Does the pharmaceutical industry exaggerate their R&D costs?

One of the principle claims for allowing pharmaceutical companies to continue their hold on current patent practices, is that research and development (or R&D) is very expensive. It just keeps coming up, and seems to be all the rage when arguing against things like the passing of Bill C-393 (which you can learn more about in this recent Boingboing post). Although the fact that there are high costs is obviously true, a recent paper published in Biosocieties would suggest that the oft cited statistics, the ones always used to support this assertion for lobbying or public relations purposes, may in fact be over inflated.

Here, the authors, Donald W. Light and Rebecca Warburton look closely at where these numbers come from:

"The most widely cited figures (by government officials and the industry's trade association for its global news network) for the cost to discover and bring a new drug (defined as a 'new chemical entity' or 'new molecular entity'; not a reformulation or recombination of existing drugs) to market are US$802 million in 2000. This has been updated by 64 per cent to $1.32 billion in 2006."

From this paper, we basically learn that the primary source of these figures come from one particular study published in 2003 and done by Joseph DiMasi, Ronald Hansen, and Henry Grabowski at the Tufts Center for the Study of Drug Development in Boston, Massachusetts. In general, there are issues of bias in how such figures were calculated, and the Light and Warburton paper systematically looks at a number of variables that would suggest that the $802 million number, as well as subsequent numbers which extrapolate from this figure, are a gross over-estimate.

The paper is definitely worth a read, having a number of points that would suggest strong mistrust for these industry figures. Examples include:

1. High potential for bias in the data that was used in the 2003 study: this includes issues related to exclusivity of access to the data (i.e. we have numbers, but it's not clear what the numbers represent exactly since only the Tuft authors know), or to the sample set itself. i.e. the pharmaceutical industry appeared to have primary control over whether they would participate as well as what data was provided to the Tufts Center.

2. The figures do not include a number of special and substantial tax provisions for R&D work. i.e. Just because such tax measures help in our R&D costs, industry feels that it should not be included in these final figures.

3. About 50% of this $802 million figure is actually due to "profits foregone." In other words, as Light and Warburton succinctly describe, this equates to a 'You owe us for all our R&D costs, plus what we would have made had we not undertaken the project in the first place.' Although this is obviously a factor for businesses to take into account, the authors ask whether this is really an appropriate way to calculate figures used to lobby for government protected pricing? Nevermind the fact, that one could argue that R&D costs are a necessity when innovation is key element for an industry.

4. That trial costs as well as time estimates are inflated. For trial costs, a number of different discrepancies occur. For example, how much exactly does it cost for human test subjects during clinical trials:

The DiMasi team refers rather opaquely to a complicated set of steps taken to arrive at the mean cost per trial and per subject. The resulting figure of $23 572 per subject is six times the average cost per subject of $3861 reported by the National Institutes of Health for 1993, at the costly (later) end of the DiMasi period (1983 to 1994)

Related to this, is the estimate of how long does the R&D process take, since longer time spans obviously equate to higher costs. Again, these estimates appear to be greatly exaggerated:

The $802 million estimate is based on 52 months for preclinical research, 72 months for trials and 18 months for regulatory review, a total of 142 months or 11.8 years (DiMasi et al, 2003a, pp. 164-166). Maximizing the length of time not only dramatizes how long and hard companies work to discover and develop a new medicine, but also maximizes the multiplication of profits foregone. Long development times are a major reason given for needing high prices. These figures, however, do not square with the lengths for trials actually reported by companies to the US FDA in the Federal Register. Trial length declined from almost 8 years for trials started in 1985 to less than 3 years for trials started in 1995 (Keyhani et al, 2006). Regulatory review times dropped from 2! years to less than a year. Thus, for medicines that started testing in 1995, total trial and review time was down to less than 4 years in the United States and even less in Europe.

Anyway, it's causing quite a stir, and the Tufts Center for the Study of Drug Development have already issued a terse press release in rebuttal. In any event, it's good reason to be more informed in such matters, because it has implications in the wider scheme of things - such as how your health dollars are spent, and also how to make policy more effective when dealing with global health issues.

LINK: Demythologizing the high costs of pharmaceutical research, BioSocieties (2011) 6, p34-50 (Note that there is full text access to this article for the month of March 2011, or you can find the article by typing "Demythologizing the high costs" into a search engine).


