A consent form owes eight elements, not a signature

A device clinical study's paperwork tends to get planned backward: pick the IRB, draft a protocol, and treat the informed consent form as the last document assembled before enrollment opens — a formality built to close out the file rather than a regulatory requirement with its own list of required content. 21 CFR 50.25 doesn't treat it that way. It names eight basic elements a subject has to be told, in language they can understand, before they can validly consent to anything — and none of the eight is the signature line most teams treat as the document's real purpose. Whether a study runs on the significant-risk or nonsignificant-risk track this blog has already traced through 812.2(b)'s fork, both land on the same consent requirement afterward.

The eight basic elements are the floor, not a suggestion

21 CFR 50.25(a) lists eight elements a device trial's informed consent has to include, and the list reads like a checklist because it's meant to be used as one: a statement that the study involves research, an explanation of its purposes and the subject's expected duration of participation, a description of the procedures involved and which of them are experimental; the reasonably foreseeable risks or discomforts; the benefits reasonably expected, to the subject or to others; a disclosure of any alternative procedures or treatments that might be advantageous; a statement on how confidentiality will be maintained and that FDA may inspect the records; and, for research involving more than minimal risk, whether compensation or medical treatment is available if injury occurs and what that treatment consists of. Two more close the list: whom the subject can contact about the research, their rights, or an injury, and a statement that participation is voluntary and can be refused or withdrawn without penalty. Eight discrete disclosures, each doing a different job — not eight ways of restating the same reassurance.

Element six only applies once risk crosses a line

Element six — compensation and treatment for injury — only attaches once a study is “more than minimal risk,” and that phrase does real work: minimal risk under 21 CFR 50.3(k) means a probability and magnitude of harm no greater than what's ordinarily encountered in daily life or routine exams. That's a judgment call sitting next to another this blog has already covered — the significant-risk determination a sponsor makes for itself under 812.2(b) to decide which IDE track a device study runs on. The two determinations aren't the same test and don't automatically move together: a nonsignificant-risk device can still expose a subject to more than minimal risk from the study procedures themselves, and a consent form built only around the IDE track's own risk label can miss the separate question 50.25(a)(6) is actually asking.

The additional elements aren't decoration either

50.25(b) adds six further elements “when appropriate” — a phrase that means the investigator has to make a call, not that the elements are optional once appropriate. They cover unforeseeable risks to an embryo or fetus if the subject is or could become pregnant; the circumstances under which an investigator can end a subject's participation without the subject's consent; any added costs from participating; the consequences of withdrawing and the procedure for doing so in an orderly way; a commitment to disclose significant new findings that might affect willingness to continue; and, where relevant, the approximate number of subjects in the study. A study with a legitimate reason to trigger any of these can't leave it out because 50.25(a)'s eight are the ones everyone remembers.

A signature doesn't cure a form that skipped an element

It's tempting to treat the signed form as the artifact that matters and the elements as drafting guidance behind it, but FDA's own guidance on informed consent describes consent as a process documented by the form — not a form that becomes valid because it was signed. A form missing a required element is deficient regardless of how many subjects signed it, and the deficiency doesn't surface until an inspector or auditor reads the form against 50.25 rather than checking for a signature. 21 CFR 50.27 even allows a short-form written consent read aloud with a separate written summary, for exactly the case where the full-length document isn't the right vehicle — but the short form still has to be backed by every element 50.25 requires, just carried in the summary instead of the form itself.

Where this meets the file

A consent-form template that maps each paragraph to its 50.25(a) or (b) element — rather than one written once for a prior study and reused as a boilerplate feel-good document — is previewed in the launch catalog, alongside the risk-determination record this blog has already traced through the IDE fork. If your program documents this differently, the shelf takes that correction directly.