Skip to main content

Business and Markets

A Grim Future? Here Are the Numbers.

Here’s a bracing look at the state of the pharma business, especially regarding R&D costs and return on investment. And let me warn you – it’s not a feel-good sort of article, but the figures are hard to refute. The author, Kelvin Stott, comes out swinging:

Here I apply a far simpler, much more robust methodology to calculate Pharma’s return on investment in R&D, which is based only on reliable and widely available high-level data on the industry’s actual historic P&L performance. This new analysis confirms the steady decline reported by others, but here I also explore the underlying drivers and make concrete projections, which suggest that the entire industry is on the brink of terminal decline.

He goes on to calculate the average return on investment (IRR) by adding up the compounded annual growth in the value of past R&D investments (as they contribute to profits, on an earnings-before-interest-and-tax basis, and adding in future R&D costs – opportunity costs, since investing in a new R&D program is one option out of many for that cash). He then has a formula for internal rate of return, based on an average investment period of 13 years per project. It’s true that projects spend more money in the latter part of that period, but there are more projects in the earlier years, so it roughly balances out.

Unfortunately, when you do this over time, you find that the IRR has been declining at a pretty steady rate. This is (as Stott shows) completely consistent with findings from Deloitte and BCG, and it also looks consistent with a McKinsey analysis from a few years back. I know what you’re thinking: “So if all of these consulting groups agree, can this conclusion actually be true?” Sadly, it probably is. The investment return on R&D has been decreasing, because the opportunities for big returns have been getting progressively harder to find, and the costs of exploiting them have been rising without interruption. This is what Bernard Munos and others have been pointing out as well; it’s a hard conclusion to avoid.

Stott’s line has the rate hitting zero in just a few years, and as he points out, there’s no reason for it to stop there. Zero IRR means that you’re breaking even, and as we all know, that’s not the bottom at all. You can go on to lose money, at ever more exciting levels, and head into negative IRR territory. In fact, if you use these figures to project profit-and-loss for the industry as a whole, Stott says that we may well be at the peak right now (see the article for the graphs). Everything from here on is down, according to this analysis.

Diminishing R&D productivity and return on investment leads to diminishing growth in sales. Eventually, growth turns negative and sales start to contract. Reduced sales then reduces the amount of money available to invest back into R&D, which causes sales growth to decline even further. And so on, until the industry is gone altogether.

Good morning to you, too. But as the article goes on to say, this doesn’t mean the end of the drug industry – just the drug industry as we know it. I’ve said something similar in recent talks I’ve given – I’ve told graduate students that I don’t know what the industry is going to look like in 20 years, but I’m sure that it won’t look just like it does today, because what we have going today is simply not sustainable. Stott makes the case that moving out of traditional small-molecule drug discovery looks like the only way out – cell-based therapies, immunotherapies, gene therapies, tissue engineering and so on are the current frontiers. And that doesn’t mean just scientific frontiers, although they are that, but business frontiers to continue to run drug companies as going concerns.

So although this article is a cup of cold water to the face, I really can’t disagree with its main conclusions, based on its premises. The only way out is to disprove some of those premises or come up with some new ones. As for the disproving part, you have things like success rates in the clinic, cost of R&D, number of good targets – we all know about the pressures on these factors. If something dramatic happens in these areas, this sort of analysis will have to look different, but if things continue as they are now (and as they have continued for the years up to now), then the picture laid out is the one that’s coming. The “find some new premises” part is what’s mentioned above – moving into totally new treatment modes and trying to make a go of them.

That’s why Stott’s graph looks a lot like the “peak oil” graphs from a few years ago. Those predictions didn’t come out as grimly as forecast, mainly because of new technologies like fracking (both to release more oil and to provide a new supply of natural gas for the energy market as well). We are going to have to break through into new technologies to survive this, too.

As medicinal chemists, we ignore this advice at our peril. Hoping that things are just going to be able to keep chugging along somehow in the small-molecule world has not worked out so great over the last twenty years or so, to put it delicately. And it’s not a good strategy for the future. Small molecule therapies definitely have a role to play, though, in these new treatment areas, and we had better be ready to pitch in, because those are the most likely areas for growth. And we could use some growth.


82 comments on “A Grim Future? Here Are the Numbers.”

  1. Stop being so negative! Provided that you keep property forecast index below 7 and don’t compromise ligand efficiency (best quantified by the seminal LELP metric), we will not be future-limited. It is clear that even undruggable targets will be tamed by coupling of enthalpically-driven binding with slow off-rate. Furthermore, apps based on game-changing techologies such as AI and Machine Learning, will disrupt pharma for the benefit of all humanity (and their pets). That’s right we’re also going use your phone to cure humanity’s dogs, cats, hamsters, goldfish and pythons.

    1. Derek Lowe says:

      If you can cure the guinea pigs, that’ll be something. Their first response seems to be to die.

