A MAUDE report proves it was filed

Search MAUDE for a product code and the results come back as a list of rows, each one the same shape: a device name, an event type, a narrative, a date. Nothing about the layout suggests that the rows behind it were filed by fundamentally different kinds of people, under fundamentally different obligations to get it right. openFDA carries that distinction in a single field — source_type — that a keyword search never surfaces on its own.

Two forms, two different obligations

21 CFR 803.11 splits adverse-event reporting by form. A manufacturer, importer, or device user facility — a mandatory reporter, working under a duty this part of the CFR defines and enforces — submits on Form FDA 3500A, the mandatory-reporting form, on a clock the rule sets. A health care professional, a patient, or a consumer who saw something and decided to say so submits on Form FDA 3500 — the voluntary-reporting form — on no clock at all, because nothing obligated them to file in the first place.

The database doesn’t grade the difference

Those two paths carry very different evidentiary weight. A manufacturer’s mandatory report is filed by a party legally accountable for the accuracy of what it submits, after whatever investigation the manufacturer’s own procedures require. A voluntary report is filed by someone who owed FDA nothing — no investigation duty, no follow-up requirement, no penalty structure attached to getting the causal story wrong. MAUDE posts both to the same table, in the same row format, distinguished only by the source_type value sitting in a field most searches never filter on.

A source_type of Manufacturer means a legally accountable party investigated and reported. A source_type of Voluntary means someone who owed FDA nothing decided, on their own, to say something. MAUDE renders both at the same width, in the same font, in the same row.

The practice

Reading a product code’s MAUDE volume as one number hides a composition question worth asking first: how much of that count is mandatory reporting — a manufacturer or user facility surfacing what its own monitoring caught — and how much is voluntary, which can move for reasons that have nothing to do with the device getting worse. Media coverage, a lawsuit, or a single widely shared account can drive a spike in voluntary reports for a device whose failure rate hasn’t moved at all. A spike concentrated in manufacturer-sourced reports is a different signal, closer to the device’s own detection pipeline finding more to report. Pulling source_type before drawing a conclusion from a MAUDE count is the difference between reading what changed in the device and reading what changed in who was paying attention.

MAUDE sits outside FDA Radar’s current source coverage — noted, with the reason, on what we monitor and what we don’t. Reading it well is a skill a regulatory or quality professional needs regardless of which tool is watching — regulatory intelligence and method, never a determination about any specific device’s safety, which is a judgment for whoever owns that product’s file.