Why Do So Many Phase III Clinical Trials Fail?

December 6, 2022

Learn on statistical explanations for a large proportion of Phase III clinical trials failures, and how to address both the overestimation of treatment effects, and the misinterpretation of P-values via Bayesian methods.


The clinical development of new drugs is known to be risky, with overall historical success rates from phase I to approval of less than 20%. The most critical phase of clinical development is the transition from phase II to phase III trials, with historical success rates around 40%. In other words, of all drugs that enter clinical development, 4 in 5 fail to get approved, and half of the failures occur in phase III. Phase III Clinical Trials Fail

  • The purpose of this talk will be to investigate statistical explanations for such a large proportion of failures, including the overestimation of treatment effects in phase II trials, the “winner’s curse”, and misinterpretation of P-values.
  • These ideas will be explored through an informal discussion between a statistician and the head of a clinical development program.
  • Moreover, Bayesian concepts will be explained in a non-technical manner and will be shown to lead to more sensible decision-making through an attenuation of the overestimation of treatment effects, and the replacement of P-values by posterior probabilities of the null hypothesis.


  • Phase II trials are likely to overestimate the magnitude of treatment effects
  • P-values are frequently misinterpreted, which leads to a “replication crisis”
  • Bayesian methods can address both the overestimation of treatment effects, and the misinterpretation of P-values



Marc Buyse, ScD, Chief Scientific Officer and Founder, IDDI

Marc holds a ScD in biostatistics from the Harvard School of Public Health (Boston, MA). He is the founder of the International Drug Development Institute (IDDI) and of CluePoints, two biostatistical service organizations based in the US and Europe. He is interested in clinical trial design, meta-analysis, validation of biomarkers and surrogate endpoints, statistical methods in oncology, statistical detection of errors and fraud, statistical monitoring of clinical trials, and medical data sharing.


Everardo Saad, MD, Medical Director, IDDI

Dr. Everardo Saad has nearly 20 years of experience in Medical Oncology and clinical-trial design. He graduated in Medicine and trained in Internal Medicine in Sao Paulo, and did his fellowship in Medical Oncology at the University of Texas M.D. Anderson Cancer Center, in Houston. After practicing for several years, he shifted his professional career towards education and research in Medical Oncology, and has a special interest in clinical trial methodology, the assessment of endpoints, and the development of novel therapies for cancer patients.

Presentation Webinar: Why Do So Many Phase III Clinical Trials Fail?

Please fill in the form to access the presentation of the webinar.