Skip to content

How We Screen for FDA Enforcement

Method and limits, published in full

Most recent completed screening pass: July 9, 2026

Every company in the Inventory Ready vendor directory must have no unresolved FDA enforcement finding as of our most recent completed screening pass. That criterion appears on our method page and in our inclusion criteria. This page documents how the screening actually works: which FDA datasets we check, how matching is done, what happens when a record surfaces, and, just as important, what the criterion does and does not mean.

A screening method is only as trustworthy as its limits are honest. Ours has real limits: matching is name-based, FDA's datasets lag events by weeks, and some FDA record types are not yet part of the systematic pass. All of them are documented below. We publish the method itself rather than per-company findings; the reasons for that are also explained below.

What we screen for

“No unresolved FDA enforcement” is one of the universal criteria every company must meet to be listed in our directory, alongside no known fraud indicators, public documentation, and verifiable capabilities. It is deliberately scoped: unresolved enforcement. A company that experienced an FDA action years ago and resolved it, with the resolution documented, can still meet the criterion. The criterion asks whether anything surfaced by our screening remains open and unaddressed as of the pass date.

The criterion is also date-bound. It speaks as of the most recent completed screening pass, which is stamped on this page and on the method page. It is not a continuous monitor, and we do not claim one.

The datasets we screen

The systematic pass runs against FDA's published recall and enforcement datasets, accessed through the openFDA program:

  • Food enforcement reports (the openFDA food/enforcement dataset), which is where dietary supplement recalls are classified.
  • Drug enforcement reports (the openFDA drug/enforcement dataset), which catches products regulated on the drug side and supplement-adjacent recalls filed there.

Both datasets are published from FDA's Recall Enterprise System, cover recall events from 2004 to the present, and are refreshed by FDA on a weekly export cycle. We consume the full bulk export rather than sampling, so a pass evaluates the complete published dataset as of the export date. The export date used by the most recent pass is recorded in the method notes at the bottom of this page.

What these datasets do not include.Recall data is one slice of FDA enforcement. Warning letters, consent decrees, seizures, injunctions, facility inspection classifications, and import refusals live in separate FDA systems, several of them in FDA's Data Dashboard, which publishes final actions only, on its own weekly cycle, organized by fiscal year with historical coverage that FDA does not fully document. Those record types are not part of the systematic pass today. They are reviewed per vendor as part of Trust Tier 4 verification, and extending the systematic pass to them is a named, planned improvement rather than a quiet assumption.

Which companies the pass covers. The systematic pass screens the companies in our vendor directory. Companies that appear on our branded ingredient reference pages as ingredient owners, and providers in our packaging materials browser, are covered by per-listing editorial review today; bringing them into the same systematic pass is a planned extension. We state this scope precisely because a screening claim that is broader than the screening actually performed would be worse than no claim at all.

How matching works, and where it can fail

FDA's enforcement records identify the recalling firm by name and address. Our pass matches directory companies against the recalling-firm field using word-boundary name matching over a set of name variations for each company (legal suffixes stripped, common abbreviations expanded). This is the same class of matching FDA's own public search tools use, and it has two well-understood failure modes:

  • False positives from name collisions. A different company with a similar name can match. For this reason, no name match is treated as pertaining to a listed company until identity is corroborated with independent identity fields such as address, state, or FDA facility identifier. An adverse conclusion is never drawn from an uncorroborated name match: a similar name is not the same company.
  • False negatives from name differences.A record filed under a different legal name, a parent or subsidiary entity, or a doing-business-as name can be missed. Matching on FDA's facility identifier numbers, which are stable across name changes, is a planned improvement.

FDA's datasets also lag the events they describe by weeks, because records enter the export after classification. Screening dates therefore reflect FDA's data export, not real time.