      1. How ungrateful of the guinea pigs! It’s especially inconsiderate of them given that the compounds were clean in the hERG assay.

      2. Anonymous says:

        Penicillin and other antibiotics are toxic to guinea pigs because of their predominantly gram-positive intestinal flora. If Fleming / Florey / Chain had tested penicillin in guinea pigs, it might have killed further development.

    2. Sean Fhearsalach says:

      Does theis actually ually mean anything? If so, what?

      1. Mister B. says:

        It looks like a sample from a scientific bullsh*it generator ! 😉

        1. Bullshitting is an art form and it would be uncultured to use a generator to create it. Nevertheless people will try to convince you that getting property forecast below 7 is A Good Thing (a couple of carboxylates should do the trick) or that LELP is meaningful (even if it looks like it’s been taken form the pages of a Mary Shelley novel). The continually expanding menagerie of rules, guidelines and metrics seems to have overwhelmed the critical thinking skills of many a drug discovery scientist.

          Drug discovery gets tougher all the time and, if we make decisions based on flawed data analysis, those who fund drug discovery may conclude that the difficulties we face are of our own making. In drug discovery (as in science) it is vital to be able (and willing) to distinguish what we know from what we believe.

          1. Lawrence Wolfe says:

            Well and truly said – specifically the last line!!!!!!

  2. biotechtoreador says:

    An interesting article, and I suppose good news for chemical biologists (biological chemists?). I wonder if there’s any differentiation in IRR by size of company? Long term there won’t be, but it seems (without having any data to back this up) that maybe there’s a sweet spot for emerging companies with a time window of X years before/after approval of their first drug? I guess applying an ‘probability of success’ multiplier would likely decrease any benefit to that of the big pharmers…..

  3. Uncle Al says:

    Bell Labs versus Lucent Technology. Corporate R&D crash and burn suggests Human Resources hired safe drinking buddies and diversity, not mutant smart autists (ewww!).

    R&D failure includes mangerial oversight by the book. So it can’t be done, so what? теория решения изобретательских задач, teoriya resheniya izobretatelskikh zadach, TRIZ. Do it the “other” way (insubordination!). Flibanserin (Addyi) was FDA never meant to work. It was meant to shut them up with side effects, especially with alcohol.

      1. Uncle Al says:

        “That is why you fail.” Warmup exercise: Dissolve 30 wt-% copper phthalocyanine in Plexiglas, monomolecular straight up. I did it the “other” way, using metallurgy, to 500g of masterbatch dye. HR calls this “deviation from professional trajectory.” To criticize is to volunteer.

        1. me says:

          I’m not sure whether you’ve stepped off the deep end or you are a Markov-chain experiment from some nascent internet-spawned “uncle ai” that happens to have a passing interest in the pharmaceutical industry.

          Either way, please continue posting, your stuff is always fascinating to read!

  4. luysii says:

    A major reason drug discovery is so hard is that we’re trying to alter a system whose workings we only dimly understand. New players crop up all the time, microRNAs, ceRNAs, piwiRNAs, long noncoding RNAs. But it just got worse, as a player as old as molecular biology itself — RNA — has just been found doing something totally unexpected. We know proteins associate with RNA in the ribosome and in RNAase P, but here they are thought to keep RNA in its proper form. The new work shows that a 2,300 nucleotide long coding RNA binds to an enzyme (GOT2) increasing its activity. Just how widespread this sort of thing is, and how drugs affect it isn’t clear, but the RNA/enzyme interaction site and is quite well preserved across species. For more on this please see —

    1. luysii says:

      Sorry ! That’s a long NON coding (for protein) RNA ! !

  5. Marcin says:

    Regardless of the therapeutic modality (molecule or device or tissue or a combination) – I think big pharma (or in future – big companies developing and commercializing therapies in various forms) will be going away from the internal R&D altogether. To some degree it has been happening already with big players setting up essentially venture funds playing on small biotechs, besides obviously scooping up academic research based on the slave-labor-modality (grad students and post-docs) and with low-overhead costs (university lab is already there).
    Many of those small biotechs came to exist thanks to belief and drive of the laid off scientists working on a project that was “de-prioritized”. Pharmas used to shy away from outlicensing their shelved projects for fear of embarrasment once a small company develops it. Now I think they embrace it. Pharma-derived and other venture money will probably wisen up and will not invest into anything before the succesfull Phase I.
    Big companies will be machines for later-stage development, manufacturing and commercializations. Those are their strengths that benefit from their natural economies (forces) of scale…

    1. To this note, Science Business by Gary Pisano basically described about ten years ago this general scenario.

      Also, I’m not sure if external funding (eg, VC, Angel investing) in startups is part of Kelvin’s model. It seems that might tilt things more toward reduced rate of return for the sector overall?

