A Class I device, a Class I recall

Pull a device recall record from openFDA and two fields sit a few lines apart, both capable of holding the string “Class I”: one describing the recall, one describing the device it recalls. They are not the same fact wearing two hats. They are two separate scales that happen to share a vocabulary, and nothing about either record warns a reader that the collision exists.

What a device’s class actually measures

A device’s regulatory class — Class I, II, or III — is a permanent attribute of the device type, assigned once by FDA’s classification regulation and unchanged until a reclassification order says otherwise. It measures the level of control needed to reasonably assure safety and effectiveness across the category as a whole: Class I devices are subject to general controls only, Class II devices carry special controls in addition, and Class III devices require premarket approval. A tongue depressor and a bandage sit in Class I not because nothing can ever go wrong with either, but because general controls are judged sufficient for the category. The number describes the device type. It says nothing about any single unit that ships.

What a recall’s class actually measures

A recall’s class is a different kind of fact entirely: a per-event severity judgment, made after the fact, about one specific defect. Under 21 CFR 7.41, FDA’s health hazard evaluation classifies a recall Class I when there is a reasonable probability that using or being exposed to the violative product will cause serious adverse health consequences or death, Class II when the consequences are temporary or reversible, or serious harm is merely remote, and Class III when adverse health consequences are not likely at all. That evaluation is about what this particular problem could do to someone, this time — not about which regulatory tier the device type was assigned to years earlier.

A device’s class describes what kind of thing it is. A recall’s class describes how badly this one went wrong. FDA reuses the same three labels for both, and the reuse is a coincidence of vocabulary, not a shared meaning.

Why the two scales can point in opposite directions

Nothing about a device’s baseline class caps or floors the severity a defect against it can reach. A contamination event in a Class I product — gauze, surgical gloves, an antiseptic wipe — can plausibly carry a reasonable probability of serious infection, which is a Class I recall sitting on top of a Class I device: the highest severity label this system has, attached to the lowest regulatory tier this system has. The reverse holds just as easily — a labelling error on a Class III implant, corrected without any plausible health consequence, closes out as a Class III recall. The device’s own class is not a prior on the recall’s class. Reading one as a hint about the other is reading a coincidence as a correlation.

The practice

FDA Radar’s own review queue is built around exactly this distinction, and the code’s own comment states the reasoning bluntly: routine recalls route to a searchable corpus rather than a reviewer’s inbox, with one deliberate exception — “Class I is the exception and the whole point: FDA reserves it for a reasonable probability of serious injury or death. That is the item a Regulatory Affairs Director must see the day it posts.” That gate checks the recall’s own classification field, never the device’s. A Class III device recalled at Class I severity reaches a reviewer immediately; a Class I device recalled at Class III severity does not, and correctly so — the device’s regulatory tier was never the question the gate is asking. Anyone reading recall data by hand should check the same field, for the same reason: what matters this week is what this defect could do, not what shelf the device type sits on.

Which specific recall is worth escalating, and what a given severity finding means for a supplier or a predicate, is a judgment for whoever owns that risk file, on the current record — this is a description of what two identically-labelled fields measure, not a determination about any device or recall in particular. Per what we monitor and how often, both fields run on every recall FDA Radar ingests, kept apart in the data exactly as they are kept apart in the rule.