Wearables can estimate blood pressure without FDA review. They cannot say what it means

The FDA revised its General Wellness: Policy for Low Risk Devices guidance on 6 January [s1], and announced it from the stage at CES [s3]. The revision resolves an argument the agency has been having with the wearables industry for years, and it resolves it in the industry's favour.

Non-invasive wearables may now estimate physiologic parameters — including heart rate, blood pressure and glucose levels — and remain outside device regulation, provided the measurements are appropriately validated and the product avoids disease references [s1]. The covered categories also include activity, recovery, sleep, pulse and fitness-related biomarkers [s1].

The regulated boundary has moved from what a device measures to what it says.

The line, in the Commissioner's own framing

Marty Makary described the test as a claims test. "We want to let companies know, with very clear guidance, that if their device or software is simply providing information, they can do that without FDA regulation," he said [s2]. The stipulation, he added, is whether a company makes claims of something being medical grade — "like a clinically appropriate, clinical grade blood pressure measurement" [s2].

He also signalled the agency's enforcement posture: "We're not going to go in there and say, 'There's one result that is inaccurate, [and] therefore we've got to shut this down'" [s2].

The exempted categories named alongside the guidance are heart rate monitors, general health and wellness wearables, software providing informational tracking of health metrics, and AI-enabled health and wellness devices [s2].

What stays regulated: products implying medical use, guiding clinical decisions, or substituting for an FDA-authorised device — even if they merely estimate a physiologic parameter [s1]. Invasive features, lasers and radiation fall outside the general wellness framework regardless of the claims made [s1].

Why this argument existed

The dispute that produced this guidance was about a blood pressure feature.

The FDA had previously issued a warning letter to WHOOP over its Blood Pressure Insights feature, which surfaced systolic and diastolic estimates; the agency determined that this crossed into medical device territory [s3]. The company's position was that the feature was wellness, not diagnosis.

The revised policy resolves that class of dispute structurally. Displaying an estimated blood pressure number is now presentation of information. Telling a user what the number means for their health is a medical claim [s3].

The problem with a claims-based line

The framing is administratively clean and physiologically strange, because the person receiving the number does not experience it as neutral information.

Someone who sees "systolic 158" on a wrist device does not treat it as an uninterpreted output. They interpret it, using whatever they know about blood pressure, and then act — by worrying, by seeking care, or by deciding the reading is fine and not seeking care. The clinical consequence attaches to the number, not to the sentence next to it.

Keeping the accompanying text non-diagnostic is a reasonable design constraint, but it does not change the underlying dynamic: the accuracy of the estimate determines whether a user is alarmed appropriately or falsely reassured, and accuracy is exactly what premarket review used to examine.

What "appropriately validated" is doing

The guidance conditions the exemption on measurements being appropriately validated [s1]. That phrase carries the entire consumer-protection weight of the policy, and it is not defined by a premarket submission any more.

The questions it leaves open are the ones that determine whether a cuffless blood pressure estimate is useful or misleading. Validated against what reference — an oscillometric cuff, or an intra-arterial line? In whom: what age range, what skin tones, what body habitus, what blood pressure range, and across what range of skin tones and perfusion states? Over what period — a single seated measurement, or through movement and across days? And with what error bounds on an individual reading, as distinct from a group mean?

Under the previous arrangement, a company that wanted to make a blood pressure claim had to answer those questions to a regulator. Under this one, it can decline to make the claim and publish the number anyway.

What this is downstream of

The changes were announced alongside an emphasis on AI adoption, and the general wellness revision explicitly permits wellness products to incorporate AI-derived insights while remaining outside device regulation, provided they avoid diagnostic or treatment claims [s1]. The agency positioned the package as aligning with a December 2025 White House executive order promoting AI adoption [s2].

That context matters for reading the guidance correctly. This is deregulation pursued as innovation policy, and it is presented that way by the agency rather than concealed.

What to watch

The practical question for the next year is whether any manufacturer voluntarily publishes the validation data the guidance now assumes exists. The exemption is conditioned on appropriate validation [s1]; nothing requires that validation to be visible to the person buying the device.