Publication date is not effective date

A Federal Register document has a date at the top, and it is tempting to read that date as the answer to “when does this bind me.” It answers a narrower question: when FDA published it. FDA Radar’s own Federal Register record keeps a second date for a reason — publication and effect are not the same fact, and for some device rules, neither one is the date that actually matters to a specific product.

The floor Congress set

5 U.S.C. 553(d), the Administrative Procedure Act’s rulemaking section, sets a minimum: the required publication of a substantive rule has to happen not less than 30 days before its effective date, with narrow exceptions for rules that grant an exemption, interpretive rules and policy statements, and cases where the agency states good cause to move faster. That is a floor, not a target. Nothing in the statute caps how much later than 30 days an effective date can land, and FDA’s own device rules regularly land much later than that.

A rule that took seven years to finish binding anyone

FDA published the Unique Device Identification System final rule on September 24, 2013. The rule itself became effective December 23, 2013 — a handful of sections on October 24, 2013 — comfortably past the 30-day floor and, on its own, a fact that sounds like the whole story. It is not. The rule did not require any device to actually carry a UDI on that date. Compliance dates ran on a separate schedule keyed to device class: Class III devices by September 24, 2014; Class I and II implantable, life-supporting, or life-sustaining devices a year after that; general Class II devices by September 24, 2016; and general Class I and unclassified devices not until September 24, 2020 — seven years after the rule that created the requirement was published, and roughly six and a half years after that rule took legal effect.

Effective tells you the rule is on the books. It does not tell you which device, if any, has to do anything about it yet.

Two fields, not one answer

FDA Radar’s own Federal Register ingestion keeps publication_date and effective_on as two separate fields on every device-relevant document, exactly because collapsing them into one would erase a real distinction FDA itself draws. Neither field, by itself, carries a class-by-class compliance schedule buried inside a rule’s own text — that schedule is not a fact the Federal Register’s API structures at all. It lives in the document’s body, in whatever table or paragraph FDA chose to write it into, and nothing short of reading that document surfaces it.

The practice

Treat a rule’s effective date as the date it exists, not the date it applies to a specific device class. For any device final rule that phases in by class, risk category, or device type — UDI is one instance of a pattern FDA reuses, not a one-off — the compliance-date table sits inside the document itself, not in either date field a feed can hand you. Read it before assuming a rule that’s been “in effect” for years has already reached your device’s class. This is regulatory intelligence about how FDA structures a rule’s own timeline, never a determination of when a specific device must comply — that reading belongs to whoever owns the file, against the rule’s actual text.