  6. Magrinho says:

    “The drug industry as we know it” plants a false notion. It is a relatively young industry (compared to Medicine) that has never remained constant; therefore, “knowing” it means recognizing constant change/churn. Look at any 5-year window since 1950 and the tectonic plates have never stopped moving.

    The huge opportunities presented by chronic CV and CNS and other drugs are gone (safe, good enough, now generic) and never coming back. We are better at moving biologicals to clinical trials. Safety assessments have become increasingly articulated (for better or worse)… And on and on.

    It really has nothing to do with “the man”/shitty mgmt, biological models or any of the details. People working under terrible management discovered great drugs without knowing anything about the underlying biology. We make up the heroic narratives after the fact.

    1. HTSguy says:

      “The huge opportunities presented by chronic CV and CNS and other drugs are gone (safe, good enough, now generic) and never coming back.”
      Really??? I can think of a number of neuropsychiatric indications where the current drugs are neither safe nor nearly good enough.

      1. Magrinho says:

        Agreed – there are massive unmet medical needs in CNS.

        For example, I think current drugs are inadequate for psychosis and dementia. And the list goes on and on. But I believe that the opportunity to address those indications with drugs is GONE. I believe that other therapies will be better and easier to bring to market.

        The time from first synthesis of haloperidol to first dose in a psychotic patient was a few months. Yes, that was wrong/unethical/crazy but that was the “golden era” of CNS drug approvals. It is very hard to get anything through GLP tox these days.

        1. Luysii says:

          Well it wasn’t an entirely golden age of drug approval. American neurologists of my generation drooled over the reports of the miraculous effects of L-DOPA on Parkinsonism coming out of Europe. This was in the mid to late 60s. The FDA wouldn’t accept the European studies, and demanded its own, with the result that L-DOPA was finally released in the states 9/1972. I actually saw a few people leave wheelchairs.

  7. TroyBoy says:

    But I worry that pessimism will keep people from even trying to develop new drugs–especially this article coming from the Portfolio Director at Novartis. There are a thousand reasons why a program/drug/experiment won’t work and you can easily talk yourself out of moving forward. But I think everyone needs to keep plugging along. One way to guarantee you won’t find a drug is to stop trying.

    When things get rough and I want to quit, I remind myself of two quotes (one of which was brought to my attention by you, Derek, on this blog):

    (1) “it’s also true that just about every successful drug faced some legitimate existential crisis along the way — at some point during its development, there was a plausible reason to kill the program, and someone had to fight like hell to keep it going”

    (2) “Wenn ich wüsste, dass morgen die Welt unterginge, würde ich heute noch ein Apfelbäumchen pflanzen”/Rough translation: If I knew that the world would end tomorrow, I would still plant an apple tree today.

    So don’t lose your optimism! Find new technologies and innovate!

    1. Design Monkey says:

      Sure. If the company’s bean counters will give the money to continue. If not – then not.

    2. Anon says:

      “So don’t lose your optimism! Find new technologies and innovate!”

      Wasn’t that his main point? That we need more radical innovation, rather than continuing as is?

      1. Kelvin says:

        Thank you!

        Hopefully this analysis shows that the *last* thing we want to do is to “keep plugging along” as we have been, because the consequences of doing that are now clear.

  8. John Tucker says:

    The conclusions may or may not be correct, but the model itself is flawed.

    If you invest $100M in return for a $200M payout 13 years later, your IRR is 5.5%. But if you invest $100M/13 every year for 13 years, your IRR is 9.5%.

    The model of attributing one year’s EBIT to a single year’s R&D expense implies a much larger investment of capital than is actually involved.

    1. Barry says:

      but applying the same flawed model consistently over time can still show a decline in ROI

    2. Calvin says:

      I’m glad somebody pointed that out.

      What lost me was the last part when we had yet more drivel around “disruptive innovation”. Has anybody who utters those words ever taken the time to actually read Clayton Christensen’s books and figure out what he’s actually talking about. Somehow I doubt it. While it’s clear that we’ve have a problem broadly in the industry the only problem in this analysis is that it lumps all companies into the same bucket. If you were to look at the mobile phone industry you’d probably see something similar because so many companies have made no money at all; it’s dominates by a small number of players. I’m sure the IRR for Apple looks somewhat different to HTC….
      So the big issue for me, is what we do about it? I’m always nervous when somebody says that we should simply dump small molecules and move to mabs, gene therapy etc. I always take the view that each technology has it’s own pluses and minuses. There are some areas where mabs are better and probably always will be (cancer, inflammation) and it is these areas where cell therapies etc seem to have the advantage. But other areas like infection are the where small molecules have an advantage (HIV, HCV would still be massive killers without small molecules). So we should concentrate on what the disease is, what the patient needs and then decide on what type of product would work for patients. They are not all the same.
      Kelvin is a nice chap, but his one foray into drug discovery was not blessed with success so perhaps he’s still scarred by that. It’s a hard business. If it was really easy everybody would be doing it and making a ton of cash. But as Google etc have discovered, human biology is somewhat tricky. We should anticipate that it is hard. I’ve never bought into the “low hanging fruit” stuff. And while I’ll agree that as SoC becomes better, new drugs have a higher bar, I’m still pretty sure there are plenty of significant diseases where lots of people die, even ones we thought we’d dealt with; we’re a looooooong way from the end of disease….Still money to be made. Some companies will be successful and some will not be. Some will fold, even perhaps some big ones.
      So all in all yet another “doom and gloom” article from the business crowd. What are the bets on Novartis doing through another “disruptively innovative reduction in headcount” to get its IRR up. Asking for a friend….

