How to Search the FDA MAUDE Database for Medical Device Adverse Events

2026-10-08 · Caduvo Team

Learn how to search the FDA MAUDE database by product code, brand, and date. Understand report types, FOIA redactions, and the limits of MDR data.

Every year, the Food and Drug Administration receives several hundred thousand to over two million medical device reports (MDRs) describing suspected device-related deaths, serious injuries, and malfunctions. Medical device manufacturers, quality managers, and regulatory affairs teams turn to the Manufacturer and User Facility Device Experience (MAUDE) database to understand how marketed technologies perform in clinical settings. When evaluated systematically, MAUDE reveals real-world failure modes, informs design inputs, and strengthens risk management files.

Yet searching MAUDE presents operational hurdles. The public search portal caps web queries at 500 results, truncates historical data beyond ten years, and relies on passive submissions. Extracting actionable regulatory intelligence requires understanding how search filters function, how mandatory and voluntary reports differ, and what MAUDE data can and cannot substantiate.

How to Query MAUDE: Search Fields, Operators, and Result Limits

The public FDA MAUDE search interface offers simple and advanced search modes. The database updates monthly to include reports received through the final day of the previous calendar month. It displays records submitted within the trailing ten years. Older reports, legacy records from the Device Experience Network (DEN, covering over 600,000 reports from 1984 through 1996), and more than 6 million Alternative Summary Reporting (ASR) records from 1999 through April 2019 are archived on the FDA MDR Data Files webpage.

The primary technical constraint of the web interface is its hard ceiling of 500 records per query. If a query matches 501 records, the web system drops everything beyond the first 500 without indicating which records were omitted.

To search the database effectively, configure the advanced fields strategically:

Product Code: Search using the three-letter FDA product code. This captures all cleared or approved devices in a classification regardless of brand name changes or corporate acquisitions. For high-volume product categories, a multi-year query will hit the 500-result cap immediately.
Date Report Received by FDA: Format dates strictly as mm/dd/yyyy. To prevent query truncation in high-volume product codes, break queries into 30-day or 90-day intervals.
Brand Name and Manufacturer: The web portal runs literal character matches. Searching "Medtronic" will miss records filed under legal entity names like "Medtronic Sofamor Danek" or "Medtronic MiniMed" unless you search those exact strings or run separate queries. Account for corporate variations, punctuation, and alternate spellings.
Simple Search Boolean Operators: In the simple search box, you can search single words, exact phrases in quotes (such as "electromechanical pump"), or multiple terms separated by and (such as bruise and throat and intubation).
UDI-DI: Entering a Unique Device Identifier Device Identifier isolates reports tied to that specific model configuration, though UDI reporting in historical MDRs remains variable.

For enterprise post-market surveillance projects requiring analysis of tens of thousands of records, bypass the web portal entirely. Download the monthly zipped raw data files or query the openFDA endpoint.

Reading an MDR: Event Types, Reporter Classifications, and Redactions

Every record in the MAUDE database stems from an adverse event filing containing publicly available information from completed MEDWATCH Form 3500 submissions. Under 21 CFR Part 803, distinct entities operate under different statutory reporting mandates:

Manufacturers and Importers: Must submit reports when they become aware of information reasonably suggesting that one of their marketed devices may have caused or contributed to a death or serious injury, or has malfunctioned and would be likely to cause or contribute to a death or serious injury if the malfunction recurred. Manufacturers submit reports of deaths, serious injuries, and malfunctions to the FDA. Importers send reports of deaths and serious injuries to both the FDA and the manufacturer, but send malfunction reports only to the manufacturer.
Device User Facilities: Hospitals, ambulatory surgical facilities, nursing homes, and outpatient diagnostic or treatment facilities must report suspected device-related deaths to both the FDA and the manufacturer (if known). They must report serious injuries to the manufacturer, or to the FDA if the manufacturer is unknown.
Voluntary Reporters: Healthcare professionals, patients, and consumers submit reports voluntarily without mandatory legal formats.

When opening an individual report, review the specific data sections:

Event Type: Categorized as Death, Serious Injury, or Malfunction. When the FDA receives a follow-up or supplemental report, the database updates to display the most recent MDR information.
Device Problem Data: Contains Device Problem Code data derived from MEDWATCH Form Sections F10 and H6.
Patient Problem Data: Contains Health Effect - Clinical Code data, also derived from Form Sections F10 and H6.
Text Data: Contains textual narratives recorded across MEDWATCH Form Sections B5, H3, and H10.

As detailed on the FDA MAUDE database overview page, certain details are protected from public disclosure under the Freedom of Information Act (FOIA). The designation (b)(4) marks redacted trade secret or confidential business information, such as product composition. The designation (b)(6) marks redacted personal or medical files, such as a patient's date of birth. Certain patient demographic details—age, sex, weight, race, and ethnicity—remain unredacted when they do not compromise personal privacy.