  1. I never post here, but after seeing this, I just had to.
    The Light article is very biased and disingenuous – they have an agenda that’s obvious from the tone of their text. For a realistic rebuttal, check out:

    There are many other realistic, balanced rebuttals to the Light article out there as well.
    I’ve worked in the the biotech and drug industry (10 years) and am currently an academic. I’ve been on both sides of the fence and have a realistic perspective on how much it costs and how tough it is to discover and develop new medicines. The Light/Warburton figure is WAY off.

  2. Light and Warburton present, in Derek Lowe’s words, “a case for the prosecution, not a dispassionate analysis.”

    I recommend his inital post and his followup (discussing the way that Light and Warburton ignore the cost of the ‘R’ part of R&D) to readers wanting to get the view from a scientist in the trenches.

    LifeSciVC has also published an interesting attempt at a crowdsourced estimate of the cost of drug development.

    As a researcher in another field, I’ve often seen Dilbert’s boss perpetrate the sort of analysis that Light and Warburton present. It usually takes the form of “we just need to FOCUS. And PLAN. If we don’t waste our time pursuing all those fruitless experiments, we’ll get things done for a few per cent of the cost!” The problem, of course, is that you don’t know what an experiment will tell you until you run it. “What? You’re a SCIENTIST, we pay you to know these things! Are you telling me you’re incompetent?”

    Yes, managers are that stupid. And that good at selling their stupidity to the public.

    I’m sorry, the process is expensive. If the public decides that it’s too expensive, and is willing to accept the burden of death and disease that comes with abandoning it, so be it.

    I now await those who will denounce me as a stooge for Big Pharma. And those who will leap in loudly blaming the victims and ignoring the fact that healthy living postpones the diseases “caused” by lifestyle by only a few years.

    1. Await no more – you are a stooge for big pharma – why is it they “rush” to market drugs that only later are found to be worthless – or worse,deadly – why is it that another country (CDN for example) can ban a drug the FDA finds acceptable? Because the FDA is also in big pharmas pocket,plain and simple. We don’t pay more for our drugs because of FDA protection – we pay more because of their complicity.

  3. I hate this because it gives wings to those anti-vaccination and “coffee enemas cure cancer!” morons.

    1. Let’s be honest: those nuts will read that into anything. “New study reveals that eating too much food is bad for you.” “OMG! THE PHARMACOMPANIES ARE KILLING US FOR MONEY!”

      Anyone who doesn’t think that the interaction between pharmaceutical companies, the government, and the public is highly corrupt and biased against the public is a special kind of gullible.

      Fools who want to believe that the entire medical industry is evil and should be opposed or ignored (except when it shouldn’t be) are going to believe that no matter what. It’s typical conspiracy logic- lack of evidence is evidence of a conspiracy, and evidence is also evidence of a conspiracy.

  4. I would like to know, then, how come Sasha Shulgin managed to invent so much stuff in his back yard for cheap?

  5. I see that the comment I made a while ago fell into moderation hell, likely because I included too many links. So I’ll direct you to just one rebuttal from a scientist in the trenches: Derek Lowe’s blog. Be sure to read the followup article, and the crowdsourced cost estimator to which it links.

  6. I work for a pharmaceutical company, doing R&D. I am not 100% gung-ho about this industry as a whole, I have a lot of problems with a lot of things about the biz.

    I don’t know all the financial data, but I can tell you this: pharmaceutical R&D is very expensive indeed, and there’s no realistic way of making it cheap by any means.

    The main reason is that everything has to be just right for a brand new chemical to go from initial inception all the way to a prescribed product. It has to be safe, effective, and profitable. I’ve been told that only one in 1000 products that are developed actually make it to market.

    The materials are expensive, the instrumentation is expensive, the real estate is expensive, the electric bill is expensive, animal studies (yecch) are expensive, human clinical trials are VERY expensive, and above all the labor is expensive; the people who are any good at making this all work are well-educated and experienced. You could hire cheaper labor, and you will end up throwing even more money away.

    Anyways, my 2 cents.

    1. Putting my Dilbert’s Boss pointy-hair on: Major Buzzkill, the whole problem with you scientists is that you insist on carrying out these experiments on compounds that don’t work! Just pick the ones that work in advance so that you don’t waste your time in the lab!

      [Uhm, there’s no way to know in advance what an experiment will say without running it…]

      But you’re a scientist! We pay you to know these things! Are you telling me you’re incompetent?

      [Science works by experimentation, that’s how we come to know…]

      You scientists! No business sense at all. You’re fired!

