How to Monitor FDA MAUDE for Adverse Events

What is MAUDE?

MAUDE — the Manufacturer and User Facility Device Experience database — is the FDA's repository of medical device reports (MDRs) describing suspected device-associated deaths, serious injuries and malfunctions. Reports come from mandatory reporters (manufacturers, importers and device user facilities such as hospitals) and voluntary reporters (health professionals, patients and consumers). It holds records from the early 1990s to the present, and the FDA receives several hundred thousand new MDRs every year.

Why monitor it

For a medical device manufacturer, MAUDE monitoring serves several purposes at once:

• Post-market surveillance. It's a primary external data source feeding your PMS system under both FDA 21 CFR 820 and EU MDR — see our guide to EU MDR Articles 83–92.
• Signal detection. A cluster of similar reports can be the earliest warning of an emerging issue with your device — or a component or material you rely on.
• Competitor and category intelligence. Because reports are public, you can watch similar devices (by product code) for problems that may affect your own design.
• Evidence for your reports. MAUDE findings feed the analyses behind your PMS Report or PSUR and any trend reporting.

Three ways to access MAUDE

The web interface is fine for the occasional search, but any real monitoring programme is built on the openFDA API or the bulk files, because they can be queried on a schedule.

Step 1: Set your monitoring scope

The single most common mistake is searching by one brand name and assuming you've covered everything. Define your scope across three axes:

• Brand / trade names — including old names, sub-brands and common misspellings. Name variations directly affect what a search returns.
• Manufacturer names — including legacy entities, acquired companies and distributor names that may appear on reports.
• FDA product codes — the three-letter classification codes are the most reliable way to capture your device type, and to widen the net to similar devices from other makers.

A robust scope watches both your own devices and the product-code neighbourhood around them.

Step 2: Build a repeatable workflow

Monitoring is a process, not a one-off search. A workable loop looks like this:

1. Query on a schedule. Because the API refreshes weekly, a weekly pull keyed to date_received catches new reports promptly.
2. De-duplicate. A single event can generate an initial report plus follow-ups; group them by report number so you don't count one event many times.
3. Triage by event type. Separate deaths and serious injuries from malfunction-only reports, and read the narrative field for context.
4. Assess for signals. Compare against your baseline — is this a new failure mode, or a rise in an existing one?
5. Document and route. Log findings into your complaint/PMS system, and where warranted, feed CAPA, trend reporting and your PSUR/PMSR.

Using the openFDA API

The device adverse-event endpoint is:

https://api.fda.gov/device/event.json

The most useful search fields are device.brand_name, device.manufacturer_d_name, device.device_report_product_code, event_type (Death / Injury / Malfunction), and date_received. A query for one manufacturer's reports received in a date window looks like:

https://api.fda.gov/device/event.json?search=
device.manufacturer_d_name:"ACME+MEDICAL"
+AND+date_received:[20260101+TO+20260630]&limit=100

To get counts instead of raw records — for example, reports by event type — use the count parameter:

...&count=event_type

Limitations you must respect

MAUDE is a passive surveillance system, and the FDA is explicit about its limits. Reading it wrong is its own compliance risk:

• It doesn't prove causation. A report existing is not an admission that the device caused the event.
• You can't calculate rates from it. There's no denominator (units in use), so raw counts can't establish incidence or be used to rank devices against each other.
• Data is often incomplete or unverified, and can be duplicated, delayed or biased. Some text is redacted as (b)(4) or (b)(6).
• There's reporting lag. Manufacturer MDRs arrive on FDA timelines (generally within 30 days, sooner for the most serious events), so recent weeks are always incomplete.

From monitoring to compliance

MAUDE is one feed among several — recalls, clinical-trial registries, literature and the open market all matter too — but it's a high-value one. The challenge is sustaining it: querying every source on a schedule, de-duplicating, cross-referencing each hit against your portfolio, and keeping a timestamped record so that when an auditor asks "how did you know, and when?", you can answer.

Done manually across disconnected tools, that's exactly where gaps appear. Automating it — continuous querying of MAUDE and the other sources, cross-referenced to your device lines and logged with an evidence trail — turns ad-hoc searching into a defensible, always-current surveillance layer that feeds your PMS plan, PSUR/PMSR and trend reporting.