A custom collaboration with: Quintiles
Ten years ago, the sequencing of the human genome promised a new golden age of medicine. Thousands of new drug targets were discovered, suggesting new weapons that might soon defeat diseases in ways never imagined. Rising to the challenge, the U.S. Congress doubled the budget of the National Institutes of Health to $30 billion. When added to the amounts spent by large pharmaceutical firms (approximately $65 billion), biotech companies (about $30 billion), universities, foundations and foreign governments, well over $150 billion is now spent each year on biomedical research.
All expectations were that a flood of novel medications would follow. Yet, surprisingly, the number of new medicines has not reacted to this investment. Indeed, recent research1 has confirmed that the tally of new molecular entities (NMEs) approved annually by the Food and Drug Administration (FDA)—between 20 and 25—is no greater today than it was 50 years ago, prompting much speculation about the reasons for this lack of response.
One key reason is that pharmaceutical companies produce NMEs at steady rates. The most innovative firms (Merck, Lilly and Roche) have averaged almost one NME per year, but the industry mean is only about one every six years. Moreover, nothing that companies have done in the past 60 years has affected their rates of new drug production: whether large or small, focused on small molecules or biologics, operating in the 21st century or in the 1950s, companies have produced NMEs at steady rates. Neither mergers and acquisitions (M&As), nor reorganizations, nor the countless process improvement programs (such as Six Sigma and Total Quality Management) have affected their mean NME output. This stasis suggests that perhaps the industry’s NME output is not depressed, as commonly thought, but reflects the limitations of its R&D model. Nevertheless, costs per NMEs have increased at a double-digit rate (about 13.4 percent) for decades, and are now well into several billions of dollars at many companies. The industry is thus caught in a vice between an NME output that is essentially constant, and likely to remain so, and research costs that are increasing exponentially.
A statistical analysis of the 1,222 NMEs approved by the FDA since 1950 shows that the NME output of a pharmaceutical company follows a Poisson distribution: its rate of new-drug production is constant, save for stochastic variations around the mean. This finding has profound implications. If the NME output of pharmaceutical companies is flat, the only way to boost the overall industry output is to increase the number of companies, which runs counter to the surge of M&A activity of the past 12 years. Indeed, the data show that reduction in the number of companies explains 95 percent of the shortfall in the expected NME output of the industry.
The data also show that the rates of NME production at pharmaceutical companies are not randomly distributed but instead fall along an exponential curve that can be modeled by the gamma distribution. One can therefore calculate the probability that the output of a company will reach two or three NMEs per year, a level often regarded in the industry as the threshold needed to meet investors’ expectations. Unfortunately, the likelihood of producing two NMEs annually is only 0.06 percent, and for three NMEs it drops steeply to 0.003 percent. It is therefore unlikely that most companies will achieve and maintain that output, unless they make significant changes to their R&D model.
The Poisson model also makes it possible to test whether bigger is really better for pharmaceutical makers, a hypothesis that has driven much of the past M&A activity. By comparing the Poisson parameters of companies before and after their M&A transactions, one can measure how much value was created, if any. An analysis of 30 deals shows that M&As only raise NME output when they involve two small companies. For large companies, M&As do not seem either to create or destroy value. They merely take out some costs while adding more bureaucracy, 2 an impact on pharmaceutical R&D that can be summarized as “1+1=1.” This led two former directors of the bureau of economics at the Federal Trade Commission to assert that “the recent mergers [in the industry] will retard the rate of pharmaceutical innovation and impose substantial consumer harm.”3
The process of developing therapeutics has multiple expense points, which will undoubtedly change as new business models replace old ones. Here is a snapshot of the current major expenditures related to the bench-to-bedside enterprise.
© 2010 Scientific American,
a division of Nature America, Inc.
All Rights Reserved.