Open Data
The datasets behind Trial Friend's data studies, free to download and reuse. Journalists, researchers, advocacy groups, and students are welcome to build on these numbers. All we ask is credit with a link back to this page.
CC BY 4.0Free for any use, including commercial, with a link backActive clinical trial counts for each of the 211 rare diseases Trial Friend tracks, split by recruitment status. Built from ClinicalTrials.gov v2 with the same condition matching our live disease pages use, deduplicated by NCT identifier within each disease and cleaned with our per-disease exclusion lists.
Where recruiting rare disease trials actually run. One file gives each tracked disease's recruiting trial count and how many US states hold at least one of its sites; the other gives every state's trial site listings, recruiting trials, and how many tracked diseases are reachable in-state. Sites are counted once per unique city and state pair per trial.
Which open rare disease treatment trials exclude people over a medication they commonly take. Every recruiting or not-yet-recruiting interventional trial for the 37 conditions with a Trial Friend medication checker was run through that checker. One file gives each condition's share of trials with a medication exclusion, split into steroids, immune-targeting drugs, gene, RNA and cell therapies, cancer drugs and other medicines; the other gives every match with the trial's own exclusion wording and any time window it names. Only matches on a named drug or its drug class count. A hand check of 100 matches found 63 clear exclusions, 23 that apply to some users (a dose limit, a recent start or a stable-dose rule) and 14 errors.
Both datasets are built from the ClinicalTrials.gov v2 API using the same disease name and alias matching that powers Trial Friend's live disease pages, then cleaned with our hand-maintained exclusion lists for known condition mismatches. Trials serving several tracked diseases are deduplicated by NCT identifier in study-level totals. Site counts merge duplicate listings of the same hospital (one row per unique city and state pair per trial), and a facility-level robustness check moved the headline single-site share by only 2 percentage points.
Known limitations: registry site lists are sponsor-reported and can lag reality in both directions, condition matching against a registry is imperfect, and for the very broadest disease categories a single query can hit the registry's 1,000-result page cap, slightly undercounting them. The census and travel datasets were generated on different dates with different status filters, so their trial totals are not directly comparable to each other. Exact figures are close estimates, not gospel.
Data: <a href="https://www.trialfriend.com/data">Trial Friend Open Data</a>, CC BY 4.0 Questions, corrections, or a dataset you wish existed? Email jason@trialfriend.com.