Skip to main content

Drug Development

Idiocy On Drug Research Costs

Here’s a piece from the Center for Economic and Policy Research that claims to have the whole high-drug-price problem figured out. It’s “incredibly inefficient research”, just so you know. How does the CEPR know, you ask? They do a comparison of the costs of research (as provided by the Center for the Study of Drug Development at Tufts) with the costs reported by the Drugs for Neglected Diseases Initiative (DNDI):

On their tenth anniversary, DNDI produced a report describing some of their accomplishments. The figure below shows some of the highlights and their price tag and compares them to DiMasi’s estimate of what it costs the big pharmaceutical companies to develop a single drug.

As the figure shows, DNDI was able to develop ASAQ, a combination drug for treating Malaria, for $17 million. More than 250 million dosages have been distributed since 2007. It developed Fexinidazole, a new drug candidate and new chemical entity, intended to treat sleeping sickness, at a cost of $38 million. DNDI developed SSG&PM, a combination therapy for visceral leishmaniasis at a cost of $17 million. DNDI’s entire budget for its first 10 years of existence was $242 million, less than one-tenth of what DiMasi estimates it costs the pharmaceutical industry to develop a single new drug.

Now, the DNDI are good people, and they’re doing good work. The Center for Economic Policy Research, however, is being either ignorant or deceptive here. They go on to say that this isn’t quite an apples-to-apples comparison, because they don’t include the cost of capital (which the DiMasi/Tufts estimate certainly does). But that completely misses the most important part of the entire comparison, which is that the examples that the CEPR cites are from known drugs that others have already spend development money on. And they also neglect to mention that the DNDI works with the pharmaceutical industry to develop the drugs it has, and that these costs are not included in the figures shown above.

Let’s take them in order. ASAQ is the combination of artesunate and amodiaquine, a treatment that had been in use (as separate pills) for many years before the DNDI became involved. Artesunate is a soluble ester of artemisinin, part of the class of drugs that won its discoverer a Nobel last month. It has been around since 1987, and is a well-known antimalarial. Amodiaquine has been around since the 1940s, when it was discovered by Parke-Davis, as part of a large partnership to discover new antimalarials during World War II. The combination of the two into a single tablet (which is a real advance for dosing under the conditions of malaria treatment) was not the discovery of a new drug; it was the reformulation of two well-known ones, in a combination that was already being used. The later stages of development and production were turned over to Sanofi, one of the larger drug companies in the world, who had been working on the same combination. Sanofi handled the largest clinical trials, the regulatory filings around the world, and the launch of the drug. They agreed to make this new combination available without patent protection, and also agreed to pay the DNDI a fixed percentage of the sales of the branded version (Coarsucam) of the combination pill for several years.

Next up is fexinidazole, which the CEPR says is a “new drug candidate and a new chemical entity”. That is false. It was a new drug candidate back in the early 1980s, when it was discovered and partially developed by the drug company Hoechst. This effort was halted later in the decade, but the drug’s revival was another partnership between DNDI and Sanofi (the company that Hoechst eventually became a part of over the years). The DNDI did excellent work by going back over a collection of nitroimidazole drug candidates and discovering that fexindazole was worth developing further, but they were looking over the products of prior drug research, and they already knew the most promising chemical class to investigate (the nitroimidazoles).

The next drug combination mentioned is SSG/PM. That’s sodium stibogluconate (yep, an antimony compound), which was discovered as an effective therapy in 1945, and paromomycin, which was introduced by Parke-Davis in 1960. The combination of the two drugs was first explored systematically during the late 1980s. DNDI conducted the Phase III trial in Africa for registration of this combination therapy – its efficacy and its safety profile were both the same as with SSG monotherapy, but the course of treatment is 17 days versus 30, which is a significant advance.

But the Center for Economic and Policy Research is completely misinterpreting what DNDI has accomplished. Their press release is all about the cost of new drugs:

“. . .it is hard to escape the conclusion that DNDI research is far more efficient than patent supported research by the pharmaceutical industry. And no one has to struggle to come up with tens or hundreds of thousands of dollars to buy the drugs developed by DNDI. They are all available as low-cost generics.”

