Issue No. 21: What It Costs to Make a Pill

A Flawed Study Fails to Refute the Basic Fact That Drugs Are Risky and Costly to Develop.

It’s Time to Focus on Supply and Demand.

Developing a drug is a risky and hugely expensive undertaking. Some 90% of publicly traded biopharmaceutical companies are not expected to make a profitthis year, and, profitable or not, such companies require massive investments in research and development.

How massive? The most thorough study of what it costs to a develop single new drug was conducted by three PhD economists: Joseph A. DiMasi, director of economic analysis at the Tufts Center for the Study of Drug Development, Henry G. Grabowski of Duke, and Ronald W. Hansen of the University of Rochester. Their papers on the subject go back to 1979 and have been cited by other researchers, including those of the U.S. government, to analyze policy questions. DiMasi and Grabowski wrote the chapter, “R&D Costs and Returns to New Drug Development: A Review of the Evidence,” in The Oxford Handbook of the Economics of the Biopharmaceutical Industry.

Tufts Research, Published in 2016, Examined 106 Drugs at Random

The most recent estimates of the three researchers were published in the May 2016 issue of the Journal of Health Economics. They looked at the research and development costs of 106 randomly selected drugs from a survey of 10 pharmaceutical firms.

These data were used to estimate the average pre-tax cost of new drug and biologics development. The costs of compounds abandoned during testing were linked to the costs of compounds that obtained marketing approval.

The researchers determined that the average out-of-pocket cost per new compound approved by the Food & Drug Administration was $1.4 billion.

They then capitalized out-of-pocket costs incurred during the process at a discount rate of 10.5% (in other words, this was the annual cost of the capital that was deployed for the research and could have earned money elsewhere). That exercise yielded a pre-approval cost estimate of $2.6 billion. They then added post-approval R&D costs and finished with an estimate of $2.9 billion, all in 2013 dollars.

DiMasi and his colleagues did not simply say: OK, here is what was spent for R&D on a particular drug that was approved by the FDA. They appropriately added in costs for drugs that underwent costly R&D but were not approved. The Tufts study, as it is often called, also appropriately adds in the cost of capital – standard operating procedure for an analysis of this sort.

Critics, who sometimes have a political agenda – see, for example, this press release from 2001, issued by Ralph Nader’s organization – have attacked the Tufts research for years, offering much lower figures for the cost of developing a drug. But the critics’ work has been consistently inadequate.

Study by Two MDs Looks at 10 Drugs That Were Out-of-the-Gate Winners

In this vein comes a study published by JAMA Internal Medicine on Sept. 11 by Vinay Prassad of the Oregon Health and Science University and Sham Mailankody of Memorial Sloan Kettering Cancer Center. Both are medical doctors (Prassad also has a master’s degree in public health), but neither is an economist. And it shows.

Prassad and Mailankody took what can only be called a quick-and-dirty approach. That is not necessarily a disparagement. Great economics papers are often simple and elegant. In this case, the researchers looked at only 10 drugs, produced by 10 companies. Each of the companies received FDA approval for its drug between 2006 and 2015 and had no drugs on the market prior to the approval. The authors used filings with the U.S. Securities & Exchange Commission to determine R&D spending by the companies during the period when the successful drugs were under development.

They found that the median cost of R&D was $648 million. They added another $109 million for a 7% per annual cost of capital or $145 million for a 9% cost of capital. Either way, their figures are less than one-third those that DiMasio and colleagues came up with.

The study, however, contained serious deficiencies. Quoted by STAT News, Bernard Munos, a senior fellow at FasterCures, said, “The study starts with a good intent, but suffers from severe flaws that invalidate its results…. An informed debate on R&D costs and drug prices must rest on rigorous analyses. This one fails the test.”

The biggest problem is that Prassad and Mailankody focus only on winners. They looked at small companies that each gained FDA approval for their first drug. Of course, such companies will have lower costs than companies that have several successful drugs and many more drugs that fail to gain approval.

Study Understates the Cost of Clinical Failure

Failure is norm for developing cancer drugs. These 10 companies were exceptions – and good for them, but bad for anyone trying to make important policy decisions based on their experience. Between 1998 and 2014, only 10 drugs were approved by the FDA for lung cancer, compared with 167 unsuccessful drugs in the clinical development pipeline. That is a success rate of just 6%; for melanoma, the success rate during the period was 7%; for brain cancer, 4%.