    3. Anon says:

      “If you invest $100M in return for a $200M payout 13 years later, your IRR is 5.5%. But if you invest $100M/13 every year for 13 years, your IRR is 9.5%.”

      OK, but that’s not what he’s doing. He counts the single lump sum R&D investment at the *midpoint* of the R&D phase, not at the beginning. If you do that with your example then the IRR is much closer (10.4% vs 9.5%). Besides, he shows that even if you shift the lump sum by a couple of years either way from the midpoint, it doesn’t change the downward trend, nor the intercept through zero.

      “The model of attributing one year’s EBIT to a single year’s R&D expense implies a much larger investment of capital than is actually involved.”

      Why? Isn’t 1 year x EBIT equal to n years x EBIT/n for any n? And 1 year x R&D = k years x R&D/k for any k?

      1. Kelvin says:

        Exactly, thank you!

        Also, the IRR itself is not so relevant as a number. The key point is that:

        1. The past downward trend (in whatever you want to call this metric) is an almost perfect straight line.
        2. There is a clear theoretical principle and rationale why it should be a straight line – and should continue to be a straight line.
        3. Projecting this linear trend into the future has direct, unavoidable consequences.

  9. a says:

    When systems fail, Question your assumptions.

    How about the unstated one: “IRR and shareholder value is ultimately a more important metric than Human Health”.

    How about a new assumption: Capitalism as practiced is devouring itself, and humanity with it. Make some backup – or at least parallel – plans.

    1. anon says:

      There are some things you usually cannot say. Taboos, wrongthink. For example that human life is not invaluable.

      Let’s do a thought experiment. Suppose that a person is suffering from a terminal illness, but there is a treatment that has a good chance of curing them. The catch is, the treatment is extremely expensive. How expensive can a life-saving treatment get, before it is not longer worth it?
      Of course, if you are the person feeling the reaper breathing down your neck, the answer is probably extremely high. But chances are you do not have that much money, so someone else have to pay. How much is a society willing to pay, in order to save the life of one of their people? A thousand USD? A hundred thousand USD? A million? Fifity million? Hundreds of million? Would your government tax everyone for an additional 33%, to save a single life? If it would, would it be worth it, considering that such a drain would impact a lot of people negatively?

      When you answer these questions, you arrive at the value of a human life in a given society, expressed in terms of money. Treatments (drugs and otherwise) can only become so expensive (with development costs and the costs of failed dev projects factored in), before societies decide they are no longer worth it, and saving certain people is not worth it. If you found a way to eradicate HIV from a person with a treatment cost of say, a million USD per person, that would mean very little in Africa for example, and people would keep dying by the truckloads.

      I used the somewhat hypothetical example of an extremely expensive cure to avoid death for rhetorical reasons, but my argument stands for less radical cases too, just substitute “cost of life” with “cost of an improvement in quality of life”.
      While I have little love for pure capitalism and “profit above all” mentality, I believe the issue of “is he worth saving?” is more fundamental, and would persist in alternative systems, because it seems to be a fundamental issue of allocation of scarce resources.
      We may be approaching a point where most diseases could potentially become treatable, if not curable. However, I find it unlikely that we (as in mankind), will have the resources to provide such state of the art healthcare for every human being in the foreseeable future. Some people will have to be left with second-rate treatments at best. The question then becomes, how will it be decided who gets what? Who lives or dies?

      Currently, this is pretty much decided by money. If you get to live in a wealthy country and enjoy the support of society through a publicly funded healthcare system (or you have good insurance, or lots of money) you are in the clear. If you live in Africa, chances are you are screwed, no *mabs for your cancer! This is probably not fair or right, but the chances of change seem rather slim.

      Hope my ramblings were not too incoherent 😉

      1. AVS-600 says:

        This is a good way of thinking about things as a starting point, but the thing to keep in mind here is that for small molecule drugs, while the fixed cost of development is very high, the marginal cost is dirt cheap. So the question isn’t really “how worthwhile is it to cure this one person’s disease”, it’s “how worthwhile is it to treat every person that suffers from this disease for as long as our civilization lasts”.

        If you operate under the assumption that our civilization will last for quite a long time into the future (you can insert a joke about current events here if you like, but I’ll refrain), then there is a VERY high threshold for when it’s worthwhile to develop a drug from a utilitarian perspective; if the high cost and uncertain nature of research prevent these drugs from being developed, that represents a genuine market failure rather than an “appropriate” diversion of resources into other endeavors.

