Pharmaceutical Research and Manufacturers of America (PhRMA), the drug industry’s Washington, D.C. lobby group, has produced two written responses to Public Citizen’s July 2001 report which found that the industry exaggerates the expense, risk and innovation of its research and development (R&D).1
The first is a paragraph posted July 23, 2001 on the PhRMA website that claims R&D costs now exceed $500 million for every new drug, according to estimates by two private consulting firms, Lehman Brothers Healthcare and the Boston Consulting Group. The second is an August 8, 2001 eight-page document prepared by Ernst & Young LLP at the request of PhRMA. Ernst & Young is a firm that has done paid lobbying for the drug industry.2
In sum, PhRMA’s responses are poorly sourced, short on specifics and misleading. Below are Public Citizen’s rebuttals. The rebuttals are organized so as to mirror the structure of PhRMA’s responses. Hopefully, this will provide journalists, policy makers and health care advocates a point-by-point guide to the mounting credibility problems of the drug industry.
I. Lehman Brothers Healthcare and Boston Consulting Group Estimates
PhRMA initially responded to Public Citizen’s R&D report by citing two estimates on R&D costs that exceed the $500 million-per-new-drug figure that PhRMA derived from a study by Joseph DiMasi at the Tufts Center for the Study of Drug Development (a research organization affiliated with Tufts University that receives funding from drug companies).3
PhRMA claimed that Lehman Brothers Healthcare Group (a division of the investment banking firm) estimated that R&D costs were now as high as $675 million per new drug.4 PhRMA also said that the Boston Consulting Group (a firm that offers strategic management advice to business) put current average development costs between $590 million to $800 million for new drugs.5
Public Citizen’s response:
In the Lehman Brothers estimate, it is unclear what the methodology is, where the numbers came from, whether the opportunity cost of capital is included, and whether the figures are pretax or after-tax. (The latter two methodological issues were used by Public Citizen to critique PhRMA’s $500 million per new drug claim.) In addition, this work does not appear to have been peer-reviewed. (One reason Public Citizen was able to deconstruct the DiMasi study is because it was published in an academic journal and its methodology was clear, as was the source of its data.)
The Lehman estimate includes a huge figure ($230 million) for screening and finding new drug leads.6 There is no mention of the key contribution the National Institutes of Health (NIH) makes to this part of the process.7
It also assigns stratospheric costs to clinical research ($169 million for each of phases II and III; DiMasi, by comparison, put all clinical and approval costs at $75 million in 1987, or $114 million in year 2000 dollars). A recent report by the Congressional Research Service found that Phase I costs were $10 million, Phase II costs were $20 million, and Phase III costs were $45 million.8
More important, this same Lehman Brothers document stresses that R&D should become more efficient and cheaper with new technologies. Specifically: "Access to these technologies has led to a growing school of thought that the cost of discovering new biological targets and the cost of creating drug leads is falling."
In addition, another Lehman Brothers document predicts that the R&D cost per new chemical entity (NCE) will drop by $150 million between 1996 and 2005.9
PhRMA can’t have it both ways. It can’t use the high Lehman Brothers estimate and at the same time ignore the fact that Lehman Brothers say R&D costs will decrease by 25 percent in the next several years.
The same basic critique applies to the Boston Consulting Group. Specifically, where did the BCG numbers come from, what was the BCG methodology, who peer-reviewed the BCG methodology, is the estimate pretax, does it include the cost of capital?
The answers aren’t clear because this report isn’t ostensibly about R&D costs. It’s a report about "The Impact of Genomics" and how genomics will revolutionize R&D, reducing costs by as much as $300 million per drug.
BCG reveals this much about its methodology: It is based on "more than 100 discussions at nearly 50 companies and academic institutions."10 (More than 100 discussions sounds impressive, but Public Citizen’s reply is: "Show us the data.") Beyond that, details about the BCG methodology are unclear.
II. Ernst & Young Report
The Ernst & Young report claims that Public Citizen deviated from "standard methodologies," presented "selective evidence" and misunderstood R&D tax deductions, the integrity of DiMasi’s data, the significance of new chemical entities, the importance of "me-too" drugs, and risks associated with pharmaceutical R&D.
