Data sources
Every claim in the register resolves to a source. Here is where they come from, under what licence, and — more usefully — what each one cannot establish.
ClinicalTrials.gov
Registered trials, pulled nightly via the v2 API. Maintained by the U.S. National Library of Medicine (NLM); public domain.
What it cannot establish: a registered trial proves a sponsor declared a study — not that it is running, nor that it will read out. An approval leaves no trace there at all: a programme that reaches market simply stops registering trials.
This site is not affiliated with, nor endorsed by, the NLM.
Base de données publique des médicaments (BDPM)
French marketing authorisations: CIS code, authorisation date, marketing status and dispensing conditions. Produced by ANSM, HAS and UNCAM and published at base-donnees-publique.medicaments.gouv.fr under the Licence Ouverte (Etalab). Files downloaded 19 August 2026.
What it cannot establish: BDPM carries no indication field. Finasteride is authorised there at 1 mg and at 5 mg with nothing in the files separating hair loss from prostate enlargement. An authorisation date is the earliest for the substance, whatever it was first authorised for.
Regulators, literature and company disclosures
FDA, EMA, ANSM, PMDA and MFDS for authorisation decisions; PubMed for peer-reviewed results; company releases where nothing better exists — and the row says so when that is the case.
Hierarchy:the regulator's own document beats the press release reporting it, and a peer-reviewed readout beats the sponsor's announcement of it. The confidence level shown on each record reflects that order.
Using this
Source data remains governed by its own licences. If you are citing the register itself, a link to the record is enough — and we would rather fix an error than defend one: report a correction.