Design: Cross-sectional study of a national database
Population: 6,308 summary reports covering 159,503 device adverse events, 2019–2025
Primary outcome: Accuracy, completeness and timeliness of summary reports
Effect: 57% no device problem identified; 55% past 30-day deadline; about 11,900 events undercounted

This BMJ cross-sectional study examined every summary adverse-event report filed in the FDA's MAUDE device database from July 2019 to June 2025. Summary reports allow manufacturers to bundle many events into one filing. There were 6,308 such reports covering 159,503 adverse events, and their use rose about 21% a year; 91% of the events involved the highest-risk class 3 devices.

The reports were often uninformative. In 57% the device problem field said an adverse event had occurred without an identified device or use problem, and 38% recorded insufficient information. Errors in the event-count fields led to an undercount of about 11,900 events. More than half (55%) missed the 30-day reporting deadline.

For clinicians, the point is that the absence of a safety signal in a device database is weak reassurance. Reporting by the people who see the harm still matters.

Report device malfunctions and device-related harm through your national system; in India that is the Materiovigilance Programme of India.
Do not take the absence of public device reports as evidence a device is problem-free.
Record device make, model and lot number in the notes when an implant or device is involved in an adverse event.
Keep a failed device where possible; investigation depends on it.

Why it matters

Device safety signals depend on reports that this study shows are often vague, delayed or undercounted.

The statistics, in plain English

These are descriptive proportions of all reports in the period, not estimates from a sample, so confidence intervals are not the issue; the limitation is that the study cannot say how many harms were never reported at all.