More specific details on the Ernst & Young report and Public Citizen’s rebuttals are below. But it is important to stress at the outset of this rebuttal that Ernst & Young did not challenge Section II of Public Citizen’s report.
That is significant because Section II used PhRMA’s own R&D spending data to show that drug companies spent roughly $71 to $150 million on R&D per drug (including failures) for drugs approved by the FDA in 1994-2000. That is roughly one-fifth of what PhRMA claims drug companies spend on every new drug.
While it is important to deconstruct DiMasi’s study because it is widely recognized as the basis for PhRMA’s $500 million-per-drug claim, it is perhaps more instructive to look at PhRMA’s own aggregate R&D spending data, which is what Section II of the Public Citizen report did.
In short, PhRMA’s data reveals that drug companies spent an average of $150 million for each new molecular entity (NME) brought to market in the years 1994-2000. However it is important to point out that there are many kinds of drugs approved each year – not just the elite group of new chemical entities (NCEs) – used in DiMasi’s study. (DiMasi’s NCEs are virtually identical to the NMEs that the U.S. Food and Drug Administration tracks in the drug approval process.) It is doubly important to look beyond NCEs or NMEs because PhRMA’s aggregate R&D spending figures – the figures it constantly touts – are for all drugs approved for market.
When PhRMA’s aggregate R&D spending (an average of $17 billion per year for 1994-2000) is divided by all new drugs approved for market (not just NCEs or NMEs) in that period (an average of 95.3 drugs per year) we see that the average R&D cost for a new drug brought to market in 1994-2000 was $179 million before R&D tax deductions; after tax deductions, the figure is $118 million. If we account for the approximate seven-year lag between R&D spending and drug approval (and use R&D spending from 1988-1994 for drugs approved 1994-2000) we see that R&D spending per new drug averaged $71 million after R&D tax deductions are accounted for.
Those estimates are generous to drug companies because they rely on PhRMA figures that have not been verified by an independent source and are reputed to contain hidden marketing costs.
Public Citizen’s response:
Summary
The Ernst & Young "Summary" starts by saying that Public Citizen did not, in some instances, use "standard methodologies" in its report. In fact, that was our stated purpose – to point out that methodologies such as DiMasi’s include "costs" that are not literally incurred in a company’s net expenditures on R&D.
Ernst & Young also claim that Public Citizen presents "selective evidence." Public Citizen notes that our key evidence comes from the following sources: The National Institutes of Health, Congressional Research Service, U.S. Food and Drug Administration, congressional Office of Technology Assessment, Fortune magazine and Joseph DiMasi. The sources of our evidence are elaborately detailed in 73 endnotes and four appendices, which include primary source documents. The sources of Ernst & Young’s "evidence" amount to four footnotes, one of which cites its client, PhRMA. Ernst & Young also make sweeping generalizations without any specific evidence. (See below on "Taxes")
Public Citizen also stresses that its R&D report was reviewed by five experts in pharmaceutical economics, including Stephen Schondelmeyer, director of the PRIME Institute at the University of Minnesota. By comparison, PhRMA’s report by Ernst & Young does not appear to have been reviewed by any independent experts and Ernst & Young LLP was hired last year to lobby for the drug industry.
Opportunity Cost of Capital
Calculating the opportunity cost of capital may be a common financial practice, but our aim was to stress to journalists and policy makers that 51 percent of the total DiMasi/PhRMA R&D estimate was a theoretical expense – opportunity cost – not a real one. Drug companies do not "pay" for opportunity cost in the literal sense.
Furthermore, the congressional Office of Technology Assessment (OTA) stressed in the first bullet-point of its 1993 report ("Pharmaceutical R&D: Costs, Risks and Rewards") the "actual cash outlay" (or out-of-pocket cost) of pharmaceutical R&D ($65 million per drug). Actual cash outlay does not include opportunity cost. This obviously means that the OTA considered the subtraction of opportunity cost to be a valid way to measure "actual cash outlay" – or real spending on R&D. Why else would it be the first bullet point in the summary of OTA’s 354-page report?