  10. Hap says:

    1) The shift to biologics seems to be to avoid elongate effective patent lifetimes and thus expected revenue, but I don’t know how sustainable it will be unless it achieves a lot. It also means the complaints about drug pricing and contribution to health care costs will likely be closer to reality – the fraction of health care costs from drugs will likely increase, and the price tags will provide very nice lightning rods.

    2) If R+D doesn’t gain any benefit from the knowledge it’s found and the experience its practitioners have, then aren’t we pretty much FUBARed? No one’s going to pay for the experience of researchers if it doesn’t make them money, at the price point that is attractive to employers, few people are going to spend lots of money and time educating themselves for short careers with poor ROI.

    1. Barry says:

      Companies expect longer exclusivity for Biologics than for small-molecules because it has been so much harder for a generic to demonstrate equivalence once the patent protection has run out. That barrier is lower with time, but I don’t see it getting to zero soon.

      1. Hap says:

        I was not coherent – I meant that biologics have a longer (effective) patent life, for the reasons you said. Given that pseudogenerics when they do come are likely to cost almost as much as the originals, increasing their discovery will probably accelerate a cost reckoning, unless they do a lot (something that small molecules can’t).

  11. b says:

    I’m sorry, but my bs detector is going off pretty loudly with this guy.

    1. Emjeff says:

      OK, I read this guy’s LinkedIn post, and have come to the conclusion that he is a crack-pot.

      1. Derek Lowe says:

        Break out the reasons why – let’s put ’em out in the open.

        1. Old Timer says:

          All right, I will take a crack at this, but I’ll keep it short since there are a lot valid points already made in the comments section of his LinkedIn post. Just for reference, I currently hold three federal grants, regularly review for PRF and NSF, and have been on numerous faculty search committees. In this context, I would say I’ve written 30+ grants and evaluated 500-800 proposals over the last 10 years. I’ve also worked in industry and started three companies (only one is still in operation).

          Problems 1-5: No details (except drug-popping toaster). It doesn’t matter how crazy or straightforward an idea sounds, you need to provide some information as to how it works. By not doing so, you demonstrate that you either haven’t thought about it, or don’t care enough about your audience. Either way, expect a chilly response.
          Problem 6: Unsubstantiated claims with hints of delusion. When the main premise runs counter to 100 years of known practice, the onus to explain is great.
          Problem 7: Poor, convoluted writing. If you’re pitching a scientific idea that you claim to have spent 12 months developing, don’t ask rhetorical questions or use two-word sentences. It is the writer’s job to explain a position (see Problems 1-5). This would not pass as an undergrad writing assignment. Consequently, it likely offended/confused the reader.
          Problem 8: Questionable logic. Instead of explaining why he’s right, Kelvin asks everyone else to prove him wrong. Since this is a common approach by conspiracy theorists, time-travel supporters, and alien hunters, it puts him in an unseemly camp.

          For the record, I have not disproven that JFK was killed by time-traveling aliens, so that thesis can still be thrown in my face.

          1. Kelvin says:

            Hi Old Timer – these are all very valid and reasonable *human* explanations for the response I got, and to be honest I was expecting no different as we live in a human world, but it does beg the question: Was it a *rationale* response?

            Even so, you will be pleased to hear that I am very quickly adapting my pitch to the human world we live in, in line with all your good points. 🙂

          2. Old Timer says:

            Let me put it this way: If you don’t have the respect to write to someone like a human, why should you expect anything but an inhumane response? Consequently, it seemed to be a perfectly rational response to your letter.

    2. Emjeff says:

      God help us, he has something called a “mini-MBA”. Derek, I think you may have been taken in by a nut-job.

      1. MoMo says:

        Ye Gods! He’s invented a machine the size of a toaster that can cure 1000’s of diseases just by adding cells! Surely Novartis has one now! I am firing all my chemists and buying one. It must be true as its on LinkedIn!

        Kelvin, if you are reading this- where can I get one of these machines? Will I have to buy one of those European to US voltage converters?

        1. Kelvin says:

          I’m here! 🙂

          Interestingly, not one of the ca. 10 scientists who have actually read my proposal (who are, by the way, very qualified and experienced in drug discovery) have been able to identify any reason why the idea can’t work in principle, or in practice. Their opinions vary from “difficult and/or unlikey, but certainly not impossible”, to “probably not for all, or even most diseases, but definitely worth a shot”.

          If an idea could potentially generate $billions in value, but cost only a couple of $million to test the concept, then the expected return on investment is huge even if the likelihood of success is greater than just 1%.

          And for those scientists who somehow know that the idea cannot work (PoS = 0%) without even knowing anything about it, well that’s not science. It’s called blind faith.