    2. Major Buzzkill: Commander Speaks conveys his respects for the quality and effectiveness of your time-on-target response, and to ask whether you would like meet and discuss sharing information, methods and resources to enhance mission effectiveness.

      Also, he wants to know if you have any good drugs.

  7. This reminds me of how AT&T, when a monopoly, would spend money like mad to justify their annual rate increases. This seems like a justification for their business model. Big Pharma is following the same business logic that record companies did. Why develop something new when you can repackage the old product?, your costs are minimal. It is what happens when your vision of your business is reduced to 30 day cycles of the stock market. The record companies paid a price, they were so in love with repackaging vinyl to cd’s that they failed to develop new artists and were so happy with the profit they couldn’t see any less money coming in when mp3’s were developed. Pharma has become addicted to repackaging drug’s (already approved) for ailments they are not necessarily suited for and is developing few new drugs. The government is stepping into this void to develop new drugs because they are so unwilling to do so. Is there a price to pay?, maybe, but considering the government’s pro big business welfare I doubt it.

  8. Given the limitations of the data for the 2003 paper, I find the figures questionable. Is the $802 million dollar figure the cost of the most expensive new drugs brought to market or is this a mean value disclosed by the company. I think that this is a legitimate question with out more companies participating in the study and more transparency in the data released.

    I would also like to see the R&D costs juxtaposed to the costs of advertising pharmaceuticals including the revenue brought in through the advertising streams and the health and monetary costs associated with supplying a larger population,than justified by the clinical trials, with drugs that they do not necessarily need.

  9. This actually reminds me of when I was a wee undergrad determined to go into research. One of my professors (who after months of knowing well) told me about why I should reconsider my direction. When you go into research, you’re basically given two main options: hope to find funding for your research by the gov/a university or go private and find a foundation or company. The problem with the latter is that you’re pressured to find certain results. Not directly stated, of course, but you’ll get lavish (read, well equipped labs with lots of fun extras in vacation areas, paid time off incentives) “gifts” bestowed upon you while always being reminded that -this- company is what is giving you this lifestyle and that the research is in the interest of the company.
    My teacher once worked for a company that was doing research on a drug’s side effects pre-fda approval. I believe he said it was in the early 90’s when all this happened. His team had a lab with great equipment, but they located him near a scenic bay with his crew getting full room and board. They got huge company discounts at restaurants, the lab was attached to a state of the art outfitted gym. It was never stated that he had to get certain results, but he would get calls asking how the research was going and how they hoped that it would turn out well lest the company suffer financial losses and become unable to support ‘members’.
    Of course, there are still problems with publicly funded research, too but seeing what you’ve written here about research expenses being disgustingly high isn’t terribly surprising. Lots of major companies take it into consideration when they plan these things out.

  10. Sorry, the Slate piece is a crock.
    I’ll agree that the discount rate used in the study seems optimistic, but then again that study was published during a boom, not a bust. If you do read the slate article, you owe it to yourself to find out how sloppy and biased their analysis was. Derek Lowe’s comments about it at In The Pipeline back on Mar. 7 is a good start. It’s a much better rebuttal than Tuft’s own for several reasons, not the least of which is the lack of words like “operationalize”.

    I’ll let Derek sum up:

    “If the cost of R&D for a new drug really were $43 million, as Light and Warburton would have it, and the financial and tax advantages so great, why isn’t everyone pouring money into the drug industry? Why aren’t VC firms lining up to get in on this sweet deal? I mean, $43 million for a drug, you should be able to raise that pretty easily, even in this climate – and then you just stand back as the money gushes into the sky. Don’t you?

    Why are drug approval rates so flat (or worse?) Why all the layoffs? Why all the doom and gloom? We’re apparently doing great, and we never even knew.”

  11. You can get some good ballpark estimates by looking as the financial reporting of specific drugs for specific companies. I’ve never seen an FDA regulated drug come to market for less than $100mm. Most are on the order of $250mm, but $1bln is not unreasonable for a drug that requires many, large clinical studies.

  12. Some bizarre things included in these costs:

    – long-term contracts on forests to extract some special chemical
    – building of new plants (~200million) that aren’t strictly necessary
    – forgone profit calculations starting from when a university discovered the drug’s potential 20 years ago even though the drug company only acquired the rights a few years ago.
    – studies for marketing purposes not FDA approval

  13. I think Anon’s comment about the old monopoly days of AT&T are worth unpacking. The biggest costs for drug development are typically in the phase II trials that measure large-popultion efficacy. If the drugs that you’re developing are just repackaging or minor variation on existing stuff, it’s going to take huge studies (or multiple studies the details of whose submissions you can cherrypick, despite the completeness rules) to demonstrate the superiority of the new formulation over the existing ones. So precisely the drugs that make the smallest advances in human health cost the most to bring to market.