That’s because almost all of them were low-cost generics to start with. Comparing these projects with current drug research is disingenuous at best, and I’m not sure how ignorant I can believe that the CEPR people really are. Alzheimer’s, the various types of cancer, drug-resistant infections, pick your therapeutic area: these things are not going to be solved by reformulating compounds that were discovered decades ago. That can be a perfectly good approach in the neglected-tropical-disease field, as the DNDI and others have demonstrated, but it builds on the work of (and often depends on cooperation with) actual drug research companies.

Pretending otherwise is idiotic. Any organization that releases such stuff cannot be taken seriously when they talk about drug research, and you have to wonder how seriously they can be taken when they talk about anything else. If you want to read something intelligent about drug prices, read Jack Scannell.

30 comments on “Idiocy On Drug Research Costs”

  1. Emjeff says:

    But, Derek, they’re academics, who, as we all know, are experts in drug development #sarcasm

  2. cirby says:

    The part that got left out?

    Those drugs were developed for the African/Third World market. Try selling ASAQ in the US. Enjoy your many long conversations with the FDA.

    It’s a lot easier to develop and sell drugs when you bypass the massive US regulation structure.

  3. Andy II says:

    I could not agree more, Derek. Excellent points, though most of the people just pay attention to the sensational title of the article only and talked about what is going wrong in pharma companies compared to the “right development steps” that DNDI has been taking. Sad.
    Another talking point. What is the cost of developing a copy drug that a generic company would spend? I heard that the budget is in Legal/IP department at a generic company and a budget is $500,000 max (a phase 1, CMC, regulatory and IP challenge). And, as everyone knows, the generic company is not liable for adverse events that patients may experience when they take their generic drugs. The brand-name drug company has to take the responsibility ( Though I am not defending the exceedingly high price of drugs recently, I felt it is not fair as many publications show that some generic drugs are not the “same” as the RLDs.

  4. Pessinest says:

    The CEPR is a economics topic’d DC policy think tank
    —take a look at their about page
    Look them up on Wikipedia —

    The politics of Econ policy rarely aligns with any kind of scientific reality

  5. johnnyboy says:

    Heavens to Betsy, what a bunch of numbskulls. Anyone with a website can pretend to be an expert these days.

  6. Erebus says:

    @1 –
    They’re not even academics. The folks who wrote that article are probably professional “junior researchers” and interns. They’re likely just out of business school, and weren’t clever enough to secure entry-level positions with McKinsey, Bain, or Goldman Sachs.

    @2 –
    Spot on. Most of the costs associated with drug development can be attributed to an inefficient and mercurial bureaucracy. Sooner or later, the regulatory framework will be entirely overhauled, streamlined, and expedited. Hopefully sooner. The “cost of production” and “R&D” that Scannell keep mentioning don’t have much to do with the cost of synthesis, screening, or formulation — but have everything to do with the [extreme] cost and [utterly excessive] duration of the clinical trials which the FDA typically demands.

    Scannell suggests that “Prices will fall if buyers increase their power versus sellers. European countries should buy as a block.” He further suggests that there should be stronger incentives to commercialize new uses of old drugs, and that costs might go down if patent law is reformed.
    …But all of these things combined pale in comparison to regulatory overhaul. I’m surprised, and I think that it might even be rather disingenuous, that he didn’t once mention that.

  7. Old Timer says:

    I would take it even a step further and say that the reason clinical trials are so expensive is because of the ridiculously inefficient health-care system in the US. This circular problem always comes back to expensive hospitals.

  8. Hap says:

    1) If overregulation were hindering drug development, then you’d expect that less regulated countries would be getting more drugs and better health outcomes from them than the US or Europe. What evidence is there that that’s happening?

    2) Something like 50% of P3 failures are for efficacy – the drugs don’t work in larger trials (those more likely to look like the treatment population). Lowering the trial requirements would mean letting lots of cheaper drugs that don’t work in real life. I’m not sure why that’s a benefit to anyone other than drug companies.

    3) The regulation apparatus has built up in part because companies oversold drugs based on benefits they couldn’t deliver for the populations they sold them to. If your word is not good, then people are going to ask for more solid evidence to back it up. If getting that evidence costs lots more money than getting the previous burden of evidence, then they probably should have thought of that before making promises they couldn’t keep.

  9. Phil says:

    Anyone else find irony in the fact that the CEPR’s findings point to the Valeant/T*ring method of “drug discovery” as the most efficient?

    DNDI is, of course, using it for good rather than evil, but it’s the same general idea.

