MAUDE Monitor
Methods & limitations
Last checked: Nov 19, 2026Source: openFDA
Where the data comes from
Adverse event reports come from MAUDE, FDA's Manufacturer and User Facility Device Experience database, read through openFDA, the public data service FDA runs for its own device databases. Recalls come from openFDA's device recall dataset. We do not edit the records. Brand names, manufacturer names and problem terms appear as FDA recorded them.
- The tool covers reports FDA received on January 1, 2015 or later.
- Reports are dated by the day FDA received them, not the day the event happened. The two can be months or years apart.
- We ask openFDA again at most every 6 hours, so this tool can lag FDA's own database by up to a day beyond that.
- Each page shows a "Last checked" date and the date of the newest record, so you can see how fresh the view is.
- If openFDA does not answer, the page says so and shows what it has. It does not show stale numbers as if they were current.
What a report is, and is not
A MAUDE report is one account of one event, filed by a manufacturer, an importer, a hospital or other user facility, a clinician or a patient. Manufacturers must report when they learn that a device may have caused or contributed to a death or serious injury, or malfunctioned in a way that could do so if it happened again. That is a low bar on purpose. A report means someone raised a possibility. It does not mean FDA or anyone else confirmed it.
- There is no denominator. The data has no count of devices sold, implanted or used. A device with 5,000 reports and ten million uses may perform better than one with 50 reports and a thousand uses. Nothing here can tell you which.
- Volume follows attention. Recalls, safety notices, lawsuits and news coverage all prompt reports of older events.
- Duplicates are common. A manufacturer and a hospital can each report the same event, and follow-up reports can be filed as new ones.
- Summary reporting hides volume. Under programs such as Voluntary Malfunction Summary Reporting, a manufacturer can file many malfunctions as one quarterly report. Counts for those device types are understated and can arrive in batches.
- Manufacturers differ. A company that reports thoroughly looks worse in a count than one that does not.
For these reasons this tool never ranks devices or companies by safety, and no number on it should be read that way.
How we count
- Event type. FDA records each report as a Malfunction, an Injury or a Death, as the reporter described it. We show those three and group everything else as Other. Other includes the types FDA labels "Other" and "No answer provided", and reports where the type was left blank.
- Reports that name several devices. In the totals and the monthly chart, a report counts once. In rankings by device type, manufacturer or brand, a report that names several devices counts once for each. Those rankings can therefore add up to more than the total.
- Manufacturer names. FDA stores names as each reporter typed them, so one company appears under many spellings. We merge spellings that reduce to the same name, and for the companies on our public-company list we also merge the names of known subsidiaries. Merging is imperfect. A company page can miss reports filed under a name we did not match, and a common name can pull in an unrelated firm.
- Brand names. Brand names are free text. One product usually appears under several spellings, model numbers and sizes, and we only merge differences in capital letters.
- Device type. Device types are FDA product codes. A product code newer than our copy of FDA's classification list shows as "Product code" with its three letters and no name.
- Specialty. A specialty view uses the product codes FDA assigns to that review panel, or the list of product codes behind a sub-specialty. Reports without a recognized product code are left out of specialty views.
- Problem terms. Device problems and patient problems are FDA's coded terms, shown as written. A report can carry several, and terms such as "Adverse Event Without Identified Device or Use Problem" are common.
- Report lists. Lists show 25 reports a page, newest first, and stop after 1,000 reports. Counts and charts are not limited that way.
Changes in report volume
The changes in report volume page compares two windows for each device type:
- Recent: the last 3 complete calendar months in FDA's data. The month of the newest report counts as complete only if that report falls in its last few days.
- Earlier: the 12 calendar months before that.
- We divide each window's reports by its number of months and compare the two monthly averages. A device type is listed as an increase when the recent average is 25% or more above the earlier one.
- Only the 100 device types with the most reports in the recent window are checked, and only those with at least 30 recent reports are shown. Small counts swing too much to compare.
- A device type with fewer than 12 earlier reports gets no percentage. It is listed separately as having too few earlier reports to compare.
This is arithmetic on report counts, not a statistical test, and it makes no adjustment for sales, seasonality or reporting practices. An increase often reflects a change in how or when reports were filed: one manufacturer clearing a backlog, a move into or out of summary reporting, or a recall that prompts reports of older events. Treat an increase as a prompt to read the underlying reports on FDA's site, and nothing more.
The newest months can still grow as FDA processes reports, so the recent window may be an undercount.
Recalls
Device type, company and sub-specialty pages list recent records from FDA's device recall database for context.
- FDA lists each recalled product, model or lot group as its own record, so one recall action can appear several times.
- "Recall" is FDA's term for any correction or removal. Many recalls are labeling changes, software updates or instructions to users, and the device stays in use.
- This dataset does not give the recall class (I, II or III), so we do not show one. Each record links to FDA's page, which does.
- Recalls are matched to a company by the recalling firm's name, with the same limits as manufacturer names above.
- We show the newest records by the date the firm started the recall. Reason and cause text is shortened to fit.
What this tool leaves out
- The reporter's description of the event. We show coded fields only. Each report links to FDA's full record, which has the narrative text.
- Anything about patients. FDA redacts identifying information. We do not try to add any back, and no one using this tool should.
- Reports FDA holds elsewhere. Some older summary reports, and reports made under past alternative programs, are not in MAUDE.
- Rates, risk scores and comparisons of safety. The data cannot support them.
Nothing here is medical, legal or investment advice, and it is not a recommendation of any company or security. Clinical decisions about a device belong with the treating physician and the device's labeling.
For information only. Nothing on this page is investment, legal, tax or medical advice, and nothing here is an offer to sell or a solicitation of an offer to buy any security. The data comes from public government sources, which can be incomplete, delayed or wrong. Check the original record before relying on it.