    If they cost so much to bring to market, why develop such incrementally useful drugs? For the same reason that movie studos bring out sequels to their big hits. Indeed, with patent protection, it’s like making a sequel that everyone who watched the original is compelled by law to go see.

    1. “[W]hy develop such incrementally useful drugs?”

      How about “different side effects profiles?” If you’re allergic to a first-line drug, or find that it’s ineffective, it’s really essential to have something to fall back on. Even if “mechanism of action” is considered to be the same, and the drugs are from the “same chemical family,” some drugs are simply ineffective or poorly tolerated in some patients. Which brings us to:

      “If the drugs that you’re developing are just repackaging or minor variation on existing stuff, it’s going to take huge studies (or multiple studies the details of whose submissions you can cherrypick, despite the completeness rules) to demonstrate the superiority of the new formulation over the existing ones.”

      Which is exactly why demonstrating equivalence, not superiority, is almost always enough. There is value in diversity. The law recognizes the value. (You can even do a study of drug B “in the population of patients intolerant of drug A” if you’re willing to accept the marketing restrictions that will come with it.)

  14. @ anon#8: Big pharma spent roughly $60 billion on R&D in 2010 and only garnered 21 approvals. Pharma is chock full of problems, but willingness to pay for R&D on new drugs has not been one of them.
    “Why develop something new when you can repackage the old product?” A: you do both. If there’s still money left on the table, pharma will reformulate an old drug to wring out that last bit of market share. If, as you point out, it’s not that expensive to do, then it’s not diverting money from new drug development.

    @#2 Kevin:I missed your comment (dell touchpad scrolling is like using a frog to navigate: jumpy) before making my own – well put.

  15. I have taken part in the ongoing commentary both at Derek Lowe’s ‘In the Pipeline’ blog (any chance of getting him as a guest blogger, boingboingers?) and on the Slate article Derek refers to, where Light and Warburton offer a defense of their paper (and where Light also attempts to shift a few copies of his book …).

    L&W seem to have confused Investigational New Drug applications, New Drug Applications, and Abbreviated New Drug Applications, all of which have differing safety/efficacy and documentation requirements. In short, ANDAs have a much lower requirement for clinical trials, being essentially an application to manufacture a generic drug based on one coming off-patent.

    Their use of median values is inherently flawed for a system where the majority of drug discovery projects never progress beyond the pre-clinical phase, with a failure rate that tapers off (but whose costs increase) towards Phase III trials (and even post-market withdrawals, something that L&W are silent on). This creates a cumulative cost distribution over time with a long front tail, spiking sharply upwards through clinical trials towards the approval/post-approval phase. To pick the median value from that highly skewed distribution is meaningless.

    It’s clear that neither Light nor Warburton really understand the drug discovery process. In particular they quote Marcia Angell verbatim on drug discovery, highlighting a key misunderstanding:

    the critical step in ‘discovering’ a new drug is understanding how the disease works and finding one or two good targets of vulnerability in the defences of a disease for intervention

    Generally speaking, this is not how drug discovery proceeds. Only very rarely is the mechanism of a disease understood, let alone the mechanism of action of the drug/target complex. Binding studies and assays in the lab might help elucidate the active sites on a protein that a ligand binds too, but only in as far as that artificial ligand blocks the natural ligand (as an antagonist) or stimulates the site better than the natural ligand (as an agonist) can that drug’s mechanism be said to be understood. And, in computational chemistry, anyone who has worked with more than one molecular modelling tool can understand the frustration of attempting to explain why two pieces of docking software, when presented with novel protein and ligand models, might come up with two entirely different modes of action or even different binding sites … (Comparison of Several Molecular Docking Programs: Pose Prediction and Virtual Screening Accuracy, Cross et al, J. Chem. Inf. Model., 2009, 49 (6), pp 1455–1474). However, if you can reproduce it, you don’t have to explain its mechanism of action – and that is actually a position supported, reluctantly, by the FDA.