          OK, granted, the approach and positioning of the idea made me look like a nut-job, but that was done deliberately to demonstrate that too many people judge based on style vs genuine interest in asking the right questions to explore true potential. Ironically, your post only demonstrates my point all too clearly.

          But in any case, can you give any fundamental reason why such an idea *cannot* work in principle, without knowing anything about the idea, and without making any false, dogmatic assumptions? I bet not.

          1. b says:

            If that is truly the case, then you should have absolutely no trouble getting some VC money, doing the work and saving the world while generating billions of dollars for yourself and shareholders. The fact that you’d rather use it as an example of why Big Pharma is so stuck in their ways illustrates that indeed, it is snake oil and you have no desire to pursue such a thing in earnest. Better to analyze and consult from 10000 feet. Don’t want to get your hands dirty. The message you are trying to deliver is completely negated by your behavior.

          2. MoMo says:

            Thanks for cogent response Kelvin. I would of listened to you before calling you a nut-job. But futurists like you have no real clue on the drug discovery process from conception of chemistry to delivering a clinical candidate to making sure the business side doesn’t derail it through ignorance and greed or lack of knowledge and arrogance. You don’t actually think a toaster device is going to alleviate mankind’s ills do you? That’s the part where one thinks “nut-job”.

            Its good to shake the tree and you have no doubt done that- but that’s when the nuts fall usually.

      2. Charles says:

        The email reads like something a nutjob would write too. He is clearly deluded that anyone would give a job. What was with that weird salesmanship about a device the size of a toaster, it reads like something a schizophrenic or snake oil salesman would write. This is not worthy of serious consideration and the executive he sent it too clearly thought the same.

        1. Kelvin says:

          Ditto my response above, hopefully this will explain.

  12. Bernard Munos says:

    This is provoking, and I command Kelvin Stott for sharing his analysis. His work and comments, are often insightful. However, given the sweeping conclusions, it’s important to make sure that they rest on sound assumptions. There is one in particular that I am uncomfortable with, and it is that costs and sales (and hence returns) follow an exponential distribution. It may sound trivial, but it is not. The exponential distribution is not “fat-tailed”, which means that it does not accurately model the blockbuster events that drive the pharma industry (for both costs and revenues).

    Blockbusters are what makes pharma viable, and gives it both its luster and extreme risk profile. Most drugs are mediocre sellers which never recover their costs. The business model is only rescued by the random occurrence of the Prozacs, Lipitors and Humiras. More importantly, blockbusters cannot statistically exist if drug sales are exponentially (or normally) distributed.

    Scientists still don’t quite agree as to the exact distribution that underpins the industry, but they generally converge around the stable (or Lévy-stable) and lognormal distributions. For the sake of rigor, the analysis ought to be re-done by sampling from these distributions (which is not trivial) instead of the exponential one. Conversely, if there is evidence that the exponential distribution adequately models the industry’s sales and costs, it must be brought into the discussion.

    1. Kelvin says:

      Hi Bernard, you raise a good point / question.

      As it happens, I also ran the IRR prioritization simulation with uniform, normal, and lognormal distributions of investments and returns (which I’d be happy to share). All simulations gave a very similar downward linear trend in the midsection, regardless of how they are distributed.

      1. Kelvin says:

        PS. I have just added some of these simulations to my blog.

  13. huh? says:

    I read this post earlier and saw at least a couple big assumptions that are often stated, but just not true.

    1. That drugs are so good now, its hard to make them better (to paraphrase). This is total BS. There’s plenty of patients with no good drugs to take. PLENTY. AD, parkisons, Bipolar, depression, pain, the list goes on forever.

    2. The oil mining analogy is silly. Its a good analogy, assuming NO INNOVATION. But that’s the whole point, to discover new things. Does this guy not realize the point of research is to discover new things? Granted, you never know when the big innovations will occur, and big pharma has problem’s innovating, but still…We aren’t in the burger business, we invent stuff.

    1. Kelvin says:

      “1. There’s plenty of patients with no good drugs to take. PLENTY. AD, parkisons, Bipolar, depression, pain, the list goes on forever.”

      Yes, and the oil drilling image in my article acknowledges this, showing that these particular diseases are still completely untapped, but the costs and complexity have been too great to generate a positive return. The success rate in AD speaks for itself.

      “2. The oil mining analogy is silly. Its a good analogy, assuming NO INNOVATION. But that’s the whole point, to discover new things.”

      Why isn’t drilling for hidden new oil reserves just like drug discovery? They both involve looking for (i.e., *discovering*) potential new opportunities, which may vary in their value, as well as their cost and complexity. And once they are exploited (drug patents expire), their commercial value is gone.

      Even folks working in the oil industry have told me just how similar it is to drug discovery, and they use similar portfolio management tools to evaluate and prioritize their investment opportunities under uncertainty.