Moreover, five of the seven academic studies on drug industry R&D cited by DiMasi do not use opportunity cost to calculate R&D expenses for new drugs. This fact is contrary to the Ernst & Young claim that "[u]nlike other studies, the Public Citizen estimates do not include the opportunity cost of capital."
Taxes
First, contrary to Ernst & Young’s claim, pre-tax value does not more accurately reflect the true value of resources devoted to R&D. As DiMasi said: "In many applications, the relevant concept is the after-tax cost."11 This is what the OTA stressed in its book-length report, which was reviewed by a panel of experts.
Second, it is inaccurate to suggest that most big drug companies are limited in their ability to take R&D deductions. As Uwe Reinhardt of Princeton University wrote in a recent article for Health Affairs, "[V]irtually all drug firms now treat the year’s total R&D outlay as an expense on their income statements."12 In large part, that’s because drug firms must do so to receive additional tax credits the federal government allows for R&D.13
In addition, this is a case where Ernst & Young make a sweeping statement without a milligram of specific evidence. What specific drug company is "limited in its ability to take tax deductions?" Based on what evidence? They provide none.
Regarding the industry’s tax burden, Public Citizen only cited – and stands by – a report by the Congressional Research Service (CRS) published in December 1999. That CRS document found that the industry enjoyed almost $4 billion a year in tax credits and had an effective tax rate of 16 percent compared to an effective tax rate of 27 percent for all other industries.14
As for the claim that the drug industry leads others in taxes paid as a percentage of revenues, the data supporting the Ernst & Young claim is unclear. It appears that this assertion is based on the tax returns of just 3 drug companies – as compared to 1,593 service companies, 1,149 wholesale and retail trade companies, 488 construction firms, 33 mining companies and 209 transportation and public utility organizations in the table shown. Clearly the three drug companies are too small a sample to be meaningful.
The DiMasi Data
Ernst &Young’s point here – that OTA found consistency between DiMasi’s estimates for R&D outlays per NCE and a similar OTA calculation based on PhRMA’s aggregate R&D data – is misleading. First, OTA is clear in stressing that the integrity of DiMasi’s data is unknown. (The OTA issued this warning: "Any company that understood the study methods and the potential policy uses of the study’s conclusions could overestimate costs without any potential for discovery. Thus, the motivation to overestimate costs cannot be discounted.") The same can be said of PhRMA’s aggregate data, which was not verified by any independent source. In other words, OTA compared DiMasi’s conclusion drawn from unverified industry data with an OTA calculation drawn from PhRMA’s unverified data. OTA then found consistency between DiMasi and PhRMA data.
Second, Ernst & Young dodge the OTA’s larger point: while DiMasi’s estimate of R&D outlays per new NCE is consistent with an aggregate analysis by OTA, DiMasi’s estimate of outlays is pre-tax, as is PhRMA’s aggregate data. Therefore, they do not represent the net expenditures of drug companies on R&D. "Adjusting for tax savings," the OTA states, "reduces [DiMasi’s] net cash outlays per NCE from $127.2 million to $65.5 million."15
(As DiMasi said in his study: "In many applications that utilize pharmaceutical R&D cost estimates, the relevant concept is the after-tax cost.")16
Third, PhRMA’s own aggregate data shows that R&D costs are far below $500 million for every new drug. Public Citizen’s analysis of PhRMA’s aggregate R&D data shows that new drugs brought to market between 1994 and 2000 had average after-tax R&D costs of between $71 million and $150 million.
Finally, Public Citizen never said that DiMasi’s R&D data included marketing costs. But a special Senate Committee investigation did reach such a conclusion about pharmaceutical industry R&D data, in general.17 Moreover, there is increasing evidence that Phase IV trials are used heavily for marketing purposes and these expenses – and other hidden promotional expenses – are included in industry R&D estimates.
