Diverging Results in Clinical Success Rates for Biologics

Clinical success rates of pharmaceuticals were recently examined in two unrelated reports, and even though both analyses reviewed data collected from the company Informa, both of the studies produced reports with slightly different clinical trial success rates for pharmaceuticals. An analysis published online on May 20, 2016 in Nature Reviews Drug Discovery (NRDD) drew from Informa’s Pharmaprojects database, while the BIO report released five days later on May 25, 2016 used information collected from Informa’s Biomedtracker.

One of the differences between the reports was the time frame studied in each analysis. The NRDD piece by authors Katarzyna Smietana, Marcin Siatkowski, and Martin Møller looked at drug success rates from 1996–2014, while the BIO report looked at success rates from 2006­–2015. The NRDD publication discusses the long-term trend in success rates, while the BIO report focuses on the recent status.

Because of reporting and dataset differences between the reports, conclusions about clinical success rates from 2014­­–2015 can’t really be made, asserts Smietana. A major underlying reason is that besides the differing time ranges, the two studies had different methodologies and datasets from which to draw. Explains Smietana, the two methodologies differ: The dataset used in the BIO study includes the success rates of all indications of each product as separate data points. On the other hand, the NRDD piece tracked by compound and did not segment based on indication-meaning every compound was counted once, regardless of how many indications it garnered. Products with multiple indications were handled differently in each report, contributing to some calculation differences in success rates. In addition, the dataset used in the NRDD analysis included a slightly different range of products according to Smietana, meaning it counted clinical trial and regulatory successes both within and outside of the United States starting from 1995. 

While Smietana et al. reported that the overall success rate for pharmaceuticals from Phase I to launch increased from 2011 to 2014 (to approximately 12%), BIO reported the overall success rate from 2006 to 2015 was about 10%. According to Smietana et al., the likelihood of approval (LOA) for biologics was 18% by 2014, but according to the BIO calculations, the LOA for biologics was 12%, potentially pointing to a decrease in the approval rate for biologics. Or, it is also possible that the number didn’t really go down, and that the difference in the success rates for biologics reported in the articles was because one report factored in multiple indications separately, while the other did not.

Importantly, the way the authors of each analysis classified a biologic differed. BIO assigns biologics in line with FDA's BLAs (biologic license applications) (excluding vaccines), and the BIO report states that “in some cases FDA will designate larger molecules, such as proteins and peptides, as NMEs” (new molecular entities). The NRDD publication classifies based on how the product is obtained (i.e., what the database calls “material of origin”); hence, some products classified as biologics by Smietana et al. will not be considered as such in the BIO study.

Because the two analyses used different definitions, time frames, and underlying datasets, Smietana admits, “we’ll likely need to wait another two or three years to see if the success rates are going to stabilize or decline again.”

“The different cuts in timeframes makes it difficult to directly compare the data, but overall I view these [reports] as quite consistent,” says Michael Hay, head of intelligence products at Informa, who served as an author for both the Biomedtracker and Pharmaprojects reports. “The BIO study purposely includes all indications to better reflect the R&D dollars spent to bring a drug to market for multiple diseases. Oncology is the most common area to see a drug studied in multiple indications and different tumor types behave very differently. This could explain the differences seen in biologics, as many of the recent oncology drugs are monoclonal antibodies,” Hay concludes.

Sources: Nature Reviews Drug Discovery, BIO