    2. Kelvin says:

      PS. Another very good (but perhaps more surprising) analogous industry to drug discovery, is the movie/film industry – huge investments (years and hundreds of $millions in filming and marketing), huge potential upside ($billions), lots of uncertainty (will people go see it?), high failure rates (box office flops), more and more incremental innovation (the same tired old plots).

      Of course, the key difference is that in the movie industry, you have almost a century of exclusivity with copyright protection vs 20 years exclusivity with patent protection in the drug industry. Hardly seems fair!

      1. Anonymous says:

        yeah but you can make entire movie without any human actors involved

  14. Billy says:

    Most drugs are difficult to discover because of biology’s complexities
    Most people don’t want to pay for them.

    Maybe we in the pharma industry should’ve stuck to something more lucrative like marketing. Ugh, my stomach turns just thinking about it.

  15. tt says:

    The analysis presented (as well as prior ones that trend the same) appears fairly sound…the business model as it stands today is not sustainable and cannot continue as is. It’s clearly not for lack of medical need or even biopharma’s ability to innovate, rather it is a function of the market forces that fund and ultimately pay for that innovation being rather unstable (i.e. the US healthcare system that pays for new drug development and subsidizes the rest of the world). Only a fool would hazard a guess as to what the industry will look like 10 to 20 years down the road, but it certainly will not look at all like it is today as changes not just in technology, but how we fund it is going to have drastic impacts on pharma. Just imagine if the US decided to go single payer and only pay for drugs with outcome based pricing, what and how would that impact how we fund and evaluate risky drug discovery programs?

  16. anon says:

    drug discovery will eventually become a fully government funded enterprise.

    1. NMH says:

      Supported by overpaid government employees or faculty, and done by cheap labor, whether it be american grad students or H1B’s from Chindia.

    2. tlp says:

      Why not by Amazon/Google/Apple/Facebook ?

      1. Scott says:

        Because Amazon/Google/Apple/Facebook can’t make any money at it.

        Fact: Most businesses aim for a minimum of 20% return on investment. Good luck pitching any project that doesn’t have at least a 20% IRR past the Veeps, the C-levels won’t even see it.

        When ROI becomes less than inflation, NO for-profit business will touch the project. This leaves drug discovery to the obscenely well-funded non-profits or the government.

  17. tlp says:

    The article represents the worst data analysis I’ve seen in this blog.

    1. Anonymous says:

      Any particular reason? It may be over-simplified, but I don’t think fundamentally flawed.

      1. tlp says:

        Well, ‘fundamentally’ it picks up author’s apocalyptic agenda (and current Zeitgeist) and finds a metric to fit data to it. After all if you beat your numbers hard enough you can make them look whatever you want.
        And it’s pretty easy to check if the prediction is true: if by 2020 there is massive outflow of capital from Pharma industry the author is right.

        PS. Will error bars be in the curriculum for MBAs anytime soon?
        PPS. It actually can be the case that majority of pharma companies are doing fine but overall the picture is disastrous. Look up Simpson’s paradox.

  18. John Wayne says:

    I read through the linked article, and I think it is a mix of parroting what many already know (business as usual for pharma will not work out) and some narrative misdirection.

    1. “In essence, drug discovery is rather like drilling for oil, …”

    Health care isn’t like drilling for oil at all. Therapies that actually move the needle aren’t priced, discovered, or thought of like a commodity in any way (ex: new Hep C cocktails). They all do eventually become cheap and commonplace. Please remember that we all make our money on the transitional period between the discovery and commoditization.

    2. I completely disagree with his chart showing the relative ease of specific disease areas. An an example, anti-infectives isn’t a solved problem unless you have been smoking too many cannabinoids. It may seem solved now, but it isn’t; most knowledgeable people know this.

    3. “The direct mathematical result of the Law of Diminishing Returns which we have already seen in our analysis above, …”

    This ‘law of diminishing returns’ is something you suggested, and doesn’t apply here; most health solutions we would like to have do not exist, and most of the ones that do exist can’t be easily replicated because we’re still figuring out what we’re doing.

    “Direct mathematical result?” What math? All I see are some graphs and an amateur attempt to get the readers to not think critically about your assumptions because you used the words ‘law’ and ‘math’ in the same sentence.

    4. “I also explore the underlying drivers and make concrete projections, which suggest that the entire industry is on the brink of terminal decline.”

    The brink? Pharma as we knew it is already gone; everybody who has been paying attention has known this for a quite a while.

    On a more positive note, I don’t think of dye, chemical, pharma or biopharma companies as being very different from each other. All these organizations attempt to use chemistry, science, and engineering to try and make life better for people. Let’s all worry less about labeling business trends, powerpoint slides, and the negative aspects of corporate culture and instead try to run experiments that will move the needle. What experiment are you doing tomorrow? I hope it is a good one.

    1. Hap says:

      If no one thinks it’s worthwhile or profitable to find drugs, we won’t be able to do useful experiments. If worrying about the problems that makes that likely keeps you from doing something, then it’s not helpful, because you’ve only got so much time.

