A Study Risk Determination locks in FDA's own call

A sponsor who reads an IRB's nonsignificant-risk finding as the end of the question has stopped one step short of where the regulation actually lands. This blog has already covered the fork 21 CFR Part 812 builds between significant-risk and nonsignificant-risk device studies, and the sponsor's own initial risk assessment, reviewed by the IRB, is what usually decides which track a study runs on. What that post didn't get into is that the IRB's agreement isn't the last word FDA recognizes. FDA's own authority over that determination survives the IRB's review, and a formal Q-Submission built specifically to settle it — the Study Risk Determination — exists precisely because relying on an IRB's read of a definition FDA can still overrule is a real, and avoidable, exposure.

The IRB's determination is a default path, not a settled one

The ordinary sequence starts with the sponsor: assess the device against the significant-risk criteria at 21 CFR 812.3(m), present that assessment to the IRB along with the supporting rationale, and let the IRB agree or disagree. Where the IRB agrees the study is nonsignificant risk, its own approval is enough to start enrollment — the IRB functions as FDA's surrogate for that review, with no IDE application required. Where the IRB determines significant risk instead, the sponsor needs an FDA-approved IDE before the study can begin. FDA's own information sheet on significant-risk and nonsignificant-risk device studies is explicit about what sits underneath that whole sequence: FDA is not bound by the IRB's determination, and can disagree with an NSR finding at any point — including after a study has already started enrolling on the strength of the IRB's own approval.

What the Q-Submission actually asks for

FDA's Q-Submission Program guidance, finalized in June 2023, names four distinct request types: Pre-Submissions, Submission Issue Requests, Informational Meetings, and Study Risk Determinations. An SRD is built around one narrow question — whether a specific planned study is significant risk, nonsignificant risk, or exempt from IDE regulations altogether — supported by a device description, the study protocol, the intended use, and the population the study will enroll. The reply that comes back is FDA's own written view of the study's risk category, not an informal comment folded into a broader strategy conversation.

Why asking first, not later, is the entire point

A sponsor who proceeds on an NSR track built entirely on an IRB's own agreement is carrying a specific exposure: if FDA later reaches a different conclusion — during a related submission's review, or an inspection, or simply a closer look prompted by something else — the study may turn out to have proceeded without the IDE it actually needed. Data gathered under that gap can become unusable for the marketing submission it was meant to support, and the sponsor is left retrofitting an IDE, and the reporting and consent obligations that come with one, onto a study already underway. An SRD filed before the first subject enrolls removes that particular bet. It doesn't change the underlying risk criteria, and it doesn't relieve the IRB of its own review — it just means the sponsor isn't staking an entire dataset on one IRB's reading of a definition FDA has the final say over.

A narrower tool than a Pre-Submission, on purpose

A Pre-Submission is built to carry a specific, answerable strategy question, and a sponsor can reasonably use one to raise almost anything about a planned study or submission, risk category included. An SRD trades that breadth for precision: it doesn't invite feedback on endpoints, statistical plan, or study design merits, and asking FDA to weigh in on those inside an SRD is asking the wrong mechanism to do a Pre-Sub's job. The reverse mistake costs something too — folding a risk-category question into a broader Pre-Sub can get an informal comment where the study actually needed FDA's own written SR/NSR/exempt determination on the record.

Where this meets the file

A study-startup tracker built around this distinction needs the sponsor's own initial risk assessment, the IRB's determination and its date, and — where one was filed — the SRD request and FDA's own written reply, kept as three separate fields rather than one collapsed “risk determination” line, since the three can disagree with each other and the file needs to show which one actually governed enrollment. An IDE risk-determination tracker built around this exact chain is previewed in the launch catalog. If your program sequences this differently, the shelf takes that correction directly.