Summary
A cross-sectional study of 6,308 summary reports in the FDA’s MAUDE database found frequent reporting errors, missing device-problem detail and delays. The reports covered 159,503 adverse events from July 2019 to June 2025.
A cross-sectional study published in The BMJ found frequent errors and delays in manufacturers’ summary reports of medical-device adverse events submitted to the US Food and Drug Administration (FDA). The researchers reviewed 6,308 reports in the FDA’s MAUDE database, covering 159,503 events from July 2019 to June 2025. Most events were associated with high-risk devices, and many reports lacked detail needed to interpret possible device problems.
The findings concern how safety information is recorded and made available for surveillance. They do not show that a device caused any individual event: MAUDE is a reporting database, not a system for determining causation or calculating event rates.
What the researchers found
The reports included 139,333 serious injuries (87% of the total), 16,985 deaths (11%) and 3,185 malfunctions (2%). Ninety-one per cent of the reported events involved class 3 devices, the FDA’s highest-risk category; 99% were related to cardiac devices.
The study identified several problems in the reports’ structured fields. In 3,589 reports (57%), manufacturers selected “adverse event without identified device or use problem”; in 2,412 (38%), they selected “insufficient information.” The researchers said these reports did not add further details in their narratives to clarify potential device problems.
In 694 reports (11%), manufacturers marked “N” in the field indicating whether the submission was a summary report. In another 423 reports (7%), they entered “n=1” as the number of events, although the narratives described more. That discrepancy led to an undercount of 11,902 events in the structured data. Researchers and regulators relying on that field alone could therefore see a lower total than the report narratives indicated.
The researchers also assessed reporting deadlines. FDA rules require manufacturers to submit reports within 30 days of becoming aware of a death, serious injury or malfunction. Fifty-five per cent of the summary reports missed that deadline. Submission took a mean of 50 days and a median of 35 days.
Use of summary reporting increased during the study period: submissions rose from 808 in 2020 to 1,757 in 2024. This is a rise in reports, not a measure of how often device-related harm occurred among patients or device users.
Why the reporting format matters
MAUDE is the FDA’s public database for medical-device adverse-event reports and is used to help identify possible safety signals. Under the exemptions examined in the study, manufacturers could submit aggregated event information drawn from specified registries instead of separate reports for each event. Aggregation can reduce the visibility of individual cases and their clinical context, such as event chronology or procedural details.
The study’s results show how errors in fields and delayed submissions can complicate the use of these records. Summary reports covered substantial numbers of deaths and serious injuries involving high-risk devices, yet often provided limited information about possible device problems. That makes the records less useful for spotting patterns and assessing them alongside other evidence.
MAUDE is a passive reporting system: reports can be affected by underreporting and reporting bias, and the database lacks total device-use data needed to calculate rates. An event listed in MAUDE is not, by itself, evidence that a device caused it. The authors also could not determine from the public reports whether missing detail originated in the underlying registries, manufacturers’ summaries or the requirements of the FDA exemptions.