The Workbench · Craft
A device trial's registration duty isn't FDA's own
This blog has already covered how most investigational device studies never produce the document most people mean by an IDE application at all — a nonsignificant-risk study runs on an IRB's own approval instead, with FDA rarely seeing the file unless something goes wrong. That fork, and the separate mechanism for locking in FDA's own risk view before enrollment starts, both answer to the same regulator, under 21 CFR Part 812. A study that clears both of those questions cleanly can still owe a registration and results-reporting duty to a different federal agency entirely, on a clock that doesn't wait for anything FDA decides.
The duty runs through HHS, not FDA
Section 801 of the Food and Drug Administration Amendments Act of 2007, codified at 42 U.S.C. 282(j) and implemented at 42 CFR Part 11, doesn't hand its registration and results-reporting duty to CDRH. It runs to the National Library of Medicine, part of the National Institutes of Health, through ClinicalTrials.gov — a database that exists independently of FDA's own device review files and answers to a different agency inside HHS entirely. A study can be moving cleanly through the IRB-only track most nonsignificant-risk device studies actually run on and still owe a separate registration record to an office that has never seen the IDE file, because the two duties were written into two different statutes administered by two different parts of the same department.
An applicable device clinical trial is a narrower population than “every device study”
42 CFR 11.22 defines an applicable device clinical trial as a prospective clinical study of health outcomes, comparing an intervention using a device subject to section 510(k), 515, or 520(m) of the FD&C Act against a control, in human subjects — along with pediatric postmarket surveillance FDA requires under section 522. The definition then excludes two specific categories on its own terms: a small clinical trial run to determine a device's feasibility, and a trial testing a prototype device where the primary outcome measure is feasibility itself rather than a health outcome. Neither exclusion turns on a subject count the regulation itself states — it turns on what the study is actually designed to measure. A study large enough to look like a pivotal trial can still fall outside the definition if its own primary outcome is feasibility; a small one can still fall inside it if the outcome it's actually measuring is a health outcome rather than whether the device basically works.
Registration runs on its own twenty-one-day clock
42 CFR 11.24 requires the responsible party for an applicable clinical trial to submit registration information no later than 21 calendar days after the first human subject is enrolled. That clock starts at enrollment and runs on its own terms — it doesn't pause for an IDE still under FDA review, and it doesn't wait on a Study Risk Determination a sponsor may have filed to lock in FDA's own view of the study's risk category before enrollment even started. A sponsor tracking one master start date for “the study's regulatory clock” has collapsed two obligations that run independently of each other, on two different calendars, to two different recipients.
Results can still be owed for a device that's never cleared
42 CFR 11.44 lets a responsible party delay submitting results by certifying, before the deadline, that the trial studies a device already approved, licensed, or cleared for a different use, with an application covering the studied use filed or expected within a year. Absent that certification, the standard deadline is no later than one year after the trial's primary completion date. A device that never gets cleared at all doesn't get an automatic pass on the strength of that fact alone: for a trial with a primary completion date on or after January 18, 2017, 42 CFR 11.48 still requires submitting the specified results information even where the device remains unapproved as of that date. A results-reporting record that treats “device not yet cleared” as a reason the obligation doesn't exist has confused a possible basis for delay with an exemption the rule never actually grants.
Where this meets the file
A device-trial compliance file needs the applicable-clinical-trial determination as its own tracked field, reasoned against 11.22's own criteria rather than copied from whatever risk determination the IDE file already reached, alongside the registration date measured from first-subject enrollment and the results deadline measured from primary completion — three dates against two different regulators, none of them derivable from the others. A ClinicalTrials.gov compliance worksheet built around that structure is previewed in the launch catalog. If your program tracks this determination differently, the shelf takes that correction directly.
The Regulatory Toolkit launches soon — a free shelf of source-mapped templates, checklists and browser-only tools for regulatory teams. Get one email when it opens, or contribute a template.