The FDA's Manufacturer and User Facility Device Experience (MAUDE) database is the single largest repository of medical device adverse event reports in the world. If you work in regulatory affairs, quality assurance, or post-market surveillance for a medical device company, MAUDE is not optional — it is the foundation of your safety monitoring program.
Yet despite its importance, MAUDE remains one of the most misunderstood and underutilized tools in the regulatory professional's toolkit. This guide covers everything your team needs to know to use it effectively.
What Is the MAUDE Database?
MAUDE is a publicly accessible database maintained by the FDA's Center for Devices and Radiological Health (CDRH). It contains medical device reports (MDRs) submitted under the mandatory reporting requirements of 21 CFR Parts 803 and 804. These reports document adverse events involving medical devices, including deaths, serious injuries, and malfunctions.
The database has been collecting reports since 1991 and currently contains millions of records. Each report includes structured fields — device name, product code, event type, patient outcome — along with unstructured narrative text describing what happened.
Who Reports to MAUDE?
Three categories of reporters are legally required to submit MDRs to the FDA:
Manufacturers must report when they become aware that their device may have caused or contributed to a death or serious injury, or when they learn of a malfunction that would be likely to cause or contribute to a death or serious injury if the malfunction were to recur. Under 21 CFR 803.50, manufacturers must submit reports within 30 calendar days, or within 5 days for events requiring remedial action.
Device user facilities — primarily hospitals and nursing homes — must report device-related deaths to both the FDA and the manufacturer within 10 working days. Serious injuries must be reported to the manufacturer (and to the FDA if the manufacturer is unknown) within 10 working days.
Importers must report deaths and serious injuries to the FDA and the manufacturer within 30 calendar days.
Voluntary reports from healthcare professionals and patients are also accepted and appear in MAUDE, though they are not subject to the same regulatory requirements.
How to Search MAUDE
The FDA provides several ways to access MAUDE data:
The MAUDE web interface (accessdata.fda.gov/scripts/cdrh/cfdocs/cfmaude/search.cfm) allows basic searches by product class, event type, manufacturer, date range, and keywords. It is functional but limited — searches can be slow, results are paginated with no bulk export, and the interface has not been substantially updated in years.
The openFDA API (open.fda.gov) provides programmatic access to MAUDE data in JSON format. This is the preferred method for any systematic surveillance program. The API supports complex queries, date-range filtering, and returns structured data that can be processed programmatically. Rate limits apply (240 requests per minute with an API key), but for most surveillance workflows this is more than sufficient.
Bulk download files are available from the FDA's MAUDE download page. These are pipe-delimited text files updated quarterly and are useful for historical analysis or building local databases.
For most regulatory teams running ongoing surveillance programs, the openFDA API offers the best balance of flexibility, timeliness, and automation potential.
The Limitations You Need to Know
MAUDE is indispensable, but it is far from perfect. Understanding its limitations is critical to using it correctly and defending your surveillance methodology during audits.
Data lag is significant. Reports can take 4-8 weeks to appear in MAUDE after submission, and some take considerably longer. The FDA processes reports in batches, and the openFDA API may lag the MAUDE web interface by additional days. This means MAUDE should never be your only source for emerging safety signals — it is a lagging indicator.
Duplicate reports are common. When both a manufacturer and a user facility report the same event, two separate records appear in MAUDE with different MDR report numbers. The FDA does not systematically deduplicate these records. For trending and signal detection, this means raw report counts can overstate the true number of unique events.
Narrative quality varies enormously. Some MDR narratives contain detailed clinical descriptions with patient demographics, timeline of events, and root cause analysis. Others contain a single sentence like "Device malfunctioned." The unstructured narrative field is where the most valuable clinical information lives, but extracting insights from thousands of variable-quality narratives at scale is a genuine challenge.
Causation is not established. An MDR in MAUDE indicates an adverse event associated with a device, not that the device caused the event. The FDA explicitly states that MAUDE data alone cannot be used to establish incidence rates or draw causal conclusions. Nevertheless, MAUDE data is the starting point for signal detection, which may then trigger more rigorous epidemiological investigation.
Product code classification can be inconsistent. Devices may be classified under different product codes by different manufacturers, and product codes can change over time. When setting up a monitoring program, it is worth verifying that your product code coverage captures all relevant devices.
Using MAUDE Effectively for Post-Market Surveillance
Given these limitations, here is how high-performing regulatory teams use MAUDE:
Establish a regular cadence. Weekly or biweekly searches are the standard for active surveillance programs. Monthly is the minimum for lower-risk devices. Document your search frequency in your post-market surveillance plan.
Monitor by product code, not just your own brand. Tracking adverse events across your entire product code gives you competitive intelligence, early warning of class-wide issues, and the ability to benchmark your device's safety profile against the field.
Build keyword strategies for narrative search. Generic product code monitoring will catch everything, but keyword searches on the narrative text can surface specific failure modes you are tracking. Combine product code filters with terms related to known risks identified in your risk management file.
Track trends, not individual reports. Any single MDR may be an outlier, a duplicate, or poorly documented. The value of MAUDE surveillance is in identifying patterns: increasing report frequency for a specific failure mode, new types of injuries not previously observed, or emerging issues with a specific lot or model.
Document everything. Your post-market surveillance plan should describe your MAUDE search methodology, frequency, and how results are reviewed and escalated. This documentation is essential for FDA inspections and EU MDR compliance (Article 83 requires systematic PMS including proactive market surveillance).
Why Manual MAUDE Searching Is Unsustainable
Here is the uncomfortable truth: the approach described above — regular searches, narrative analysis, trend tracking, competitive monitoring — is the right approach. But doing it manually is unsustainable for any team monitoring more than a handful of product codes.
A single product code can generate hundreds of new reports per month. Reading and classifying the narratives, identifying duplicates, tracking trends over time, and documenting the entire process for audit readiness — this work can consume an entire FTE for a modest device portfolio.
This is exactly the problem that modern surveillance platforms are designed to solve. Automated tools can pull new MAUDE reports on a scheduled basis, apply natural language processing to classify narratives by failure mode and severity, flag emerging trends, and maintain the audit trail that regulators expect.
DeviceWatch, for example, connects directly to the openFDA API to ingest new adverse event reports weekly, then uses AI to analyze the clinical narratives and generate structured summaries. The human review step remains — regulatory professionals must review and acknowledge the AI-generated analysis — but the hours of manual search and data entry are eliminated.
Looking Ahead
The MAUDE database is evolving. The FDA's ongoing migration to the Adverse Event Management System (AEMS) will consolidate MAUDE and other reporting databases into a single platform with improved data quality, better deduplication, and more structured reporting fields. This transition is expected to improve the data that surveillance teams work with, but it also means that surveillance systems need to be adaptable to new data formats and APIs.
Whether you are just establishing your post-market surveillance program or looking to modernize an existing one, understanding MAUDE — its strengths, its limitations, and how to work with it systematically — is non-negotiable. The database is imperfect, but it remains the most comprehensive source of real-world device safety data available.