Drug development

Why Do Phase 3 Clinical Trials Fail?

Roughly 4 in 10 Phase 3 programs fail to reach a marketing application. Most fail because the drug's benefit doesn't hold up in a larger, more rigorous trial, often because the Phase 2 result was overstated.

Key points

  • In the 2011–2020 BIO industry data, 57.8% of Phase 3 programs advanced to a marketing application, so about 42% did not.
  • In a study of 640 novel therapeutics that entered pivotal trials, 57% of failures were due to inadequate efficacy, 17% to safety concerns, and 22% to commercial reasons.
  • A common pattern is a promising small Phase 2 result that shrinks or disappears in a larger trial.
  • Changes between Phase 2 and Phase 3 in endpoint, population, comparator, or dose are frequent sources of failure.
  • Only 40% of failed pivotal trials in that study were published in peer-reviewed journals, so much of the evidence from failures is never fully seen.

Phase 3 trials are the large, expensive studies designed to confirm that a drug works. By the time a drug reaches this stage it has usually passed safety testing and shown promise in Phase 2, yet a large share still fails. Phase 3 failures are among the most costly events in drug development, for the companies involved and for the patients waiting on the results.

How often Phase 3 fails

In the 2021 BIO, Informa Pharma Intelligence, and QLS Advisors analysis of 1,928 Phase 3 transitions from 2011 to 2020, 57.8% of programs advanced to a marketing application. About 42% did not.

Rates varied by disease area. Oncology had the lowest Phase 3 success rate of the major areas in the report, at about 48%, and ophthalmology (51.2%) and neurology (53.1%) were also below average. Hematology was highest, at 76.8%.

The main reasons

A study in JAMA Internal Medicine (Hwang and colleagues, 2016) followed 640 novel therapeutics that entered pivotal trials between 1998 and 2008. Of these, 344 (54%) failed in clinical development. The reasons:

Reason for failure Share of failures
Inadequate efficacy 57%
Commercial reasons 22%
Safety concerns 17%

Source: Hwang et al., JAMA Internal Medicine (2016). The remaining failures had other or unclassified reasons.

Why efficacy disappears in Phase 3

The Phase 2 result was too good

The most common story behind a Phase 3 failure begins in Phase 2. Small trials produce noisy results. When dozens of Phase 2 programs are run, some will look impressive purely by chance, and those are exactly the ones selected to move forward. When a larger trial measures the same effect more precisely, it tends to come out smaller. This is a form of regression to the mean.

This pattern is not limited to drug development. In a 2005 study in JAMA, John Ioannidis examined 45 highly cited clinical studies that reported an effective intervention. Later research contradicted 16% of them and found weaker effects for another 16%. Among the randomized trials, those later contradicted or weakened tended to be smaller.

Phase 3 tested something different

Phase 3 often changes the question in ways that seem minor but are not:

  • A different endpoint. Phase 2 might show an effect on a biomarker or symptom score, while Phase 3 must show a benefit on an outcome patients and regulators care about more, such as survival or hospitalization.
  • A broader population. A Phase 2 signal from a narrow, carefully selected group may be diluted when Phase 3 enrolls a wider population.
  • A stronger comparator. Beating placebo is easier than beating the current standard of care.
  • A different dose or regimen, sometimes chosen for convenience or safety rather than effect.

The trial was not big enough

A trial powered to detect a large effect will often miss a smaller real one. If the Phase 2 result overstated the effect size, a Phase 3 trial designed around that estimate can be underpowered from the start.

The biology was never right

Sometimes the drug does what it was designed to do but the target does not drive the disease as much as believed. Phase 2 may have shown a change in a biomarker, but Phase 3 reveals that the change does not translate into clinical benefit.

Why safety problems appear late

Phase 3 trials typically enroll 300 to 3,000 participants and follow them for longer than earlier studies. An adverse effect affecting one person in a thousand may not appear at all until a drug is given to that many people. Phase 3 is also the first time a drug is compared directly against alternatives at scale, which can reveal that its side effects are unacceptable relative to existing options.

Commercial failures

More than a fifth of the late-stage failures in the Hwang study were attributed to commercial reasons. These include a competitor reaching the market first, a smaller expected market, a change in company strategy, or a lack of funding. A commercially terminated drug may still have worked. The trial simply did not continue long enough to find out.

The hidden cost: unpublished failures

The same study found that pivotal trial results were published in peer-reviewed journals for only 40% of the failed drugs. Among drugs abandoned for commercial reasons, just 8.1% of trials were published. When failures go unpublished, other researchers cannot learn from them, and the evidence base for every drug with the same mechanism is weaker. This is one reason Spectral Genomics records clinical trial results from registries, and trial terminations, alongside publications.

Warning signs before a Phase 3 readout

None of these guarantee failure, but each is associated with higher risk:

  • The Phase 2 result came from a small or uncontrolled trial.
  • The Phase 2 effect appeared only in a subgroup or a secondary endpoint.
  • Phase 3 uses a different primary endpoint, population, or comparator than Phase 2.
  • Other drugs with the same mechanism have failed in the same disease.
  • The disease area historically has a low Phase 3 success rate.
  • The sponsor has changed the trial's primary endpoint or sample size mid-study without a clear prespecified reason.

Sources

  1. Hwang TJ, Carpenter D, Lauffenburger JC, Wang B, Franklin JM, Kesselheim AS. Failure of investigational drugs in late-stage clinical development and publication of trial results. JAMA Internal Medicine. 2016;176(12):1826–1833.Peer-reviewed
  2. Biotechnology Innovation Organization, Informa Pharma Intelligence, QLS Advisors. Clinical Development Success Rates and Contributing Factors 2011–2020 (2021).Industry report
  3. Ioannidis JPA. Contradicted and initially stronger effects in highly cited clinical research. JAMA. 2005;294(2):218–228.Peer-reviewed
  4. U.S. Food and Drug Administration. Step 3: Clinical Research.Regulator

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.