What is the FDA MAUDE database?
The FDA MAUDE database (Manufacturer and User Facility Device Experience) is the public record of medical device adverse event reports submitted to the U.S. Food and Drug Administration. Manufacturers, importers, device user facilities, and the public file reports when a device may have caused or contributed to a death, serious injury, or malfunction. MAUDE is where those reports become searchable.
It is run by the FDA's Center for Devices and Radiological Health and is free to use. Reports typically appear weeks to months after the event, so MAUDE is a trailing record rather than a real-time feed.
What MAUDE contains (and what it does not)
Each MAUDE record describes one report: the device involved, the manufacturer, an event type (death, injury, or malfunction), dates, and a free-text narrative describing what happened. Millions of reports stretch back to the early 1990s.
What MAUDE is not:
- Not a recall database. Recalls are tracked separately by the FDA. A device can have many MAUDE reports and no recall, or a recall with few reports.
- Not a clearance database. 510(k) and PMA approval records live in their own FDA systems.
- Not a complete census of harm. MAUDE is the surface of reports that were submitted, which is different from everything that actually happened.
Who files MAUDE reports
Reporting obligations differ by party. Manufacturers and importers must report under federal regulation. Device user facilities such as hospitals report deaths and serious injuries. Clinicians and patients can file voluntary reports. Because the rules differ, the same kind of event can be reported at very different rates depending on who witnessed it, which is the first reason raw counts can mislead.
How to search MAUDE
There are three reasonable front doors, each good for different work:
- The FDA web search on accessdata.fda.gov is the canonical interface. It supports brand name, manufacturer, product code, date range, and event type. It is authoritative but slow to paginate.
- openFDA exposes MAUDE as a JSON API at
/device/event.json. Use it when you need to script queries or pull more than a few hundred records. - A curated lookup layer like Claripulse, which normalizes names, deduplicates reports, and groups devices so one device reads as one device.
For a step-by-step walkthrough with real queries, see how to search the FDA MAUDE database.
The fields that actually matter
MAUDE has dozens of fields, most of them sparse. The ones that carry almost every real query are:
brand_nameandgeneric_name: the marketed and FDA-side device names.manufacturer_d_name: the manufacturer, with spelling that varies across filers.product_code: the three-letter classification code, the most reliable way to search a device category.event_type: death, injury, malfunction, or other.date_received: when the FDA logged the report, the field to use for date ranges.mdr_text: the free-text narrative, where the clinical story lives.
The limits you have to know
MAUDE answers "what was reported" well and "how often does this happen" poorly. Five structural limits explain why:
- No denominator. A device with forty reports tells you nothing without knowing how many were in use. Rising sales raise counts on their own.
- Underreporting. Reporting is uneven across parties and event types, so counts undercount true events by an unknown amount.
- Inconsistent names. One device may appear under several brand and manufacturer spellings, so a naive search misses reports.
- Duplicate filings. A single event can produce an initial report, supplements, and a final report under different keys.
- Cross-code devices. The same physical device can be reported under multiple product codes, splitting the picture.
These are exactly the gaps a normalized layer is built to close. See how safety signal detection works for how analysts turn messy counts into something defensible.
MAUDE vs other FDA data sources
MAUDE is one of several FDA systems. Recalls track corrective actions after a problem is confirmed. 510(k) and PMA records track market clearance and approval. Adverse event narratives in MAUDE are the earliest public signal that something may be wrong, often before a recall exists, which is why post-market teams watch it closely. To see how the pieces fit together, read our guide to post-market surveillance, or skip straight to the free MAUDE lookup tool to browse a category.