How this works
One NCT ID, two kinds of answer
Paste a trial's NCT ID and we fetch its full record from ClinicalTrials.gov and resolve any linked publications on PubMed — nothing more than these two public sources is ever used. What you get back depends on the trial's current status. A trial that was terminated, withdrawn, or suspended gets a reasoning pass: ranked hypotheses for why it likely stopped. Every other trial — completed, recruiting, or still planned — has no failure to explain, so it gets a plain factual snapshot instead, with no AI reasoning involved at all.
If the trial failed: ranked hypotheses
A structured reasoning pass classifies the likely cause against a fixed taxonomy, and ranks each hypothesis by confidence:
- Recruitment / enrollment shortfall
- Efficacy (missed primary endpoint)
- Safety / toxicity
- Funding / business decision
- Operational / protocol issues
- Strategic / competitive deprioritization
- Futility (interim analysis)
Every claim is labeled by how certain it is. Fact means it's directly in the trial record. Inference means it's a reasonable read of those facts. Hypothesismeans it's plausible but unproven. Every ranked hypothesis also comes with its strongest counter-argument, so nothing is presented as more certain than it actually is.
If it didn't: a plain factual snapshot
Completed, recruiting, and planned trials don't have an ambiguous "why" to reason about, so we don't force one — we skip the AI pass entirely and pull together what's already on the public record:
- Phase, study type, and condition(s)
- Sponsor
- Timeline — start date and (estimated or actual) completion date
- Enrollment — target or actual, labeled accordingly
- Who can join — age range, sex, healthy volunteers
- Trial sites — how many, and in which countries
- Results, if posted — outcome measures and adverse event counts
- Linked publications, if any
If a completed trial has posted results, we show what was measured — outcome titles and counts — not a computed "did it work" verdict. The full statistical comparison always stays on ClinicalTrials.gov, linked from every snapshot.
What this isn't
This tool doesn't use any data beyond ClinicalTrials.gov and PubMed, has no accounts or saved history, and its output is not medical or investment advice — always verify anything decision-relevant against primary sources.