Guide · Post-market data
How to search MAUDE, and read what it says
MAUDE (Manufacturer and User Facility Device Experience) is FDA's database of medical device reports: adverse events and malfunctions reported by manufacturers, importers, device user facilities and voluntary reporters. openFDA's copy held 26,136,889 reports in its export of 2026-09-15. Searching it well comes down to three things: search by product code, read the counts as counts, and check for duplicates.
1. Find the product code
Every device type FDA regulates has a three-letter product code, such as DQY for percutaneous catheters or DXN for non-invasive blood pressure monitors. Look it up from a competitor's 510(k) in PredicateScope, browse the product code index, or check the Device Data Snapshot.
2. Search by the code, not the brand
Find the device's three-letter FDA product code first (from its 510(k) or the classification database), then search MAUDE by that code rather than by brand name. Brand and model names are free text and vary from report to report; the product code is a coded field, so a code search finds the reports a name search misses.
MAUDEScope runs that search against openFDA and shows, for one product code, how many reports there are, who filed them, the event types (malfunction, injury, death), how late reports arrive, which fields are filled in, and which quarters stand out. The exact openFDA query is printed under every figure so anyone can re-run it.
3. Read the counts as counts
No. MAUDE contains counts of reports, not rates: there is no public denominator of devices sold or used, reporting is incomplete, one event can be reported more than once, and a report does not establish that the device caused the event.
The useful questions are about the record, not the device: Is the volume of reports changing? Which problem codes dominate? Did a spike follow a recall or a change in who reports? How complete are the reports?
4. Watch for duplicates
The same event can arrive as several reports: from the manufacturer and the hospital, or as an initial report and follow-ups. In our deduplicated slices, built with a clustering method that is still a draft without measured accuracy, the share of reports that looked like duplicates ranged from 4.9% to 25% across 5 product codes. For DQO (diagnostic intravascular catheters), 20,198 reports clustered into an estimated 18,102 events. Snapshot 2026-09-15, openFDA data as of 2026-09-08.
| Code | Device | Reports | Estimated events |
|---|---|---|---|
| LIT | Catheter, Angioplasty, Peripheral, Transluminal | 35,907 | 33,808 |
| DQO | Catheter, Intravascular, Diagnostic | 20,198 | 18,102 |
| EOB | Nasopharyngoscope (Flexible Or Rigid) | 8,127 | 7,732 |
| LRC | Instrument, Ent Manual Surgical | 561 | 524 |
| OWO | Drug-Eluting Sinus Stent | 48 | 36 |
5. Write it up so it can be checked
For a risk file or a clinical evaluation, record the query, the date of the data and the counts, and keep the statements to what the record shows. The Device Dossier exports that as a traceable document, and the MAUDEScope method page explains every figure.