What's Pharma R&D Really Producing? Another Take on the Weak Drug Pipeline
(As Ed Silverman is on vacation this week, we’re trying something new and getting some op-ed guest writers. Dan Hoffman, Ph.D., a partner in Clinical Trial Recovery Specialists, does his own analysis of a study that Sanford Bernstein analyst Tim Anderson said showed drug pipelines are still trending downward.)
Last week Sanford Bernstein’s Tim Anderson reviewed some data from KMR, a benchmarking and performance analytics group, and identified what he characterized as a “potentially disturbing” trend in pharma’s R&D productivity. Anderson’s commentary received a fair amount of attention (see Pharma news), but some subsidiary data points that haven’t been explored seem potentially more important than Anderson’s top line conclusion.
In brief, Anderson states that “pipeline success rates across all phases of development have been slowly worsening” or remaining flat, thereby requiring the industry to enter more compounds into clinical development to yield an equivalent number of approvals. Moreover, the amount of time required to go from preclinical development to registration has been steadily increasing. He concluded that if the KMR figures mean what they seem to suggest, “then drug companies will continue to limp along, desperately seeking alternate solutions to try and deliver on their stated growth objectives.”
Anderson offered some plausible reasons for this declining productivity. First, the ballooning costs of clinical trials in recent years has made pharmas more eager to cull out subpar pipeline candidates earlier, certainly before the companies have to endure the vastly higher Phase 3 costs. That may explain why success rates for Phase 3 compounds declined from 70% in 2003-2007, to 65% in 2007-2011, while Phase 2 success fell more steeply over the same period, from 34% to 22%.
Anderson speculates that a second reason may account for a higher failure rate: pharma’s move away from me-too compounds in favor of “more novel, high-value,” pricier products that are inherently riskier to develop.
But both the data and their assessment remain open to additional interpretations. Earlier this year an associate worked through some FDA data to get a read on pharma’s drug development productivity. Looking at the years 1999-2004 and 2005-2011, he compared for each period the number of compounds in development by phase, the number of approvals, the number of NMEs (new molecular entities), NMEs as a percentage of approvals, and then the number and approval percentage represented by Priority Review NMEs. He used this last metric as a surrogate, a questionable one to be sure, of “innovativeness.”
What he found was that the number of NDAs/BLAs, NMEs, and PR-NMEs hasn’t changed substantially since 1999, despite the growth in the number of compounds during each development phase. Instead of reflecting an industry-wide move to “more novel” compounds, NMEs and Priority Review NMEs as a percentage of approvals declined from 31% and 14%, respectively, during 1999-2004, to 25% and 10% from 2005 to late 2011.
To read the remainder of this article, go to Pharmalot.