  10. Andy II says:

    @Phil: I respectfully disagree. Valeant/T*ring method is to identify and license an old drug yet with solid/stable sales figures (small enough other generic companies would take but large enough to make profit) that preferably a single manufacture so that they can have an exclusive supply agreement. They then raise the price. Ackman touted Pearson/Valeant business strategy here as you know already ( No new formulation or clinical study to add value to the old drug/compound.

    DNDI is to take an old drug/compounds and does a R&D work in collaboration with the originator of the drug/compound, and turn them into a useful drug in a new form for mainly for neglected diseases. I think it is more than “DNDI is, of course, using it for good rather than evil, but it’s the same general idea.”

    1. Phil says:

      @Andy II

      Fair points, and they are not the same business model by a long shot. I did not mean to imply that the DNDI has done anything but seize a win-win opportunity for patients and the companies who began developing the drugs in the first place.

      Where I find irony is this statement in the OA:
      “This massive discrepancy in costs should generate some interest in emulating the DNDI model for other drugs.”

      Sadly, they’ve already gotten what they wished for. I’m sure companies like T*ring and Valeant would LOVE to emulate the DNDI model in order to minimize R&D investment and maximize their profits. They’ve already hijacked the regulatory protections intended for neglected diseases, and once the low-hanging, fully FDA-approved fruit is gone, their next step will almost certainly be reformulation (just like ASAQ). I just hope they are both bankrupt by then.

  11. Ted says:

    The title says it all:

    “Incredibly Inefficient Research”

    Talk about being tarred with your own brush…


  12. qetzal says:

    Everything Hap said above.

    IMO, the biggest factor is that we are (collectively) so dismal at identifying which drugs will work, prior to large clinical studies.

    I’m sure we could find ways to make clinical development less expensive. But as long as 90% of candidates entering the clinic continue to fail, and as long as we also continue to demand that only proven drugs get approved, I don’t expect major improvements.

  13. cookingwithsolvents says:

    sodium stibogluconate…horrible compound for a horrible disease. I shuddered at both and am glad that at least something works, but still. wow.

  14. sk says:

    “Alzheimer’s, the various types of cancer, drug-resistant infections, pick your therapeutic area: these things are not going to be solved by reformulating compounds that were discovered decades ago.”

    Is there any medical reason for this?

    Under the current system, a fraction of R&D money is spent on discovering second uses and reformulations of off-patent drugs due to the fact that it is difficult or impossible to enforce a monopoly price. So this has nothing to do with medical efficacy and everything to do with the current pharmaceutical business model that relies on patent exclusivity over NCEs, and charging payers an exorbitant monopoly price.

    Perhaps this is highlighting that alternative incentives for drug development (e.g. public funding/prizes for successful clinical trials) could be a more efficient use of funds?

  15. Erebus says:


    Re: 1-
    Given that much of the developed world has adopted the US model — for e.g. Japan and much of Europe — I am not convinced that one should reasonably expect that at all. The rest of the world has neither the financial clout nor the cognitive capital to sustain innovative home-grown drug industries. And I don’t think that Pfizer’s going to rush to release drugs in Africa and South America before it releases them in the USA.
    …Besides, the current regulatory structure benefits these large drug companies most of all — it is an almost insurmountable barrier to entry for new firms, thus dramatically reducing competition & distorting the competitive environment.

    Re: 2-
    The way Phase III studies are run is absurd. New drug candidates shouldn’t be compared against the “gold standard” treatment — and, in any case, shouldn’t be expected to surpass it by a significant margin. Drug candidates should be compared to no treatment. If a drug candidate is safe, and if a drug candidate is more effective than no treatment, then it stands to reason that it should be allowed to see the light of day. The market finds its own uses for things, and clinical trials don’t tell us everything there is to know about these drugs.
    …Besides, the process is mercurial and unpredictable when it isn’t openly corrupt. At the very least, there should be a consistent, standardized, and much more predictable process.

    Re: 3-
    That’s a rather selective interpretation of history. The drug amendments of 1962 — the Kefauver Harris Amendment — were passed explicitly in response to thalidomide. That they have nothing to do with actual drug safety, but a heck of a lot to do with tightening the FDA’s grip on the pharmaceutical R&D, is rather ironic.

