What the instructions for use actually has to say, and to whom

Two instructions-for-use documents can describe the same device, carry the same warnings, and still be built to satisfy entirely different tests. Section 502(f)(1) of the FD&C Act, applied through 21 CFR 801.5, asks whether the directions are adequate — whether a layperson could use the device safely and for its intended purpose from what's written — a question FDA answers by judgment, not by a checklist. EU MDR Annex I Section 23.4 asks something narrower and easier to audit: does the document contain each of roughly twenty specified items the section names. A file built to satisfy one test can look thin measured against the other, and knowing which test is actually in front of you changes what “adequate” requires.

A performance standard, not an ingredient list

21 CFR 801.5 defines adequate directions for use as directions under which a layperson can use a device safely and for the purposes for which it's intended, and a device that lacks them is misbranded under section 502(f)(1) of the FD&C Act. The regulation doesn't enumerate what has to appear in those directions — the same requirement applies whether the device is an adhesive bandage or a home glucose meter, satisfied differently depending on what the device actually asks its user to do correctly. A document can't earn a pass by matching a list of section headings; it has to work for the reader the standard has in mind, which is a usability question before it's a content question.

The prescription-device swap, not an exemption from writing anything

21 CFR 801.109 exempts a device from the layperson standard when it's unsafe to use except under the supervision of a licensed practitioner — but only under specific conditions: it has to be sold to, or on the order of, such a practitioner, and dispensed accordingly. In exchange, the label or dispensing package has to carry professional labeling — indications, effects, routes, methods, frequency and duration of use, along with relevant hazards, contraindications, side effects, and precautions — written for a practitioner audience instead of a layperson one. 801.109 changes who the content has to work for. It doesn't remove the content requirement.

Annex I turns the same question into a named list

MDR Annex I Section 23.4 takes a structurally different approach: it requires the instructions for use to contain a defined set of particulars, including several carried over from the general labeling requirements in Section 23.2, plus items specific to the IFU itself — the device's intended purpose with a clear specification of indications, contraindications, and target patient and user groups, and, where applicable, the clinical benefits to be expected. Checking a draft IFU against that list is a real, useful gate, and one 801.5's open-ended standard structurally can't offer on its own, because it never names a fixed set of items to check for presence in the first place.

The gap neither checklist closes

A completeness pass against Annex I's list, or an internal template built around 801.5's known content areas, only proves the sections exist — not that a layperson, or under 801.109 a practitioner, can actually follow them to use the device correctly. That's the same distinction a trustworthy checklist has to hold onto between an item that's present and a claim that's actually true, and the same one a Refuse-to-Accept review draws between a submission that's complete and one that's correct. An IFU can pass every item on a completeness list and still fail the standard the list was only ever a proxy for — that gap closes through usability testing with representative users, not through a second read of the table of contents.

Where the wording traces back further

A warning or precaution in an IFU is frequently not just labeling copy — under ISO 14971 it can be the risk control measure itself, the thing that reduces a hazard's estimated risk from unacceptable to acceptable in the risk management file's own chain. A warning added to the instructions with no corresponding entry in that file's control list is asserting a risk control the risk file can't show was ever verified — which makes the IFU and the risk file two views of some of the same underlying evidence, not two documents that can be drafted independently and reconciled later.

An instructions-for-use template built around both tests — the 801.5/801.109 branch and the Annex I 23.4 checklist, cross-referenced back to the risk file's control measures — is previewed in the launch catalog. If your team has a usability finding this structure would have missed, the shelf takes that correction directly.