ClinicalAtlas · a mode
PhaseOdds
Most drug candidates that start clinical trials never reach patients. Before you read any headline about an AI-found drug, it helps to have the usual odds in mind. Start with a hundred candidates entering Phase I and guess how many are left after each phase.
A hundred candidates
Three headlines: seen, predicted or hoped?
Each headline below is written from an entry in the ClinicalAtlas tracker. Mark each one, then see what the tracker records and where it comes from.
- Seen: a trial result was reported.
- Predicted: a computational claim.
- Hoped: a stated aim.
Honest notes
- The rates are averages over thousands of programs in all disease areas. They differ a lot by disease, by kind of drug and by company. A single candidate's odds are not the average.
- The counts are expected numbers from a hundred, rounded to whole candidates at each step. Real batches of a hundred would vary.
- Why three headlines and not twenty: we only print what we can source and re-check against the tracker. The tracker never claims that AI caused a trial's outcome, and nor does this page.
- Nothing here is medical advice.
Sources
- BIO, Informa Pharma Intelligence and QLS Advisors (February 2021). Clinical Development Success Rates and Contributing Factors 2011–2020. PDF
- Jayatunga MKP, Ayers M, Bruens L, Jayanth D, Meier C (2024). How successful are AI-discovered drugs in clinical trials? A first analysis and emerging lessons. Drug Discovery Today 29:104009. doi:10.1016/j.drudis.2024.104009
- The ClinicalAtlas tracker: each headline's own sources are shown after you mark it.