NCEs
Ernst & Young claim that Public Citizen criticizes the DiMasi study for only including new chemical entities (NCEs). That is not accurate. Rather, Public Citizen noted that DiMasi’s R&D estimates were based only on the cost of NCEs – which are the most expensive of new drugs approved for market. This is particularly important because when PhRMA touts total industry R&D spending it includes spending on all drugs, many of which are not NCEs. Thus, PhRMA emphasizes, on one hand, industry spending on all new drugs because it inflates overall industry R&D spending. Then, on the other hand, PhRMA suggests that R&D estimates should be based only on NCEs when critiquing the Public Citizen study. In addition, many NCEs are "me-too" drugs and are not as innovative as the industry would have the public believe. Instead, they mimic already existing drugs and are intended to capture a share of a burgeoning market. Thus, the FDA categorizes these drugs as lacking significant therapeutic benefit.18
Me-Too Drugs
Dr. Marcia Angell, former editor-in-chief of The New England Journal of Medicine, agrees with Public Citizen on this point. "We already have many Claritin-type drugs for allergies," Angell recently said in an article in The American Prospect magazine. "How many do we need? They are all very similar, but what is happening is that the drug companies focus their resources on marketing ‘copycat’ drugs to convince doctors and patients that somehow theirs is better than all the others." For Angell this emphasis on marketing (Claritin was more heavily advertised in 1998 than Bud and Coke.19) as opposed to new therapeutic discovery is troubling. "It is very hard to justify enrolling human subjects in a trial if the trial is fundamentally trivial," she explained.20
Furthermore, Ernst & Young make a dubious claim that the proliferation of me-too drugs leads to increased competition and lower prices within therapeutic classes. Let’s look at the evidence. Celebrex and Vioxx are very similar drugs (anti-arthritic Cox-2 Inhibitors). Yet in the year 2000, the competition between Celebrex and Vioxx did not lead to lower prices – as Ernst & Young suggested.21 The average price of a Celebrex prescription increased 11 percent over 1999; for Vioxx the increase was 7 percent. In the class of oral antihistamines, the average price of a Claritin prescription increased 6 percent last year; the price of Allegra, a similar drug, rose 9 percent.22
Risk
First, Ernst &Young assert that thousands of compounds are screened for every one brought to market; but where’s the risk to the industry if taxpayer-funded scientists are doing much of the screening? An internal NIH document obtained by Public Citizen reinforces the fact that taxpayer-funded scientists are conducting much of the basic research that leads to the discovery and development of drugs. (That NIH document found that 55 percent of the published research that led to the discovery and development of the five best-selling drugs in 1995 conducted by publicly-funded researchers not industry scientists.)
Second, a recent Lehman Brothers document stresses that R&D should become more efficient and cheaper with new technologies, mergers, and re-engineering within companies. So, while thousands of compounds are screened for every successful drug, the time it takes to screen thousands of drugs will be dramatically reduced.
For instance, Boston Consulting Group predicts that the number of samples screened per day will soon increase from 1,000 to 100,000. In turn, BCG predicts that development success rate will improve by 100 percent and companies will increase the number of NCEs they produce annually five-to-tenfold by the year 2003.
Third, Ernst & Young claim that a small percentage of NCEs tested on humans (14 percent from 1964 to 1975) are actually approved for market. What they don’t mention is that almost one-third of such "failed" drugs – according to a study by DiMasi – are not really failures at all.23 Instead, companies stop research on them strictly for "economic" reasons – in other words, because they will not be as profitable as other drugs in the development pipeline.
Specifically, DiMasi’s study (published in the journal, Clinical Pharmacology & Therapeutics) shows that 29 percent of so-called NCE "research failures" are due to "economics" rather than efficacy (41 percent) or safety (26 percent).24 Those statistics cover the period 1985-1989. The trend in the reasons for abandoning research on NCEs, shows "economic" reasons increasing over time, from 18 percent for 1964-1969 to 29 percent for 1985-1989. So it's fair to say that, according to DiMasi, profitability is increasingly a reason for stopping research on otherwise safe and efficacious drugs.
A final note on risk: OTA stressed that the returns on pharmaceutical investments were more than enough to induce more investment in R&D – so where’s the risk if that’s the case?