Pay attention to exemption codes. Entries marked with RWDYYXXXXX indicate reports submitted under 21 CFR 803.19 alternative reporting exemptions derived from real-world data sources like registries, electronic health records, or claims data.

What MAUDE Data Cannot Tell You: The Denominator Problem

Teams often misuse MAUDE by treating raw report totals as statistical incidence metrics. The FDA's documentation on openFDA's MAUDE dataset clarifies that MDR data alone cannot be used to establish rates of events, evaluate changes in event rates over time, or compare event rates between devices. The number of reports cannot be interpreted in isolation to reach conclusions about the existence, severity, or frequency of device problems.

This restriction stems from several structural factors:

The Absence of a Denominator: MAUDE records the absolute number of reported events, but provides zero data on total units manufactured, distributed, implanted, or procedures performed. A device with 500 reported malfunctions across 2,000,000 clinical uses has an observed failure rate of 0.025%. A competing device with 50 reported malfunctions across 1,000 clinical uses has an observed failure rate of 5.0%. Evaluating raw counts alone leads engineering teams to false conclusions.
Under-Reporting and Reporting Bias: Passive surveillance captures an incomplete fraction of real-world events. High-profile recalls, safety alerts, or media coverage often trigger sudden surges in reporting for a specific device without any shift in physical failure frequency.
Duplicate Reporting: Because multiple entities hold reporting obligations, a single clinical incident can generate multiple entries: distinct records from the user facility, distributor, manufacturer, and voluntary submitters. In the raw downloadable files, all source records share an MDR REPORT KEY, but web queries do not aggregate them into a single incident.
Unverified Causation: Submitting an MDR is not an admission that the device or manufacturer caused or contributed to the event. Confirming causation is difficult if circumstances surrounding the event have not been verified or if the device has not been directly evaluated.

Using MAUDE in Design Inputs, Risk Management, and Post-Market Surveillance

Despite its epidemiological limitations, MAUDE serves as an engineering benchmark when applied to identify failure mechanisms rather than calculated event frequencies.

1. Generating Design Inputs and FMEA Controls

When developing a device, engineering teams must identify foreseeable misuse and mechanical failure modes under risk management procedures. Query the MAUDE database using the product code of your intended predicate device before finalizing your 510(k) clearance documentation or drafting a design failure mode and effects analysis (DFMEA).

If you are designing an electromechanical pump, an advanced search across competitor brands will show whether existing systems encounter motor stalls, fluid pathway occlusions, or battery casing fractures. Documented field failures feed directly into design inputs, allowing teams to establish verified bench-testing acceptance criteria before building verification prototypes.

2. Identifying Predicate Weaknesses

When evaluating an asset for predicate device selection, cross-referencing clearance history in the FDA 510(k) database against adverse events in MAUDE prevents regulatory missteps. If a candidate predicate has accumulated dozens of serious injury reports over the past 24 months due to connector fractures, FDA review staff during premarket review will evaluate that specific failure mode. Discovering this during predicate selection allows you to execute comparative bench testing in advance.

3. Post-Market Surveillance (PMS) Benchmarking

Under post-market surveillance requirements, manufacturers must assess whether a device's risk-benefit profile remains acceptable. Reviewing competitor narratives in MEDWATCH Form Sections B5, H3, and H10 reveals how other manufacturers evaluate device performance issues. When competitors describe engineering evaluations or root-cause findings, that data provides a point of comparison for internal complaint handling.

Consolidating these datasets manually requires processing millions of rows of monthly raw text files. Caduvo aggregates MAUDE event counts by product code alongside 58,225 FDA recall records, giving regulatory and market-access teams visibility into post-market device performance without spreadsheet limits.

Actionable Rules for MedTech Teams Searching MAUDE

Search by product code before brand name. Always query the three-letter FDA product code first. Brand names change after corporate transactions, and voluntary reporters regularly misspell proprietary trade names.
Date-slice queries that hit 500 records. If an FDA search returns exactly 500 records, the interface truncated your data. Narrow your search window to single quarters or individual months until result counts drop below 500 per query.
Read narrative fields, not just problem codes. Examine the text narratives recorded across MEDWATCH Form Sections B5, H3, and H10 alongside coded entries in Device Problem Data (Sections F10 and H6). These narrative sections provide technical context that categorical problem codes omit.
Account for FOIA redactions. Expect (b)(4) in place of confidential commercial details like product composition, and (b)(6) in place of identifiers like a patient's date of birth. Demographic fields like age, sex, weight, race, and ethnicity remain available.
Never present MAUDE counts as an incidence rate. In regulatory submissions, design dossiers, or marketing collateral, use MAUDE findings strictly as qualitative identification of observed failure modes, never as an incidence rate or proof of superior safety over a competitor.

Sources

General information, not regulatory, legal or reimbursement advice. Company and product names belong to their owners; a mention does not imply affiliation or endorsement. Report an error. Data and corrections policy.