Methodology & coverage
How this data is sourced, how current it is, how documents are scored, and what the AI does and does not do.
Catalyst Wire Dossier organizes two decades of public regulatory and clinical-trial documents by disease and drug, condensing each into a plain-language summary with an importance score. This page is the receipt: every number below is computed live from the database.
Where the data comes from
Every document is ingested from an official public source and links back to the original record — we add structure and summaries, never invent facts.
- ClinicalTrials.gov — 582,783 documents
- FDA (openFDA) — 73,073 documents
- EMA (European Medicines Agency) — 6,106 documents
Coverage & freshness
- 500,233 documents analyzed and searchable (of 661,962 ingested)
- 33,066 diseases · 87,177 drugs indexed
- Roughly 20 years of history, refreshed by a daily ingest
- Most recent document analyzed: 2026-08-29
How importance scoring works
Each analyzed document is assigned a single category and an importance score from 0–100 by the language model, so a routine administrative filing and a first-in-class approval don’t look the same. Scores rank what surfaces first on disease and drug pages and what triggers alerts; they are a heuristic, not a clinical or investment judgment.
Categories:
- Approvals
- Safety alerts
- Label changes
- Trial results
- Research
- Guidance
- Withdrawals
The AI analysis — and its limits
- What it does: categorize each document, score importance, and write a short summary. Analysis runs on gpt-5-nano.
- Source-linked: every document keeps a link to the original FDA / ClinicalTrials.gov / EMA record so you can verify any claim.
- It can be wrong: summaries and scores are automated and may misread or omit detail. Treat them as a fast index into the primary sources, not a substitute for reading them.
- Not medical or investment advice. Dossier is an information tool; decisions should rest on the original records and professional judgment.
Why this beats checking the sources by hand
FDA, ClinicalTrials.gov, and EMA each have their own search, formats, and update cadence. Dossier pulls all three into one disease- and drug-keyed view, scores and summarizes every item, and can email you when something new lands — so you stop stitching together five tabs to answer “what changed here?”
Questions about sources or scoring: support@catalystwire.org