About, Data & How to Cite

Last updated: June 2026

How to cite trialcat

Using trialcat in a paper, poster, grant, or report? Please cite it — and cite the underlying ClinicalTrials.gov data too. Copy-paste blocks below. Replace the access date as needed.

APA (7th ed.)

The Real Cat AI Labs. (2026). trialcat: Clinical trial enrollment intelligence [Web application]. The Real Cat AI Labs. https://trialcat.ai

IEEE

The Real Cat AI Labs, "trialcat: Clinical trial enrollment intelligence." The Real Cat AI Labs, 2026. [Online]. Available: https://trialcat.ai

MLA (9th ed.)

The Real Cat AI Labs. "trialcat: Clinical Trial Enrollment Intelligence." The Real Cat AI Labs, 2026, trialcat.ai. Accessed 16 June 2026.

Chicago (author-date)

The Real Cat AI Labs. 2026. "trialcat: Clinical Trial Enrollment Intelligence." https://trialcat.ai.

BibTeX

@misc{trialcat2026,
  title        = {trialcat: Clinical Trial Enrollment Intelligence},
  author       = {{The Real Cat AI Labs}},
  year         = {2026},
  howpublished = {\url{https://trialcat.ai}},
  note         = {Data from ClinicalTrials.gov (U.S. National Library of Medicine)}
}

Citing the underlying data (ClinicalTrials.gov)

U.S. National Library of Medicine. ClinicalTrials.gov. National Institutes of Health. https://clinicaltrials.gov (accessed 16 June 2026).

How trialcat works

trialcat turns the public ClinicalTrials.gov registry into an interactive map and a set of enrollment-pattern statistics. You filter by geography, therapeutic area, phase, status, intervention type (drug vs. device vs. biologic, and beyond), and — for devices and drugs — a product-type drill-down. The map recolors live; click any country or U.S. state for a stats popup you can export to CSV.

Where the data comes from (the pipeline)

1. Source. Every record originates from ClinicalTrials.gov API v2, the U.S. National Library of Medicine / NIH registry of FDA-regulated studies. The data is public domain.
2. Fetch & parse. A backend ETL pulls studies, paginating through the API, and parses each into a normalized schema (trial, locations, interventions, conditions, outcomes). Partial dates and missing fields are handled defensively.
3. Derive. Therapeutic area is mapped from the trial's MeSH terms; an approximate per-month enrollment rate is computed from start and primary-completion dates; intervention type comes straight from the registry.
4. Enrich (FDA). Device interventions are best-effort matched to the openFDA device-classification database for an FDA class (I/II/III) and review-panel category; drug interventions are matched to openFDA NDC / Drugs@FDA for a pharmacologic class and Rx/OTC status. These are ballpark, name-based mappings, not verified per-product determinations.
5. Refresh. A scheduled job pulls trials updated in the recent window so the database stays reasonably current. trialcat does not guarantee real-time accuracy; sponsors update their own records on varying schedules.

What to trust, and what not to

• Enrollment rates are approximate (start → primary-completion proxy; overestimates duration for long endpoints).
• Therapeutic area is heuristic (MeSH-tree mapping; may differ from the sponsor's framing).
• FDA class / product category and drug class are advisory (heuristic openFDA name-matching; verify against the FDA 510(k)/PMA, Orange Book, or Drugs@FDA databases for anything that matters).
• Coverage is a subset, not the whole registry — and it grows over time. Numbers are landscape-level, not audit-grade.

In short: trialcat is for research, education, and landscape orientation. For decisions, go to the primary sources. Full legal terms, including the Massachusetts governing-law and commercial-use disclaimer, are on the Terms & Disclaimer page.

trialcat is a project of The Real Cat AI Labs, Inc., a Massachusetts 501(c)(3) nonprofit, with support from Biotech Mentor LLC. Open-source (MIT). Data is public domain.