Key points
- Roughly 9 in 10 drug candidates that enter clinical trials are never approved.
- An analysis of industry programs from 2011 to 2020 found only about 7.9% of drugs entering Phase 1 went on to approval.
- Phase 2 is historically the hardest step: fewer than a third of programs advance from Phase 2 to Phase 3.
- Lack of clinical efficacy is the most common reason for failure, followed by unmanageable toxicity, poor drug-like properties, and commercial or strategic decisions.
- Success rates vary widely by disease area. Oncology has historically had some of the lowest rates.
Most experimental drugs fail. That is not a sign that the system is broken. It reflects how difficult it is to predict, from laboratory and animal data, how a molecule will behave in thousands of human patients. Understanding why drugs fail is the foundation for estimating which ones are likely to succeed.
How often drugs fail
Several large studies have measured success rates by following thousands of development programs over time.
The 2021 BIO, Informa Pharma Intelligence, and QLS Advisors analysis of clinical programs from 2011 to 2020 reported these average phase transition rates:
| Transition | Share of programs that advanced |
|---|---|
| Phase 1 → Phase 2 | 52.0% |
| Phase 2 → Phase 3 | 28.9% |
| Phase 3 → application submitted | 57.8% |
| Application submitted → approval | 90.6% |
| Phase 1 → approval (overall) | 7.9% |
Source: BIO, Informa Pharma Intelligence, QLS Advisors (2021). Rates are averages across all disease areas and drug types.
Different studies count programs differently. For example, a 2019 analysis in Biostatistics by Wong, Siah, and Lo used a different method and dataset and estimated a higher overall success rate of about 13.8%. The broad conclusion is consistent across studies: most drugs that enter clinical trials are never approved, and Phase 2 is the largest single point of attrition.
The main reasons drugs fail
A 2022 review in Acta Pharmaceutica Sinica B summarized analyses of clinical failures and attributed them roughly as follows:
| Reason | Approximate share of failures |
|---|---|
| Lack of clinical efficacy | 40–50% |
| Unmanageable toxicity | About 30% |
| Poor drug-like properties (for example, absorption or metabolism problems) | 10–15% |
| Lack of commercial need or poor strategic planning | About 10% |
1. The drug does not work well enough
Efficacy failures are the most common, and they have several root causes:
- The target was wrong. The biology looked convincing in cells and animals, but the target does not drive the disease in humans.
- The models misled. Animal models often fail to reproduce human disease, especially in neurology and psychiatry.
- The dose was wrong. Too low a dose shows no effect. Too high a dose causes side effects that mask any benefit.
- The wrong patients were enrolled. A drug may only help a subgroup, such as patients with a particular genetic mutation, and that effect is diluted in a broad population.
- The endpoint was wrong. A trial might measure an outcome the drug does not affect, or one that is too noisy to show a real difference.
- The comparator was strong. Beating an effective existing treatment is much harder than beating placebo.
2. The side effects are unacceptable
Some safety problems appear only in larger, longer trials. A side effect that affects one person in a thousand may never appear in a Phase 1 trial of 50 people. Toxicity is judged relative to the disease, so a side effect that would be acceptable in a cancer drug may halt a drug for a less serious condition.
3. The drug behaves badly in the body
A molecule can be potent in the laboratory and still fail because it is poorly absorbed, cleared too quickly, interacts with other medicines, or does not reach the tissue where it is needed.
4. Business and strategic decisions
Not every discontinued program failed scientifically. Companies stop drugs because a competitor reached the market first, the expected market is too small, priorities changed after a merger, or the company ran out of money. These terminations are real failures for the program, but they say little about the underlying science.
Why Phase 2 is the hardest step
Phase 2 is usually the first time a drug is tested for effect in the target patient population at a meaningful scale. It is where the assumptions made during discovery are tested directly. A drug that was safe in Phase 1 has passed only the first and easiest test. Whether it can actually help patients is still unknown.
Phase 3 failures are fewer, but they are more costly. A common pattern is a promising but small Phase 2 result that does not replicate in a larger, more rigorous Phase 3 trial. Small trials are more likely to produce positive results by chance.
Success differs by disease and drug type
Average rates hide large differences. In the 2011–2020 BIO analysis, oncology programs had one of the lowest overall rates of success from Phase 1 to approval, while hematology programs had one of the highest. Rare-disease programs and programs that select patients using biomarkers have tended to perform better than average.
The quality of the original target matters too. Programs whose mechanism is supported by human genetic evidence have historically been about 2.6 times more likely to reach approval than those without it, according to a 2024 analysis in Nature.
How to read a failure
When a trial stops, the reason is often more informative than the fact that it stopped. On ClinicalTrials.gov, terminated, suspended, and withdrawn studies include a brief "why stopped" explanation provided by the sponsor. These explanations range from "lack of efficacy at interim analysis" to "business decision" to "slow enrollment," and they mean very different things for a drug's prospects.
When Spectral Genomics records clinical drug trial results, we distinguish:
- efficacy failures, where the primary endpoint was not met,
- safety failures, where the trial was stopped because of harm,
- operational failures, such as slow enrollment, and
- strategic terminations, which the sponsor describes as business decisions.
Each category affects the outlook for a drug, and for other drugs with the same mechanism, in a different way.
What this means for probability estimates
Historical success rates are a starting point, not an answer. A Phase 2 cancer drug with a validated biomarker and a strong randomized result is not the same bet as an average Phase 2 program. Our approval probability methodology explains how we use historical benchmarks, and why we do not simply multiply industry-wide phase rates together and call the result a prediction.
Sources
- Biotechnology Innovation Organization, Informa Pharma Intelligence, QLS Advisors. Clinical Development Success Rates and Contributing Factors 2011–2020 (2021).Industry report
- Sun D, Gao W, Hu H, Zhou S. Why 90% of clinical drug development fails and how to improve it? Acta Pharmaceutica Sinica B. 2022;12(7):3049–3062.Peer-reviewed
- Wong CH, Siah KW, Lo AW. Estimation of clinical trial success rates and related parameters. Biostatistics. 2019;20(2):273–286.Peer-reviewed
- Minikel EV, Painter JL, Dong CC, Nelson MR. Refining the impact of genetic evidence on clinical success. Nature. 2024;629:624–629.Peer-reviewed
- National Library of Medicine. ClinicalTrials.gov Data Element Definitions (Why Study Stopped).Registry
This article is for education. It is not medical advice and not investment advice. Figures are drawn from the sources above and describe historical averages, not the prospects of any particular drug. See our editorial standards and disclaimer.