    I just saw this & thought it tangentially relevant. The guy has a point. Fortunately, federal government inefficiencies don’t typically affect industry — unless, of course, you’re in an FDA-regulated business.

    1. Jonathan says:

      There a number of industries in the USA affected by federal regulatory inefficiencies. One glaring example is the NRC (Nuclear Regulatory Commision). They charge a MASSIVE fee for approving new nuclear reactors, currently around 1 billion dollars, and charge it for every reactor, of any size – so all research into safe compact reactors that utilities companies aren’t paying billions of dollars to build must happen outside of the US. I’ve heard of ridiculous requirements MSHA places on mines where they just ignore all other options until the company caves and goes along with them, resulting in months or years of delays. Look the the 7 years that the Keystone XL pipeline languished without an approval under the current administration. Numerous times they ignored legally binding review timelines until the company gave up and pulled its application.
      I could go on, but you get the idea – it is one reason companies are doing more work off shore because on shore in the US they are subject to the whim of bureaucrats whenever they actually try to do something physical instead of, say, writing a computer program or app.

  16. Matthew Todd says:

    The ASAQ and fexinidazole stories were told at our recent open source pharma meeting by someone involved in both programs, Els Torreele (was at DNDi, now at Open Society Foundations). The stories were of interest there because they are examples of development in the absence of patent protection, and examples of effective public-private sector collaboration. Els was at pains to point out that the quoted costs hide the substantial work performed by pharma prior to the DNDi component (though the fexinidazole clinical work had to be re-done to meet modern standards, I believe) and, of course, that in these projects there was *zero attrition*.

  17. Hap says:

    1) You don’t have to build the infrastructure abroad – we built it and have been sending it abroad. If a foreign subsidiary develops a drug that works well but can’t be sold here, but can be sold somewhere else, then people there (and perhaps here) will buy it, and their health outcomes will differ from here.

    2) Letting drugs with greater than zero effect means lots of people will take drugs that don’t really work, which costs money and probably lives. The drugs will be cheaper (probably) and there’ll be more, but whether it costs less or gives better outcomes overall doesn’t seem obvious. (In addition, setting up trials versus zero effect seems impossible – who’s going to be the control? I don’t think doctors can give no treatment if it’s not the best practice, so you won’t have a control set.)

    3) I though thalidomide had something to do with safety, but in any case, I was thinking particularly of Vioxx, which was sold to people who (may) not have been able to benefit more than it harmed them. The size and costs of trials have increased significantly in that span, and while some is harder diseases requiring larger trials, some is probably because the FDA can’t trust drug companies to be honest with them, and so requires higher standards of evidence instead.

  18. Anthony says:

    Derek Lowe is a respected chemist, who just happens to derive most of his income from the patent-driven pharmaceutical system; Dean Baker is a noted economist, perhaps the earliest to warn, in 2002, of the existence of a US housing bubble and the potential damage it would cause to the US economy. He is also an expert on patent law, particularly as it pertains to the development of drugs and its effect on the efficiency of the pharmaceutical industry.

    Lowe quite obviously has a vested interest in defending the patent system, although he is at a loss to explain why it is superior to a public system, where all research is shared and open source. Yes, the patent system works, but with enormous waste, not to mention outright fraud in many cases. Few of the new drugs produced each year are the result of new chemical entitities, which Lowe, naturally, is quite interested in. But would these new entities go undiscovered without the patent incentive?

    Baker’s main point is that simply funding public agencies up front to perform this basic research would be more efficient, and obviously cheaper, than handing giant corporations a patent, which in effect is a licence to charge whatever the company wishes for a new drug (since normal market-driven pressures simply don’t exist in the patent system, which works like a massive subsidy to pharmaceutical corporations).

    Lowe can name-call, but the basic question still stands: in the absence of patent incentives, will important new drugs still be discovered? Answering in the negative says more about the ambitions of research scientists and popular bloggers than it does about economists fretting about spiralling drug costs.

    1. Perfect synopsis. ty.

  19. Erebus says:

    1. This does happen occasionally. I know that, for a time, Italy and Russia had pharmacopoeias which were very different from the USA’s. I don’t know how things look today. It might be interesting to examine a relatively affluent non-EU country and see what factors are at play. Know of any potential candidates?

