Why MAUDE report counts are not device-risk rates

A device adverse-event chart can prompt a useful question without answering how often a problem occurs. For ENT readers evaluating a device claim, the distinction matters before a report total becomes a risk statement. FDA uses reports in its Manufacturer and User Facility Device Experience database, or MAUDE, to help identify potential safety issues. The agency also warns against using those reports alone to calculate event rates or compare adverse-event rates between devices.

MAUDE receives mandatory reports from manufacturers, importers and device user facilities, alongside voluntary reports from health professionals, patients and consumers. These are reports of suspected device-associated deaths, serious injuries or malfunctions. They form part of postmarket surveillance; they are not a count of every device use followed systematically for the same outcome.

Start by checking the unit behind the number. FDA’s Total Product Life Cycle documentation distinguishes device-problem counts from report counts and event counts. One report can contain several problem codes or describe more than one event. A chart of problem codes therefore cannot automatically be read as a chart of affected people, separate incidents or individual reports. Ask which field was counted before interpreting the size of a total.

Then check how the source handled later information. MAUDE report records can receive supplemental information, and FDA’s Total Product Life Cycle documentation notes preliminary, supplemental and duplicate reports in its source data. A later database observation is not automatically a new incident. The reporting method should state its observation date, counting unit and treatment of those records.

The missing denominator is a separate obstacle. FDA identifies underreporting and the lack of information about how often devices are used among MAUDE’s limitations. A count of submitted reports does not supply the number of exposed people or uses needed to estimate incidence. Expressing one category as a percentage of retrieved reports only describes that retrieved set. It does not turn the percentage into the chance that a device user will experience the event.

Search coverage also needs scrutiny. FDA’s openFDA MAUDE description warns that product, trade and company names can vary; a search returns records containing the supplied terms. A neatly labeled total may therefore reflect the chosen search language. It should not be presented as a complete account of everything associated with a product merely because a query returned a number.

For a claimed safety comparison, ask for the supporting design before accepting the ranking: which populations and uses were counted, over what period, with which outcome definitions and how events were identified. These are editorial questions about whether the comparison has a defensible basis. Matching two search windows or dividing two report totals does not, by itself, resolve the limitations FDA describes.

Causation requires another distinction. FDA says a submitted report does not establish that the device caused or contributed to the reported event. That limit does not make reports useless or show that a concern is unfounded. It means the report is a starting point for investigation, while a conclusion about cause or comparative safety needs evidence beyond the tally.

The practical reading order is to identify the counted unit, examine coverage and record handling, find the exposure basis for any rate, and separate a suspected association from a causal conclusion. If those details are absent, describe the limitation alongside the number. Do not use a larger total to label a device more dangerous or a smaller total to label it safer.

Source note: This independent reference explainer uses FDA documentation observed October 15, 2026 UTC, still October 14 in U.S. Eastern time. The selected pages’ original publication and whole-page update dates were not verified. This is not a new FDA action, finding or safety alert. No individual adverse-event report, patient narrative, case export or device-specific comparison was used.