not evidence.
• A population that matches the intended use. This is where wearable studies fail most often. Optical sensing performance varies with skin pigmentation, and studies that enroll only light-skinned, young, healthy subjects will draw questions. Enroll across skin tones, body sizes, ages, and the disease states you claim to work in.
• Real-world conditions. Motion artifact, poor fit, sweat, low perfusion, cold hands. A device validated only at rest in a lab will be challenged.
• Pre-specified endpoints and a locked algorithm. Freeze the algorithm before the pivotal data collection. Tuning it on the same data you use to prove it works invalidates the result, and reviewers ask about this specifically.
• Defined failure behavior. What the device does when the signal is not good enough. A device that outputs a confident wrong number is far more dangerous than one that says "insufficient signal."

Whether this requires a formal study under an investigational device exemption depends on risk; the thresholds are covered in clinical trials for medical devices. Many wearable validation studies are non-significant-risk and can proceed with institutional review board approval alone, which is dramatically cheaper.

Continuous monitoring raises the bar

Spot-check measurement and continuous monitoring are different products to a reviewer. Continuous claims bring alarm behavior, data completeness, and the question of what happens when the phone is out of range or the app is killed by the operating system. If your wearable is meant to catch an intermittent event, characterize how much of the time it is actually watching and what a missed window means clinically.

Connectivity and cybersecurity are inseparable from this. A connected wearable that transmits clinical data needs threat modeling, secure update capability, and documentation the agency now expects as part of the submission — see FDA cybersecurity requirements for connected medical devices.

Practical sequencing

1. Write the intended use statement in one sentence, in month one. Everything else follows from it.
2. Get a regulatory opinion on which side of the line that sentence puts you, and what a small wording change would cost or save.
3. If you are regulated, search for predicates before finalizing the claim — an achievable predicate can be worth reshaping the product around.
4. Run a pre-submission before spending on pivotal data.
5. Design the hardware for the regulated case even if you launch a wellness version first.

That last point matters commercially. Launching a wellness wearable and adding a cleared claim later is a sound strategy only if the first-generation hardware was engineered so the second generation is a submission rather than a redesign — including the physical side covered in wearable product design.

Projects House develops wearables on both sides of the line, with the regulatory question settled before the architecture is frozen. Describe what your device measures and what you want to claim through the contact form.