Data · Device events
Medical device adverse event reports, by product code and quarter
Which kinds of medical devices draw the most adverse event reports to the FDA, and how those reports split into deaths, injuries and malfunctions, quarter by quarter. We count reports with the FDA's own openFDA API, read no report text or patient detail, and check the counts against a second query shape and the live totals.
What is in this dataset?
Adverse event reports the FDA received about medical devices, counted by device product code and calendar quarter from 2025 Q1: reports, deaths, injuries, malfunctions and other, for the 300 product codes with the most reports and for all devices together. Counted with the openFDA device event API (licence CC0); no report row, narrative, patient or manufacturer is ever read, only counts. Reports are filed up to weeks after an event and are driven by litigation and recalls as well as by harm, so a count is not an incidence rate.
Use it to see which device types draw the most adverse event reports and how that moves quarter to quarter: a sudden jump in injury reports for one product code is often the first public sign of a recall or a mass tort. Product-liability and mass-tort practices can watch the codes behind the devices they litigate. Pair it with the mdl dataset. Agents can look up the national total or a product code, compare quarters, or search a device by name.
- Records
- 2,079
- Look up by
- 7
- quarter
- Primary source
- 2,079
The letters are explained in what P, S, D and U mean.
Where does the data come from?
Counted with the openFDA device event API: one count call per quarter and event type (Death, Injury, Malfunction, Other, No answer provided) by device product code, plus an all-types call per quarter; only counts are requested, never a report row or narrative. The 300 product codes with the most reports are kept; names come from the FDA classification database. A gate re-queries random product-code quarters and the national totals live and ties each code's quarterly records to a single whole-window query.
- Period covered
- January 1, 2025 to September 30, 2026
- Places
- The whole US
- How often we refresh it
- Every month
- The source's next release
- November 2026
- Records read
- October 29, 2026
Official sources
- U.S. Food and Drug Administration (openFDA), by official API. CC0 1.0 (https://open.fda.gov/license/): openFDA data are free to use for any purpose, with no restriction..
- U.S. Food and Drug Administration (openFDA), by official API. CC0 1.0 (https://open.fda.gov/license/): openFDA data are free to use for any purpose, with no restriction..
- U.S. Food and Drug Administration (openFDA), by official API. CC0 1.0 (https://open.fda.gov/license/): openFDA data are free to use for any purpose, with no restriction. Used for product code names only..
- U.S. Food and Drug Administration (openFDA), by official web pages. The openFDA licence page itself (CC0 1.0)..
The sources the records cite most
- openFDA license (CC0): 2,079 records
- openFDA, device adverse event API (MAUDE): 2,079 records
What does a record look like?
-
All medical devices, 2026 Q2: 923,529 adverse event reports received by the FDA (3,477 death, 318,321 injury, 601,660 malfunction and 71 other).
Status P · checked October 29, 2026 · source: openFDA, device adverse event API (MAUDE)
-
All medical devices, 2026 Q1: 811,618 adverse event reports received by the FDA (4,022 death, 292,256 injury, 515,289 malfunction and 51 other).
Status P · checked October 29, 2026 · source: openFDA, device adverse event API (MAUDE)
The full records add the dates, the quoted source text and every field in the CSV and JSON files below.
What each of the 24 main columns means
jurisdiction- US for every record (the data are national)
level- product-code or us (all device product codes together)
productCode- FDA's three-letter device product code (empty for the national record)
deviceName- Device name for the product code from the FDA classification database
specialty- FDA medical specialty of the product code
deviceClass- FDA device class (1, 2, 3, or the code U/N where the FDA lists none)
quarter- Calendar quarter of the date the FDA received the report, such as 2026Q2
year- Calendar year of the quarter
qtr- Quarter number 1 to 4
periodStart- First day of the quarter
periodEnd- Last day of the quarter
status- P: counted from the primary source
text- One plain sentence with the quarter's counts
value- Same as reports
unit- What value counts
reports- Adverse event reports (MDRs) the FDA received in the quarter for the product code; counted once per report that lists the code
deaths- Reports whose event type is Death
injuries- Reports whose event type is Injury
malfunctions- Reports whose event type is Malfunction
other- reports less deaths, injuries and malfunctions: Other, No answer provided, and the few reports with no event type
effectiveFrom- Same as periodStart
effectiveTo- Same as periodEnd
lastChecked- Date the openFDA API was last read
id- Stable record id: dataset, place and period
How do I use it in my own tools?