    2. More drugs, cheaper drugs… that doesn’t sound bad to me. Hardly sounds worse than the alternative.

    You make a good point re: clinical trials. I think that, at least in certain cases, there should, however, be enough historical data to compare against, and that a control group may even be entirely unnecessary. (A grave heresy, I know.) Safety should be the FDA’s primary concern, in any case. I don’t believe that they should be the ultimate judges and arbiters of drug efficacy. That burden should rest on the far broader shoulders of the scientific and medical communities.

    …Besides, as I mentioned previously, the current drug approval process is a mess. Fixing it would solve a lot of the problems that currently beset the industry. We shouldn’t need to modify its essential structure — but surely it must be possible to streamline, expedite, cheapen, and standardize the process. Surely it must also be possible to prevent lobbying groups from influencing regulatory decision-makers. In any case, the decade-long, shockingly expensive, and unabashedly corrupt process we’ve got isn’t serving anybody’s interests very well.

    3. When you said “companies oversold drugs based on benefits they couldn’t deliver for the populations they sold them to”, I thought that you were referring to hucksterism, i.e. violations of the False Claims Act. Seems to have been a misunderstanding.
    I see what you’re saying about Vioxx. But increasing the size and scope of trials in response to the Vioxx scandal is exactly analogous to passing the drug amendments of 1962 in response to the Thalidomide scandal — it looks good, it makes it appear as though the government has the best interests of the people at heart, but it’ll ultimately do far more harm than good.

  20. Hap says:

    Lowe quite obviously has a vested interest in defending the patent system, although he is at a loss to explain why it is superior to a public system, where all research is shared and open source.

    You mean, other than being able to produce drugs? The ability to make drugs without making money from them has been around for a while, yet there aren’t a whole lot of drugs from it (and, as noted, the ones that do exist depended on for-profits to make and test). And, while it hasn’t been the public system’s purpose to develop drugs (but knowledge), they haven’t developed any, either. So it seems a bit premature to disassemble the patent and profit system for something else.

    If your system depends on people being something they aren’t (and haven’t been, consistently, since the beginning of time), it can probably be moved from the “viable economic system” column to the “fantasy” column.

    1. Anthony says:

      Patent protection is not an incentive to produce drugs; it is an incentive to market them. The system has never been particularly good at developing drugs that won’t be in general use, but it has been enormously effective at developing nearly udentical copies of already acailable competitor medications. The research and development side has become a sorry joke; if it hadn’t, this debate wouldn’t be necessary.

      The pharmaceutical industry is practically the only industry that is handed a government-granted right to sell its products thousands of times above their marginal cost of production. Yes, chemists don’t fret about these things, but economists do, particularly when such things are dragging down the economy of an entire country.

      1. matthew cryer says:

        bang! agin.

  21. Andy II says:

    There is a very encouraging regulatory system for antibacterials. In 2012, in response to the emerging drug resistant bacteria and few new antibiotics under development, FDA established QIDP (Qualified Infectious Disease Product) as a part of GAIN (Generating Antibiotic Incentives Now) Act that will extend the Hatch-Waxman provisions related to data exclusivity by 5 years while maintaining the current paradigm for an abbreviated NDA paragraph IV certification. ( Because of this “incentives,” more companies are now working on discovery/development of new antibiotics.
    Well, I still think that generic companies who go after lucrative RLD with a copy version should be accountable for any events that are caused by their drugs.
    All big pharma’s are now working on biosimilars while defending their own RLD biologics…. Something funny.

  22. Erebus says:

    What’s most interesting is when the FDA breaks its own rules. From today:

    “The pills are not approved by the Food and Drug Administration, but the agency has chosen not to take enforcement action because of the lack of alternatives.”

    …But surely that company is not allowed to market that product as a drug, with drug-like claims of efficacy? Odd, isn’t it?

    Aside: We like to criticize annoying companies like Valeant and T***ing, but they’re a product of our absurd regulatory system. Were the system saner — and, preferably, optimized — that sort of business model would not exist.

  23. Thomas says:

    There is the problem of testing for efficacy. We find many second uses, or differences, in broad usage of new drugs.
    So wouldn’t it be good to approve drug A which suppresses B in the body, but B is just correlated to but not the cause of disease C?
    Suppressing B might be helpful for some other purpose. Like we found a second use or two for aspirin. OTOH new APIs may be so selective that this is not likely to happen. Perhaps too high selectivity is the cause of things which just do not work?

Comments are closed.