Coverage & data integrity

Exactly what pdufa.bio tracks, how complete and how fresh it is, and — stated plainly — what it does not cover. Regenerated automatically on every data refresh.

1151

catalysts tracked

PDUFA / FDA decisions

1000

trial readouts

device milestones

469

companies (tickers)

98%

rows with a primary-source link

46%

PDUFA recall vs reference set

How we source it

ClinicalTrials.gov	958
SEC EDGAR (8-K/6-K/10-Q)	85
Earnings (FMP)	72
Company press (FMP)	15
Curated device seed	12
Curated foreign PDUFA	4
curated_pharma;colist	3
sec_edgar;colist	2

Date precision (we never fake a hard date)

Every date is tagged by how precisely it is known. Month/quarter/half items are estimates that shift and are shown as “(est.)”.

month	958
day	135
half	31
quarter	22
year	5

What we do NOT cover (known limitations)

• No per-drug approval probabilities. We refuse to invent one — here's why.
• Foreign / mega-cap PDUFAs that don't file clean US press releases are covered only via a small curated seed, not exhaustively.
• Pre-IND and early-pipeline assets without a registered trial or filing are out of scope.
• Quarter / half-year readout estimates are ClinicalTrials.gov primary-completion windows and move; treat them as ranges.
• Options / dark-pool / IV fields are best-effort vendor snapshots (ORATS ~15-min delayed), price-only, and labeled as such.
• We drop our own junk. This refresh filtered 46 stale-alias rows and flagged 268 rows with no drug name rather than publish them.

Update cadence

The dataset refreshes on a schedule (~5×/day via cron) plus a full primary-source crawl. It is a snapshot, not a live tick — a shown value can be a few hours old, and vendor options data carries a ~15-minute delay. See /sources.

Want the underlying data? The research datasets publish n-disclosed cohorts you can cite. A read-only feed / API is on the roadmap.

Generated 2026-06-24 from the live catalyst set.