The authors say they accounted for failures by including R&D costs associated with all the not-yet-approved drugs in the portfolios of the 10 companies. But even if all those drugs fail, the companies in the study would achieve an astounding success rate of 23%.

As DiMasi himself, interviewed by STAT News, put it:
 
There are a number of serious flaws with this study…. Most importantly, it includes only a small set of companies that were relatively successful in development during the period they analyzed. As a result, it inadequately adjusts for risks across the industry and so undercounts the costs of investigational cancer drug failures.

Derek Lowe’s blog In the Pipeline, published by Science Translational Medicine, writes that neglecting to account adequately for the cost of failure “sinks the entire paper.” He continues,

We have in this business somewhere around a 90% failure rate in the clinic. Picking companies’ first approvals disproportionately selects for the fortunate ones who succeeded their first time out. In other words, this estimate ignores (as much as possible) the cost of clinical failure, and that cost is one of the central facts of the entire drug industry.

Evidence of this “central fact” is that companies are willing to pay huge sums in order to avoid failure risks. In 2011, for example, Gilead Sciences paid $11 billion to purchase Pharmasset, a company that was developing – but had not even gained approval for – a Hepatitis C drug.

Out of 127 Attempts, Just 4 Medicines for Alzheimer’s Symptoms

By downplaying failure, the Prassad-Mailankody study misses the main dynamic of drug discovery. Only 0.2% of molecules show enough promise for testing in humans and only 20% of medicines starting phase I human clinical trials receive FDA approval. Since 1998, there have been 127 projects started by the biopharmaceutical industry for treatment of symptoms associated with Alzheimer’s Disease, and only four medicines have been approved. But failures provide lessons that lead to innovation – as recent approvals for lung-cancer medicines have shown.
 
Another example of how years of failed research can lead to improved care is in Chronic Myelogenous Leukemia (CML). After the approval of the first breakthrough therapy for CML and multiple agents afterwards, the five-year survival of patients afflicted with this deadly disease has increased from 31% to 89%. This achievement is due to the resiliency of researchers undeterred by initial failure.
 
To gain FDA approval for a single product, researchers, on average, test more than 10,000 promising compounds. More than 200 scientists work for close to 15 years, conducting more than a dozen human clinical trials involving over 3,000 patient volunteers.

Other Problems With the Prassad Study

There were other problems with the Prassad-Mailankody study as well. It did not include early R&D costs, nor the post-approval costs that were cited in the 2016 Tufts study. In addition, it understated capital costs at 7% to 9% while the Tufts research used 10.5%, based on a widely accepted capital-asset pricing model. And, by focusing on smaller companies, the researchers gave a distorted picture by minimizing failure. Large, established biopharmaceutical companies have significantly higher failure rates because they stay in business after several iterations of unsuccessful tries. On the other hand, smaller companies that do not succeed immediately tend to quickly faze out of the arena of medical research.

More important, unlike Prassad-Mailankody, the Tufts study is a serious piece of research, built on detailed evidence developed over many years. In a section of their 2016 paper, the authors carefully examine and respond persuasively to the arguments of their critics. The 2016 paper is the fourth in a series of analyses of drug costs by DiMasio and his colleagues, and it found that out-of-pocket costs had increased at an annual rate of 9.3% between the 1990s and the 2000s and early 2010s.

It is an undeniable fact that developing a new drug is expensive and getting more so. Companies that develop successful drugs are rewarded, and then plow their profits into more R&D in the hopes of developing future successful drugs.

The three largest U.S. pharmaceutical companies by revenues last year produced average returns on total capital of 15% and returns on share equity of 20%. Those returns are good but hardly out of line with other well-run companies, such as Microsoft, with capital returns of 20% and equity returns of 31%, or Procter & Gamble, with returns of 15% and 19%, respectively.

Drug-company returns are absolutely necessary for new research. After all, the funding (the nearly $3 billion per drug) comes from private investors. They won’t invest – and innovation won’t be forthcoming – without the generation of adequate profits.

In the end, we can’t have a meaningful discussion of health-care costs unless we approach the basic facts in a rigorous, dispassionate fashion. The Tufts study has done that, and, try as they might, critics can’t lay a glove on it.