Device enforcement records by recall class

FDA report dates in calendar 2025

Class II2,433
Class I380
Class III95
2,908 report records mapped to 972 event IDs; rows are not unique recall events.

The FDA's open device-enforcement data contains 2,908 records with report dates in 2025. Those records map to 972 event IDs in MedDataUSA's extract.

The difference is not a mistake. A recall event can involve multiple products, models, lots, or report records. Counting rows as “recalls” can inflate the apparent number of distinct events by nearly threefold in this period.

Quick read

  • Class II accounted for 2,433 records, Class I for 380, and Class III for 95.
  • 2,869 records were labeled firm-initiated; 38 were FDA-mandated and one lacked that field.
  • Enforcement reports are not a live alert feed or a safe denominator for comparing company quality without product and event deduplication.

What the classes mean

FDA uses Class I when there is a reasonable probability that use of or exposure to a violative product will cause serious adverse health consequences or death. Class II applies when consequences may be temporary or medically reversible, or when the probability of serious consequences is remote. Class III applies when adverse health consequences are not likely.

Those classifications describe health-hazard assessment, not the number of affected units or the operational burden of a correction. A Class II record can still involve a widely used product and substantial remediation.

Records, products, and events

The openFDA enforcement endpoint is organized around report records. Records carry a recall number and can share an event ID. The 972 unique event IDs in this snapshot offer a closer event-level denominator, but even that is not a universal count of every real-world corrective action.

Events can evolve, product descriptions differ in granularity, and report dates are not necessarily recall initiation dates. An analysis must choose which date and identifier answer its question.

Most records were firm-initiated

The “voluntary/mandated” field labeled 2,869 records as firm-initiated. That aligns with FDA's explanation that recalls are usually initiated by firms and overseen by the agency. It does not mean FDA played no role, nor does it measure how quickly a firm detected or corrected a problem.

Likewise, 38 FDA-mandated report records do not necessarily represent 38 distinct mandated events; the same deduplication issue applies.

What not to infer

Raw record totals should not become a company league table. Companies differ in product mix, installed base, reporting structure, and how a single event is divided into records. A higher count is not, without further denominators and context, proof of worse product quality.

The endpoint also carries an FDA warning: openFDA should not be used to issue public alerts or track the full recall lifecycle. Current safety decisions should use FDA recall communications and product-specific notices.

Methodology and limitations

MedDataUSA retrieved every openFDA device enforcement record returned for report_date from January 1 through December 31, 2025. We paged in 1,000-record blocks, retained recall number, event ID, classification, report date, initiation type, and status, and counted report records and distinct event IDs.

This is a report-date snapshot. It does not count affected units, weight hazards, determine current recall status, identify unique companies, or replace FDA safety communications. openFDA data can be corrected and its endpoint update date was September 30, 2026 at review.

Sources