Changes in R&D Costs
Ernst & Young claim that several factors, "including longer clinical trials," have increased the costs of R&D. This statement is flat wrong according to a recent study by the Tufts Center for the Study of Drug Development (where DiMasi works). The Tufts Center – which receives drug industry funding – reports that clinical trial time has decreased 19 percent in recent years.25
Furthermore, the biggest, most efficient drug companies average far less clinical trial time than the industry average. For example, Glaxo takes just 4.1 years while the industry median is 5.7 years. Merck takes 4.5 years, Pfizer 4.6 years and Abbott 4.6 years.26
Ernst & Young also assert that attrition rates for R&D projects remain high. This is another cynical stab on their part. Attrition rates may remain high, but if companies can soon screen 100 times more compounds a day – as the Boston Consulting Group predicts – then they will soon enjoy greater success in generating new small-molecule drugs, and reap greater financial rewards from their R&D.27
Government Research
Ernst & Young say only that the pharmaceutical industry and the public sector play complimentary roles. That conclusion is obvious and does not conflict with Public Citizen’s report.
What is more germane is the importance of NIH research and the degree to which it subsidizes the industry, reduces its risk, and cuts its cost.
Drake and Uhlman capture the basic dynamic of public-private research interaction in their 1993 book ("Making Medicine, Making Money"):
"The riskiest and most costly research is the very basic research that has no clear-cut applications. It is the fundamental research that leads to the discoveries that win Nobel Prizes. This is the research that often begins with childlike questions…Commercially motivated drug companies rarely can afford the luxury of asking them, though, because no one can predict if, let alone when, the answers will come…That’s why drug companies rely heavily on tax-supported academia and the National Institutes of Health, which focuses on basic research. That also helps to explain why seventy-four NIH-supported scientists went on to win the Nobel Prize, compared to three from drug companies, even though the industry spends more on R&D than the National Institutes of Health."
For its R&D report, Public Citizen unearthed a NIH document that attempted to quantify NIH’s contribution to top-selling drugs. That NIH document showed that more than half of the published research that led to the discovery and development of the five top-selling drugs in 1995 was conducted by scientists who were funded by U.S. taxpayers.
That finding shows the degree to which public research compliments the pharmaceutical industry, and it suggests that PhRMA should do more to compliment taxpayers for their contribution to the staggering profits of the industry.
Endnotes
1. The report can be found at http://www.citizen.org/congress/drugs/R&Dscarecard.html.
2. Public Citizen, "The Other Drug War: Big Pharma’s 625 Washington Lobbyists," July 2001. According to lobby disclosure reports, one division of Ernst & Young (Washington Council Ernst & Young) was paid $160,000 to lobby for the drug industry in the year 2000. Ernst & Young LLP received another $60,000 to lobby for the drug industry in the year 2000. Ernst & Young clients include Pfizer, Eli Lilly, Aventis and Johnson and Johnson. http://www.citizen.org/congress/drugs/pharmadrugwar.html
3. According to the Tufts Center for the Study of Drug Development website, corporate sponsors include Merck, Pfizer, Bayer and other large drug companies. Corporate sponsors get to review the center’s activities and "help shape strategic objectives." http://www.tufts.edu/med/csdd/images/benefits.pdf
4. Lehman Brothers, "Drug R&D Costs, Success Rates, and Emerging Technologies: A Look At Three Future Scenarios," 1997, published in PAREXEL’s Pharmaceutical R&D Statistical Sourcebook 2000.
5. The Boston Consulting Group, "A Revolution in R&D: The Impact of Genomics," June 2001. http://www.bcg.com/publications/files/genomics.pdf
6. Lehman Brothers, "Drug R&D Costs, Success Rates, and Emerging Technologies: A Look At Three Future Scenarios," 1997, published in PAREXEL’s Pharmaceutical R&D Statistical Sourcebook 2000.