deviceevents_lookup- For one calendar quarter (as 2026Q2): the national total and each of the top 300 device product codes, with adverse event reports, deaths, injuries, malfunctions and other. Counts only, from the openFDA device event API.
deviceevents_compare- The device event records of up to 10 quarters (as 2025Q4, 2026Q1, 2026Q2) side by side: the national and product-code counts so quarters can be compared.
deviceevents_search- Device adverse event records whose text matches the words given, such as a device name ("insulin pump"), a three-letter product code ("DZE") or a quarter ("2026 Q2").
How to cite this dataset
You can reuse, adapt and republish this data, including commercially, under the Creative Commons Attribution 4.0 license (CC BY 4.0). The only condition is that you credit Probative and link to the source.
- Version
- 1.0.0, released October 29, 2026
- Every version
- 10.5281/zenodo.23752987 (always points to the newest)
- License
- CC BY 4.0
Download the data
- CSV: device-events.csv 1261 KB · one row per record (columns in README.txt)
- JSON: device-events.json 2997 KB · every record with its sources
Free, no email or login. The files are hosted on this site and Zenodo, a public research archive.
Suggested citation
Probative. (2026). Medical device adverse event reports by product code and quarter (FDA MAUDE via openFDA) (Version 1.0.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.23752988
Credit line for a chart, table or article
Source: Probative, Medical device adverse event reports by product code and quarter (FDA MAUDE via openFDA), version 1.0.0, https://doi.org/10.5281/zenodo.23752988, licensed CC BY 4.0.
Add to a reference manager (Zotero, Mendeley, EndNote)
Not legal advice. A report to the FDA is an allegation that a device may have caused or contributed to an event, not a finding that it did, and a count of reports is not an incidence rate or proof of a defect. Questions about an injury from a medical device are for a licensed product-liability attorney and the person's doctor.
What changed recently?
- 2026-10-29 · FDA MAUDE adverse event reports by device product code and quarter, 2025Q1 to 2026Q3: 5,812,617 reports, with deaths, injuries and malfunctions for the 300 product codes with the most reports. (source)
The full log is in changes.json.
What are the limits?
- A report is an allegation or a manufacturer's own notice, not a finding that a device caused harm. Reports are not incidence: they depend on how many devices are in use, how often each is reported and who is reporting.
- Counts are driven by litigation, recalls and press as well as by harm. A single product code can jump by orders of magnitude in a quarter (for example when lawyers or manufacturers file batches), and the same event can be reported more than once by different sources.
- Reports arrive up to weeks after an event, so the newest quarter is flagged preliminary and will grow; older quarters can still be revised by the FDA.
- A report that lists several devices is counted once under each product code it lists, so product-code counts can add up to more than the national total.
- Only the product codes with the most reports over the quarters are included, plus the national total. Other codes exist.
- Narrative text, patient details and manufacturer names are never read or stored; only counts by product code, event type and quarter.
- Records older than 130 days since their last check are marked
stalein the files until we re-check them. Every confirmed error we fix is listed in the data corrections log.
Found a wrong value? Report an error with the record id and the source that shows the right value.
Running a firm? You don’t need the raw data: we handle your state’s ad rules for you. See where your firm stands first.
Free
Get your free review.
Send us your firm’s website. Probative Growth records a 90-second video of what clients see when they look you up, and emails it with a one-page summary within 2 business days. No logins, no call needed.
Not sure where to start? Answer 7 questions Ready to talk? Talk to us Or see what AI says about your firm
Your free review: It covers your Google listing, website, phone and ads, from public information.