A qualified tool has a context, not a blanket

A sponsor building a case for a new device sometimes needs someone else’s measurement — a biomarker, a clinical outcome assessment, a computational model — and proving that measurement means what it claims to mean is its own research project, before the device research even starts. FDA built a program to let a sponsor skip that proof, once, for everyone, inside one named use: the Medical Device Development Tools program. What it hands back is not a blanket endorsement of the tool. It is a qualification scoped to the context FDA actually reviewed.

A qualification decision, not a clearance

FDA announced MDDT the way it announces most voluntary programs — with a draft guidance, not a new regulation: Medical Device Development Tools; Draft Guidance for Industry, Tool Developers, and Food and Drug Administration Staff, published at 78 FR 68459 on November 14, 2013, under docket FDA-2013-D-1279. CDRH finalized it in 2017 and reissued the current version on July 17, 2023, still under the same docket. No statute created the program; CDRH built it, and keeps rebuilding it, by guidance alone.

The process runs in two phases, and either can end the conversation early. In the proposal phase, a tool developer submits the tool, a description of what it measures, and a Qualification Plan — the evidence it intends to gather to support reliance on the tool for one named purpose. CDRH decides, typically within about ninety days, whether that purpose is worth building evidence for. Only then does the developer move to the qualification phase, executing the plan and submitting the resulting data as a full qualification package. Neither phase carries a user fee.

The qualification is the context, not the tool

What comes out the other end is scoped to a Context of Use (COU) — FDA’s term for a written statement of exactly how the tool is meant to be used, for what purpose, and at what point in device development. A qualified COU means a CDRH reviewer accepts the tool for that purpose without re-litigating whether it works. It does not mean the tool is validated, period, for whatever else someone might want to use it for.

The qualification decision for the Image Viewer Integrity Evaluation System, a non-clinical assessment model qualified under submission U240030/S001, shows how literally FDA draws that line. The tool compares a third-party whole-slide-image viewer against a predicate viewer, pixel by pixel, to determine whether the two decode and render the same file identically. FDA’s qualification covers exactly that comparison — the viewer component. It says nothing about the computer the viewer runs on or the display it renders to, both of which sit outside the qualified COU entirely, even though both affect what a pathologist actually sees on screen.

A qualified tool is a verified answer to one named question. Ask it a different question and the qualification doesn’t travel with the name.

What citing it buys a sponsor

Any sponsor — not only the one who developed the tool — can rely on a qualified MDDT inside its qualified COU, citing it in a 510(k), De Novo, PMA, or HDE submission by its MDDT submission number, either in the cover letter or in Section F of FDA’s Premarket Review Submission Cover Sheet, Form 3514. FDA publishes a Summary of Evidence and Basis of Qualification Decision for every qualified tool, naming the COU in writing, so the scope is never a matter of asking the reviewer what it meant.

The practice

Finding a qualified tool that sounds like a match for your device program is the easy part; FDA’s qualified-tools list and each tool’s SEBQ are public. The work is reading the COU statement as narrowly as FDA wrote it — the device type, the measurement, the development phase it was qualified for — and confirming your intended use sits entirely inside that boundary, not adjacent to it. A tool qualified for one imaging modality, one population, or one stage of development does not extend itself to a neighboring one by resemblance. Whether a specific tool’s qualified COU actually covers your specific intended use is a determination for your own regulatory and nonclinical team, read against the SEBQ’s own language; this is a description of how the qualification is scoped, never a determination that any tool, named here or anywhere else, fits your program. FDA Radar’s five monitored sources don’t carry a feed for MDDT qualification decisions today — per what we monitor and how often, it isn’t one of the five, and it isn’t on that page’s own list of named gaps either: unlike a pending 510(k) or a Form 483, there is no structured feed yet to decide whether to add.