7. Public Citizen, "Rx R&D Myths," July 2001. Public Citizen obtained an internal National Institutes of Health document which reports that 55 percent of the published research that led to the discovery and development of the five top-selling drugs in 1995 was conducted by publicly-funded scientists in the U.S., most of which were underwritten by the NIH.
8. Congressional Research Service, "Pharmaceutical Research and Development: A Description and Analysis of the Process, April 2, 2001.
9. Lehman Brothers, "The Changing Economics/Dynamics of Pharmaceutical R&D due to Enabling Technologies: R&D Cost Per NCE to Fall," December 1997, published in PAREXEL’s Pharmaceutical R&D Statistical Sourcebook 2000.
10. The Boston Consulting Group, "A Revolution in R&D: The Impact of Genomics," June 2001. http://www.bcg.com/publications/files/genomics.pdf
11. Joseph DiMasi et al., "The Cost of Innovation in the Pharmaceutical Industry," Journal of Health Economics 10:107-142, 1991.
12. Uwe E. Reinhardt, "Perspectives On The Pharmaceutical Industry," Health Affairs, Volume 20, Number 5, September/October 2001.
13. Ibid.
14. Congressional Research Service Memorandum, "Federal Taxation of the Drug Industry," December 1999.
15. Office of Technology Assessment, "Pharmaceutical R&D: Costs, Risks and Rewards," 1993, page 16.
16. See n. 11, page 133.
17. Special Committee on Aging, U.S. Senate, "The Drug Manufacturing Industry: A Prescription for Profits," September 1991, pg. 5. According to the report: "[T]he ongoing Aging Committee investigation has found that many of the dollars that drug manufacturers claim are spent on research of new pharmaceutical products are actually spent on marketing research…These so-called ‘research’ activities help the companies collect the data they need to design their lavish marketing and promotional campaigns."
18. DiMasi’s study focuses only on NCEs, which he defines as drugs that have never been tested before in humans. His definition of NCE differs only slightly from the Food and Drug Administration definition of a new molecular entity, or NME. DiMasi explains: "One difference is that the FDA includes diagnostic agents and we do not. Another difference is that we include a few therapeutically significant biologics, where the FDA does not include any."
19. Gardiner Harris, "Drug Firms, Stymied in the Lab, Become Marketing Machines," The Wall Street Journal, July 6, 2000.
20. Jennifer Washburn, "Undue Influence," The American Prospect, vol. 12 no. 14, August 13, 2001.
21. National Institute for Health Care Management, "Prescription Drug Expenditures in 2000: The Upward Trend Continues," May 2001. http://www.nihcm.org/spending2001.pdf
22. Ibid.
23. Joseph DiMasi, Clinical Pharmacology & Therapeutics, 1995, 58 (1): 1-14, cited in PAREXEL’s Pharmaceutical R&D Statistical Sourcebook 2000.
24. Ibid.
25. Kenneth I. Kaitin and Elaine M. Healy, "The New Drug Approvals of 1998, 1997, and 1996: Emerging Drug Development Trends in the User Fee Era," Tufts Center for the Study of Drug Development, January 2000, published in PAREXEL’s Pharmaceutical R&D Statistical Sourcebook 2000.
26. CenterWatch, "The Speed Demons of Clinical Drug Development: The Fastest Major Pharmaceutical Companies," 1999. Published in PAREXEL’s Pharmaceutical R&D Statistical Sourcebook 2000.
27. Boston Consulting Group, "The Pharmaceutical Industry into Its Second Century: From Serendipity to Strategy," January 1999. According to this report (pages 51-56), drug companies will enjoy five-to-tenfold increases in the number of NCEs they produce annually by the year 2003. This success owes to three main innovations: 1) Genomics will allow drug companies to identify far more disease targets for drug interventions and increase, in the next decade, the number of small-molecule targets from 500 to between 3,000 and 10,000. 2) Combinatorial chemistry will allow companies to generate 10,000 times more compounds in a month while reducing the cost per compound by a thousandfold. 3) Another technological advance, high-throughput screening, will soon increase the screening capacity of companies (that is, the ability to screen targets against new compounds in order to find compounds that might be effective drugs) by a factor of 50.