    For what it’s worth, here’s DiMasi’s rebuttal:

    In their commentary, Light and Warburton restate arguments about the methods and data used for our R&D cost study that they have had published elsewhere. We have thoroughly rebutted each and every one of their prior claims in two published responses that Light and Warburton do not cite in their current paper (DiMasi et al., 2005a, 2005b). In the current piece, Light and Warburton also attempt to operationalize their criticisms by making adjustments to our published estimates that purport to demonstrate that R&D costs are much lower than we estimated. They make five such adjustments, all of which are erroneous—- they inappropriately mix median values reported for individual drugs with what are mean values for the costs of clinical failures and preclinical fixed costs, and for which the concept of a median has no meaning; they misconstrue the nature of the corporate income tax and incorrectly consider manufacturing tax credits; they use discount rates that are meant for other contexts but that are inappropriate here; they treat line extension approvals as separate and independent units of observation alongside their original approvals; and they grossly misstate the meaning of and misuse figures in our paper on industry-reported data on expenditures on self-originated drugs, licensed-in drugs, and already-approved drugs. Detailed discussions of these issues can be found in our above-mentioned rebuttals. In short, every one of Light and Warbuton’s adjustments are invalid. Furthermore, two peer-reviewed papers by current and former FTC economists, also not cited by Light and Warburton, validate our work using other methods and public data (Adams and Brantner, 2006, 2010). They find that R&D costs are likely as high or higher than our estimates

  16. It’s also pretty easy to debunk the Light/Warburton analysis just by reading a newspaper occasionally. If you could really discover and bring a new drug to market for $40-60M, as Light/Warbuton claim, and since a new drug is worth about $1B/yr in revenue, then drug discovery would be incredibly profitable. In fact, that’s what we would all be doing with all of our money, and the existing pharmaceutical companies would be expanding R&D like gangbusters. Needless to say, this is not happening. The industry has been in an incredible decline for at least a decade, laying off R&D staff and merging shrinking businesses. Since the Tufts numbers only measure how much existing drugs cost to discover, they may in fact be too optimistic to apply to new drug discovery.

  17. I am a researcher at a mid-level cancer research University.

    What bothers me the most about this argument is that it discards a very important fact. We’ve already paid for these drugs once.

    My lab, and the 40 other labs near me all operate using government grants. Money granted by taxpayers that probably totals about 15-20 million a year just for our tiny corner. I would say that these labs generate maybe one good new drug every 5-10 years. You as a tax-payer just paid 100-200 million for that new drug. These drugs are essentially given as a gift to the drug companies in return for a kickback to the university. The drug company generally does the phase 2 trials (probably on the order of another 200 million dollars).

    You as a taxpayer are now offered the privilege of paying for the research on these drugs a second time, in addition to already having gifted the basic research to the drug companies.

  18. The component of their critique that accounts for the majority of the discrepancy, “profits foregone”, makes no sense from a business perspective. Say a drug company spends $6 billion for R&D in a year, and develops 6 new compounds make it past the pre-clinical stage, and one is approved for use by the general public. Light and Warburton argue that pricing the one successful drug should only reflect its development costs, and not the failures? Any business employing this strategy would cease to exist.

    I’m highly skeptical of their skepticism. Anyone who follows the sector in the financial markets can tell you aggregate Pharma R&D investment in the U.S. alone is huge (around $80 billion) given the small number of approved new molecular entities each year (like 20 -30). The pharma sector has had more layoffs than any other in the U.S. since the start of the recession in 2008. It’s hard to believe the sector would be in so much trouble if R&D costs were really that low.

  19. Not surprised, as this is basically the same kind of shit that big media is pulling time after time…

  20. Consider this. I worked for 15 years at a pharma research site in the UK. The site itself was opened in 1984, and closed in 2006 – so around 22 years. The cost of building the site was around $200MM. The annual running cost of the site at the end was around $50MM. Let’s assume that the average annual running cost of the site over the time it was open was $25MM. Over 20 years, the total investment made in the site was to a first approximation $700MM. During that time, that investment resulted in two marketed drugs. And, the site was a discovery site only, so the cost of doing all the clinical trials and safety studies is not included in the $700MM. The average cost of *discovering* those two compounds, just in discovery, was then $350MM.

    Of course, the argument can be made that the site worked on 20 projects, and so the cost *per project* over that 20 year project was 1/20th of the total investment. In absolute terms, it’s true that the cost of each individual project was much less than $350MM. But based on an investment v. return analysis, the investment made by the parent company was $700MM, and the return was two marketed drugs.

Comments are closed.