How we show drug approval odds
Every development program on a Spectral Genomics drug profile shows a historical benchmark: how often comparable programs at the same stage went on to FDA approval. This page explains where those figures come from, how to read them, and the rules we follow.
In short
- The odds we publish are historical benchmarks: average outcomes for past programs at the same stage, labeled as such.
- They are not predictions for any individual drug, and we do not publish drug-specific probabilities.
- Each benchmark applies to one development program, a specific drug for a specific disease, never to a drug as a whole.
What each profile shows
A drug tested in several diseases has several programs, each with its own evidence and its own stage. For each program, the “Development programs” table on the profile shows two figures for the program’s current stage:
- Eventual approval: the share of past programs at that stage that were eventually approved.
- Next step: the share of past programs at that stage that advanced to the next one, such as Phase 2 to Phase 3.
Where a benchmark would mislead, for example after approval for that use or after a failed pivotal trial, the table says so instead of showing a number.
The benchmark figures
The figures are industry-wide averages for programs from 2011 to 2020, published by the Biotechnology Innovation Organization, Informa Pharma Intelligence, and QLS Advisors.
| Current stage | Next step | Advanced | Eventually approved |
|---|---|---|---|
| Phase 1 | Phase 1 → Phase 2 | 52.0% | 7.9% |
| Phase 2 | Phase 2 → Phase 3 | 28.9% | 15.1% |
| Phase 3 | Phase 3 → application | 57.8% | 52.4% |
| Application submitted | Application → approval | 90.6% | 90.6% |
Source: BIO, Informa Pharma Intelligence, QLS Advisors. Clinical Development Success Rates and Contributing Factors 2011–2020 (2021). All therapeutic areas and drug types combined. “Eventually approved” figures are the report’s likelihood-of-approval table.
Odds by therapeutic area
Success rates differ widely between diseases. The same report breaks likelihood of approval down by area. We show these figures on our therapeutic area pages.
| Area | From Phase 1 | From Phase 2 | From Phase 3 | Programs |
|---|---|---|---|---|
| All indications | 7.9% | 15.1% | 52.4% | 12,728 |
| Cardiovascular | 4.8% | 9.6% | 45.6% | 651 |
| Metabolic | 15.5% | 25% | 55.7% | 399 |
| Endocrine | 6.6% | 15.2% | 57.1% | 887 |
| Oncology | 5.3% | 10.8% | 43.9% | 4,179 |
Drug profiles use the all-indication figures. The report does not publish which diseases it assigned to each area, so matching a specific program to an area rate would be our guess, not the report’s.
Why a benchmark is not a prediction
A benchmark describes a group of past programs. A common shortcut is to multiply the rates for each remaining phase and call the result a drug’s chance of approval. We do not, for three reasons:
- It ignores the evidence. A drug that just produced a strong, randomized Phase 2 result has different prospects from one with ambiguous data, even though both are “in Phase 2.”
- It ignores context. Success rates differ by disease area, by drug type, and by whether patients are selected with biomarkers.
- It assumes the phases are independent. In reality, the reasons a drug passes Phase 2 affect its chances in Phase 3.
That is why each profile pairs the benchmark with our sourced reading of the evidence and a list of what could change the outlook. For the background, see Why Do Experimental Drugs Fail? and How Are Drug Approval Probabilities Calculated?
Rules we follow
- Every figure is labeled as a historical benchmark and cites its source and period.
- We never present a benchmark as the chance that a specific drug will be approved.
- We publish a drug-specific probability only if the method behind it has been tested against the real outcomes of past programs and beats these benchmarks. We do not publish any today.
- Approval odds are not an expected stock return and never a recommendation to buy or sell a security.
Approval odds on Spectral Genomics are statistical summaries of past drug development programs. They are not medical advice and not investment advice. See our disclaimer.