What happens when a record surfaces

Every record surfaced by a pass receives a documented internal disposition before the pass counts as complete. In plain terms, a surfaced record ends up in one of four places:

  • Not the same company. Identity review shows the record belongs to a different firm. The match is recorded as a false positive and the listing is unaffected.
  • Identity unconfirmed. The match cannot yet be corroborated or ruled out. The record stays open with a due date for follow-up, and no adverse action is taken while identity is unconfirmed. A pass with open identity questions is not a completed pass.
  • Resolved, with evidence.The record pertains to the listed company but is documented as resolved. We apply record-appropriate evidence standards: a recall requires a documented terminated status in FDA's data; a warning letter requires FDA's close-out; an inspection finding requires reclassification. “The company says it was handled” is not a resolution standard.
  • Unresolved. The record pertains to the listed company and is not documented as resolved. Listings do not remain active with unresolved findings.

Dispositions are recorded in an append-only internal log: corrections are written as new entries that reference the old ones, never edited in place. Companies can raise questions or corrections about their own listing at any time through claim your listing.

What “no unresolved enforcement” means, and what it does not

It means: in our most recent completed screening pass, run 2026-07-09 against the datasets named above with the matching method described above, every surfaced record received a documented disposition, and no active listing carried an unresolved finding.

It does not mean:

  • FDA approval or certification. The FDA does not certify, approve, or endorse manufacturing facilities, and nothing on this site should be read as implying it does.
  • A clean history forever. The criterion concerns unresolved findings. A resolved historical action, documented as resolved, does not by itself disqualify a company.
  • Coverage of every enforcement surface. The systematic pass covers the recall and enforcement datasets named above. It does not yet systematically cover warning-letter text, inspection classifications, or import refusals, and it does not cover FTC advertising enforcement at all.
  • A real-time guarantee.The criterion speaks as of the pass date, and FDA's own data lags events by weeks. Something could happen tomorrow that no dated screen can capture.
  • Anything about companies not listed. Ours is a curated directory. Companies are absent, or become inactive, for many unrelated reasons: certification confirmations lapse, information goes stale, businesses change focus, or our review is simply not finished. An absent or inactive listing is not evidence of an enforcement finding, and no one should infer one from it.

Why we publish the method, not findings

FDA enforcement records are public government records. FDA publishes them, maintains them, and corrects them; anyone can search them directly using the links below. We see no public value in republishing per-company enforcement narratives, and we see real potential for harm: name-based matching has known false-positive modes, datasets lag, and a mistaken association between a company and an enforcement record is difficult to fully retract once published. Our screening therefore feeds listing decisions, not public storytelling. The criteria are public; the per-company work stays internal; and the protective action when something is unresolved is a listing decision, not a publication.

Check the sources yourself

Every dataset we screen, and several we do not yet screen systematically, is publicly searchable. A brand evaluating a manufacturing partner can and should look directly:

When searching, keep the same limits in mind that we document for our own pass: search name variations, corroborate identity before drawing conclusions, and remember that absence of a record in any one database is not proof of a clean history across all of them.

Method notes

Most recent completed pass: 2026-07-09.

Sources scanned in that pass: openfda-food-enforcement (bulk export 2026-07-07); openfda-drug-enforcement (bulk export 2026-07-07). Coverage statements on this page enumerate exactly what the pass scanned, nothing broader.

Match method (as recorded): name-search: word-boundary vendor-name-variation match over recalling_firm (bulk export).

Entity scope: companies in the vendor directory. Branded-ingredient owner companies and the packaging materials browser are covered by per-listing editorial review; bringing them into the systematic pass is a named, planned extension.

Known coverage gaps: warning-letter full text, consent decrees, seizures, injunctions, inspection classifications, import refusals, and FDA Form 483 observations are not part of the systematic pass; FTC advertising enforcement is out of scope entirely. FDA data lags events by weeks.

Cadence: screening passes are dated and published as they complete. This page reflects the most recent completed pass and makes no forward-looking refresh promise.

Related reading

How we assess companies overall, including the criteria beyond enforcement screening, is documented in our method, and current verification coverage is published in how we assess. For ingredient-level sourcing questions, the branded ingredient reference covers ownership, licensing, and dated availability, and the certifications index explains what each credential covers and how to verify it.

Citing this page

Journalists and researchers are welcome to cite this method with attribution to Inventory Ready and a link to this page. For questions about how the screening works, contact media@inventoryready.com or see the press page.