Clinical Development Success Rates: What the Numbers Really Say
"Nine in ten drugs fail" is close, but it hides where they fail and why. What the large studies show, how devices differ, and the part people play.
FloodGate Medical · October 20, 2026 · 3 min read
"Ninety percent of drugs that enter clinical trials fail" is repeated in pitch decks and keynotes everywhere. It's roughly right, but a single number hides most of what matters: where in development programs fall away, which kinds of programs fare better, and how much of the outcome comes down to execution rather than science.
What the large studies show
The most widely cited figures come from the Biotechnology Innovation Organization's analysis of development programs from 2011 to 2020, produced with Informa Pharma Intelligence and QLS Advisors. It found that a program entering Phase I had a 7.9% chance of reaching approval. The odds of moving from one stage to the next were:
| Transition | Success rate |
|---|---|
| Phase I to Phase II | 52.0% |
| Phase II to Phase III | 28.9% |
| Phase III to submission | 57.8% |
| Submission to approval | 90.6% |
Source: BIO, Clinical Development Success Rates 2011 to 2020
A different method gives a different answer. A study by Wong, Siah and Lo, which followed individual drug and indication pairs through trial data from 2000 to 2015, estimated an overall success rate of 13.8%. Both studies agree on the shape: most programs that enter the clinic never reach patients.
Phase II is where most programs fail
Phase II is usually the first real test of whether a therapy works in patients with the condition, rather than in healthy volunteers or animal models. Fewer than three in ten programs clear it. Many that looked convincing in early data fail to show a meaningful benefit at this stage.
Success varies widely by therapeutic area
Averages flatten big differences. In the BIO analysis, hematology programs had the highest chance of approval from Phase I, at 23.9%, and oncology among the lowest, at 5.3%. Tumor biology is hard to predict from early data, and cancer trials often compare against strong existing treatments.
Why programs fail
Most failures come down to the therapy itself. In a study of novel therapies that entered Phase III, published in JAMA Internal Medicine, more than half of the failures were for lack of efficacy and about one in six for safety. Commercial and strategic reasons accounted for most of the rest.
Not every failure is scientific, though. Programs also stall for reasons that have nothing to do with whether the therapy works:
- Trial design: endpoints, patient populations or sample sizes that can't produce a clear answer.
- Enrollment: sites that can't recruit the patients the protocol needs, on time.
- Protocol changes: amendments mid-trial that add cost and delay.
- Funding and priorities: programs paused or dropped when money or strategy shifts.
- People: clinical operations, biostatistics and regulatory teams stretched too thin to execute well.
How medical devices differ
Device development follows a different path. Many moderate-risk devices reach the US market through the 510(k) pathway by showing they're substantially equivalent to a device already on the market, often without a new clinical trial. Higher-risk devices that need premarket approval must show reasonable assurance of safety and effectiveness, usually through clinical studies run under an investigational device exemption.
Device trials bring their own challenges. Blinding is often impossible when the treatment is a procedure. Outcomes can depend on the physician's skill and experience with the device. Designs change between iterations, so evidence must keep up with the product. And because the device is used by a clinician, training, procedure support and site selection matter as much as the protocol.
The part people play
A trial is ultimately run by people. Strong clinical operations leaders keep enrollment on schedule. Biostatisticians design studies that can answer the question being asked. Regulatory professionals engage the FDA early enough to avoid surprises. In device trials, clinical specialists and field clinical engineers support the procedures that generate the data.
Those people are in short supply, and the best of them are rarely looking for a new role. Companies that start building their clinical and regulatory teams early, before a program falls behind, give themselves the best odds of getting a therapy or device to patients.
FloodGate Medical is a retained search firm specializing exclusively in MedTech, including clinical, regulatory and quality roles. See how a search works.
Frequently asked questions
What percentage of drugs in clinical trials get approved?
BIO's analysis of development programs from 2011 to 2020 found that a program entering Phase I had a 7.9% chance of reaching approval. A study by Wong, Siah and Lo, using a different method, estimated 13.8%. Both agree most programs never reach patients.
Which phase of clinical development has the lowest success rate?
Phase II. In the BIO analysis, 28.9% of programs moved from Phase II to Phase III, because Phase II is usually the first real test of whether a therapy works in patients with the condition.
Why do clinical trials fail?
Mostly because of the therapy itself. In a study of novel therapies that entered Phase III, more than half of failures were for lack of efficacy and about one in six for safety. Trial design, enrollment, protocol changes, funding and stretched teams also stall programs.
How is medical device development different from drug development?
Many moderate-risk devices reach the US market through the 510(k) pathway, often without a new clinical trial. Higher-risk devices need premarket approval, usually through clinical studies under an investigational device exemption, and device trials face their own challenges such as blinding and operator skill.
Traditional search is built to reward transactions, not outcomes.
Building a clinical or regulatory team?
Tell us about the role. A Talent Director in your specialty will come back to you.