MIT study finds that the probability of clinical trial success is nearly 40% higher than previously thought

Allan B. Haberman, Ph.D


NIH Clinical Center


On December 7, 2017 we published an article on this blog entitled ”Improving Candidate Selection: Translating Molecules into Medicines”. This article was based on a December 4, 2017 symposium sponsored by Aptuit entitled “Improving Candidate Selection: Translating Molecules into Medicines”. The focus of the meeting was on improving drug candidate selection in order to improve development success.

Our article stated that “Only about 10% of drug candidates make their way from first-in-humans trials to regulatory approval. The greatest amount of attrition occurs in Phase 2. Approximately half of candidates fail at that stage, mainly due to lack of efficacy.” As we also stated in that article, drug attrition numbers have not changed since our 2009 publications, “Approaches to Reducing Phase II Attrition” and “Overcoming Phase II Attrition Problem”.

However, especially since the year 2000, drug developers have been working with increasingly newer classes of drugs. They attribute continuing high attrition rates to difficulties in working with ever-changing classes of drugs designed to treat complex diseases. Attrition thus continues to be a moving target.

Several more recent estimates of clinical trial success are comparable to those cited by participants in the Aptuit symposium, and in our own 2009 publications. For example, as pointed out by Endpoints News, BIO (the Biotechnology Innovation Organization) in a recent publication analyzing clinical development success rate from 2006 to 2015, determined that the overall likelihood of approval from Phase 1 for all drug candidates was 9.6%, and 11.9% for all indications other than cancer. (The likelihood of approval for oncology candidates was 5.1%; this is about the same as the figure for oncology success cited in our 2009 report.) Meanwhile, AstraZeneca cited a 5% success rate for its own candidates in a January 2018 analysis.

Now comes a January 2018 study by Andrew W Lo, Ph.D. and his colleagues at MIT that concludes that 13.8% of all drug development programs eventually lead to approval. This study was discussed in a February 1, 2018 article in Endpoints News by John Carroll. Dr. Lo is the Director of the MIT Laboratory for Financial Engineering.

As with earlier studies, the success rates depend on the particular indication. For example, infectious disease vaccines have the highest rate of success, 33.4%. Oncology drugs—as in most such studies—have the lowest rate of success—3.4%.

Dr. Lo’s study represents a Big Data approach to determining drug development success rates.The MIT group analyzed a large dataset of over 40,000 entries from nearly 186,000 clinical trials of over 21,000 compounds. To analyze this dataset, the researchers developed automated algorithms designed to trace each drug development path and compute probability of success (POS) statistics in a matter of hours. If generating POS estimates had been done by traditional manual methods, it would have taken months or years.

Despite the intense focus of the biopharmaceutical industry, investors, and the general public on cancer, the POS for oncology drugs has been consistently abysmal for years—as shown by our 2009 report, the 2016 BIO report, and the Lo et al. 2018 MIT study. However, according to the MIT study, although the POS for oncology drugs had the lowest overall approval rate of 3.4% in 2013, it rose to 8.3% in 2015. Both Dr. Lo’s group and John Carroll of Endpoint News attribute this sharp rise to the advent of immuno-oncology drugs.

As we discussed in our February 22, 2018 blog article, “JP Morgan 2018 (JPM18) panel optimistic for new breakthrough immuno-oncology therapies despite a crowded field”, leading researchers in academia and industry believe that because of the strong emergence of immuno-oncology therapies, now is probably the best time for progress in oncology in several decades. This is consistent with the findings of Dr. Lo’s group. However, as we stated in our previous blog article (based on the conclusions of the JPM18 panel), “This historic opportunity would be maximally capitalized if people from academia, industry, regulatory agencies, and nonprofit organizations work together, especially in adopting novel collaborative study design, aimed at bringing the promise of cancer immunotherapies to patients, sooner rather than later.”

Another issue discussed by Dr. Lo and his colleagues in their study is role of biomarkers in the success of clinical trials. The researchers compared POS estimates for trials that stratified patients using biomarkers to those that did not use biomarkers. They found that trials that utilized biomarkers tended to be more successful (by nearly a factor of 2) than those that did not. However, biomarker-stratified trials studied by the MIT group were nearly all in oncology. Therefore, it was not possible for the MIT researchers to obtain valid conclusions on the role of biomarkers for therapeutic areas outside of oncology.

Nevertheless, with the continuing development of oncology biomarkers, coupled with breakthrough R&D results in immuno-oncology, the MIT researchers expect that the rates of approval of cancer drugs will continue to improve.


Dr. Lo’s group intends to provide continuing information on the success rates of clinical trials, beyond this initial study. The goal is to provide greater risk transparency to drug developers, investors, policymakers, physicians, and patients, order to assist them in their decisions.

Moreover, our book-length report, Cancer Immunotherapy: Building on Initial Successes to Improve Clinical Outcomes can help you understand the role of advances in immuno-oncology in the current and expected increases in drug development success in the cancer field.


As the producers of this blog, and as consultants to the biotechnology and pharmaceutical industry, Haberman Associates would like to hear from you. If you are in a biotech or pharmaceutical company, and would like a 15-20-minute, no-obligation telephone discussion of issues raised by this or other blog articles, or of other issues that are important to your company, please contact us by phone or e-mail. We also welcome your comments on this or any other article on this blog.

Leave a Reply

Your email address will not be published. Required fields are marked *

Post comment