About the data
Data sources, methods and limitations
Drugs and approvals (Drugs@FDA)
The drug catalog is every NDA (new drug application) and BLA (biologics license application) in FDA's Drugs@FDA database, downloaded from openFDA and rebuilt regularly: 5,899 applications as of Nov 18, 2026. Generic drugs (ANDAs) are left out; their reference drug is listed. Each application is one row, named after its first product's brand name.
The approval tracker lists original approvals and efficacy supplements (which add a new indication, population or dosing) since 2000. Labeling, manufacturing and other supplements are left out. Drugs@FDA covers CDER; most vaccines, blood products and cell and gene therapies are approved by CBER and are missing.
Novel means CDER's novel drug approvals list (read from fda.gov for 2021 on) or, for earlier years, a Type 1 (new molecular entity) NDA. Novel biologics before 2021 are not all marked.
Drugs in development (pipeline)
The pipeline lists investigational drugs and biologics from ClinicalTrials.gov: every drug or biologic tested in an industry-sponsored, interventional Phase 2 or Phase 3 trial that is recruiting, active or enrolling, or that completed with a primary completion date in the last two years. Sponsors list the same drug under code names and generic names, so names given together for one intervention are merged into one entry; two different generic names, or two approved drugs, are never merged. Entries whose names match an approved ingredient in Drugs@FDA or FDA's NDC directory are left out, since those are new uses of approved drugs (their trials are on each drug's page). Placebos, standard care and chemotherapy backbones are left out, and a name seen in only one trial is kept only if it is a code or generic name.
Phase is the highest phase among those trials, not a regulatory status, and the registry can lag a company's own announcements. Merging is automatic, so an alias can occasionally belong to a related compound, and a drug approved outside the US can appear. Drugs in development only in Phase 1, or only in academic trials, are not listed.
Drugs by condition
A drug is listed for a condition when the Indications and Usage section of its current FDA label (from openFDA) names one of the phrases shown on the condition page, or the words typed in the search. Only brand-name prescription drugs and biologics (NDAs and BLAs) are listed; generic labels are counted. A label can name a condition it treats only in part (a subtype, a stage, or as add-on therapy), and a word match can occasionally pick up a label that mentions the condition in another way, so read the label before relying on it.
Expedited programs
Breakthrough therapy and accelerated approval come from CDER's published lists of approvals (PDFs on fda.gov). Only the application and submission numbers are read from them and matched to Drugs@FDA. Priority review and orphan come from the Drugs@FDA submission record. Fast track has no public list and is not shown. Designations that have not yet led to an approval are not public and are not shown.
PDUFA calendar (AI-compiled)
FDA does not publish target action dates. Once a week, Claude (Anthropic's AI) searches company press releases, SEC filings, FDA's advisory committee calendar and trade press and compiles the dates it finds, linking the source of each. It can miss dates, especially for private and small companies, and can misread a source. Always check the linked page.
Adverse events (FAERS)
Reports come from the FDA Adverse Event Reporting System through openFDA. On drug pages, a report counts when it names that application. On the FAERS Monitor, a free-text search matches brand, generic and reported product names. Any drug on a report counts, suspect or not, because openFDA cannot tie a condition to one drug on a report.
Reports are voluntary (mandatory for manufacturers), unverified and often duplicated. Counts are not rates and cannot show that a drug caused an event. The proportional reporting ratio (PRR) compares how often a reaction appears in one drug's reports with all other reports; it is flagged as a signal at PRR of 2 or more, chi-squared of 4 or more and 3 or more reports (Evans et al., 2001). A signal is a reason to look closer, nothing more.
Medicare spending
From CMS's Medicare Part D Spending by Drug and Part B Spending by Drug datasets, calendar years 2020 to 2024. Spending is gross: it does not subtract manufacturer rebates, which are large for many brand drugs, so it overstates net cost. CMS suppresses small counts. Drugs are matched to the catalog by brand name, so some drugs CMS lists under a different name are not linked.
Clinical trials
Live from the ClinicalTrials.gov API, limited to studies with a drug or biologic intervention. On drug pages, trials are found by the first active ingredient named as an intervention, so trials using a code name may be missed. Company pipelines use the lead sponsor names listed for that company. Pivotal-looking is a heuristic: Phase 3, "pivotal" in the title, or randomized with 100 or more participants.
Therapeutic areas
Drugs are placed in areas by keyword matches on their FDA pharmacologic class, CDER's approved-use text and ingredient; trials by their listed conditions. A drug can be in several areas or none.
AI-written text
Drug briefs, the PDUFA calendar and the news digest are written by Claude from the data on the page plus, for briefs and the calendar, web searches. Each is written once and stored: briefs are rewritten after 30 days, the calendar weekly and the digest daily. AI text can be wrong; every brief links its sources.
For medical devices, see our sister site MedTech Resources.
For information only. Nothing on this page is investment, legal, tax or medical advice, and nothing here is an offer to sell or a solicitation of an offer to buy any security. The data comes from public government sources, which can be incomplete, delayed or wrong. Check the original record before relying on it.