FDA: Class I, II, III
The FDA classifies every recall it documents into one of three classes. The classification is determined by the FDA after the manufacturer reports the issue, based on the severity of the potential health hazard. Importantly, the classification doesn’t change retroactively — once a recall is Class II, it stays Class II even if later reports show worse outcomes.
- Class I: reasonable probability that using the product will cause serious adverse health consequences or death. Examples: Listeria contamination of ready-to-eat refrigerated foods, undeclared peanut in a baked good, the wrong drug in the right bottle. These are the recalls you should treat as urgent.
- Class II: may cause temporary or medically reversible adverse health consequences, where the probability of serious harm is remote. Examples: out-of-specification drug potency that wouldn’t cause permanent injury, allergen mislabeling on a less-common allergen.
- Class III: not likely to cause adverse health consequences. The product violates an FDA rule (labeling, UPC, documentation, software calibration) but the issue is administrative rather than clinical.
The FDA publishes Class I/II/III classifications on the openFDA enforcement endpoint. We carry the classification through verbatim on every FDA recall page in our database.
USDA FSIS: Class I, II, III
The USDA’s Food Safety and Inspection Service uses the same three-tier system, with parallel definitions but a narrower scope — FSIS regulates meat, poultry, and processed egg products only. If an FSIS recall is Class I, the hazard profile is the same as an FDA Class I: reasonable probability of serious harm. Pathogens like Listeria, Salmonella, and E. coli O157:H7 in ready-to-eat protein products almost always trigger Class I FSIS recalls.
FSIS recall notices also include the establishment number (the unique ID of the production facility, visible on the inspection legend on packaging) and the production date range. We carry both through into the recall detail page so you can match a recall against your own freezer.
CPSC: voluntary joint announcements
The Consumer Product Safety Commission doesn’t use the Class I/II/III system at all. Instead, CPSC recalls are issued as joint press releases between the agency and the manufacturer, with the hazard described in plain English (fire risk, choking hazard, entrapment risk, lead-paint violation) and a remedy spelled out (free replacement, refund, repair kit).
Almost every CPSC recall is technically voluntary — the manufacturer initiates the recall and CPSC reviews and publishes it. CPSC does have mandatory-recall authority under Section 15 of the Consumer Product Safety Act, but it’s used very rarely because manufacturers generally cooperate once the agency identifies a hazard.
Our severity score for CPSC recalls is computed from the hazard type and the presence of reported injuries — see the methodology page for the rubric.
NHTSA: campaign numbers, not classes
NHTSA vehicle, tire, child-seat, and equipment recalls each get a campaign number — a unique identifier like 25V123000 (year, product code, sequence). The product codes are V for vehicle, T for tires, C for child seats, E for equipment. NHTSA doesn’t classify recalls into severity tiers; instead the recall notice describes the defect and the remedy directly.
NHTSA recalls are unique in two important ways. First, every NHTSA recall is mandatory for the manufacturer to remedy at no charge to the owner — there’s no expiration on the obligation. If you bought a 20-year-old vehicle with an open recall, the manufacturer still has to fix it. Second, the recall description names the Federal Motor Vehicle Safety Standard (FMVSS) that the defect violates (FMVSS 208 for airbags, FMVSS 105 for brakes, FMVSS 138 for tire pressure monitoring, and so on).
We compute severity for NHTSA recalls from the consequence of the defect (does the airbag fail to deploy? does the steering lose assist?) rather than from a Class designation, because NHTSA doesn’t publish one.
How our 1–5 severity score maps to all four systems
Because every agency uses a different classification, we compute a single normalised severity score 1–5 across every recall in our database. The rubric is fully published on the methodology page. In short:
- FDA/USDA Class I + reported deaths → Severity 5 (Critical)
- FDA Class I or USDA Class I → at minimum Severity 4 (Severe)
- FDA Class II + hospitalization reports → Severity 4
- FDA Class II without reported illness → Severity 3 (High)
- FDA/USDA Class III, voluntary low-risk → Severity 1–2
- NHTSA campaigns with reported deaths → Severity 5
- NHTSA campaigns affecting structural safety (airbag, brake, steering, fuel) → Severity 4