      At least for some things small molecules reign – antivirals, for example, and antibiotics.

      1. John Wayne says:

        I completely agree that, at the end of the day, some people with money need to be convinced that scientists can do something useful. I believe that the current hype/reality cycle is a net negative to our collective credibility. It makes us look like we can’t get anything to work, and it convinces young people that there isn’t anything new to discover out there.

        The only part of science that is a process workflow is designing experiments and looking at the data.

  19. dearieme says:

    “In essence, drug discovery is rather like drilling for oil, …” How absurd.

    In essence, drug discovery is rather like making love to a beautiful woman.

    1. Wistful says:

      I’ve done the latter; hope to do the former before I retire.

    2. Diver Dude says:

      The first reference to Swiss Tony I’ve seen in any scientific discussion. Chapeau, sir!

    3. MoMo says:

      You should have this quote made into a sampler and hang it up in your labs-

      “Fornication is easier than discovering drugs, and talking about both is even easier”

  20. Che-mist Guevara says:

    No worries, folks.
    Robots are taking a large chunk of low-end jobs, a nuke may hit our country or fake news may take over any hopes of remaining a respected people – and society (eg, us) had better considered these “inconveniences” as more tractable that curing Alzheimer’s disease.
    The impact of such factors will occur long before major breakthroughs in public health – unless something changes.
    Too bad there’s no app to fix these…. yet!

  21. a says:

    Even if this analysis is correct, we already know how pharma can survive: the same way it did up until the 90’s. That is, manufacturing commercial products with a high margin, then using the profits from that to fund development of new drugs opportunistically. Selling powders, baby formula and skin cream under a highly regulated manufacturing environment AND a trusted brand that together make a barrier to entry for start ups. This is what JnJ is now, right? One of the only big pharmas to still have commercial products and a trusted brand. Seems to me that you either go this way, or go the Amgen route and just make Abs. I think this is what Witty was going for when he choose a Loreal exec to succeed him.

    1. Thomas McEntee says:

      re ‘a says’: I was in my physician’s office yesterday getting a check-up on my lingering congestion/cough and remarked how the cefuroxime axetil I had picked up at my local national chain drug store came from a company in India that was the recipient of FDA letters noting its failure to maintain adequate quality control and GMP plant cleanliness… We talked about probability that the national drug store chain analyzes all of its foreign-sourced drugs (LOL) and how much more a brand name version would cost. The doc mentioned that his father-in-law was a retired pharmacist who owned his own drug store and who had said how hard it was even to sell drugs at the retail level and still make money. The only was he stayed afloat was to resort to selling half-gallon bottles of Coke/Pepsi and hard liquor.

      1. Emjeff says:

        This is absolutely true. Pharmacists are not making money (reimbursments are awful, and the chain stores survive by selling other crap in the front end). Doctors aren’t really making money, and the state of Pharma is obvious to us all. Which begs the question, who is making the money in health-care these days?

        1. grim says:

          There are too much middle men in the system: insurers, lawyers, price negotiators, accountants, etc. It’s painfully clear if you look at who actually get paid how much for any expensive prescription drug.

          1. Thomas McEntee says:

            For a Sankey diagram of US healthcare expenditures by age group, aggregated condition category, and type of healthcare, see paper by Dieleman, Baral, Birger, et al. in JAMA, published 27 December 2016 (JAMA. 2016;316(24):2627-2646. doi:10.1001/jama.2016.16885). It’s available in its entirety once you register on the JAMA site (free). Very interesting but still doesn’t explicitly address the issue of WHO in the healthcare racket are getting the big bucks (but dear reader, we know it’s the C-suite crowd)

        2. Thomas McEntee1 says:

          My guess as to who is making the money are the C-suite people at the very top. When Gary Cohn recently asked a roomful of CEOs how many would take the money from a reduced corporate income tax rate and use it to create jobs, only a few raised their hands (he said ‘where are all the hands?’…). Most will use it to raise dividends or do stock buy-backs…all of which will justify them paying themselves more. We’ve all read about the obscene increases in the ratio of CEO pay to ‘average employee’ pay/compensation.

          Emjeff’s comment made me think of Sankey diagrams depicting inputs and outputs. I’ve seem these for energy but never for health-care. Nice thing about Sankey diagrams is their capability to include all of the numbers. For energy, they’re amazing in way they show waste.

  22. whydontuhirelegitimatescientists says:

    Pharma shot itself in the foot when the business types and HR decided to staff and outsource using the same people that fill our box stores with a bunch of useless junk that doesn’t work.

    All this talk of pharma R&D being toast is nonsense.

  23. doogs says:

    I’ve been out of the chemistry world for a while, so take this with a grain of salt. That said, I’d like to give an alternative, very optimistic view.

    It seems to me that we are on the verge of some serious breakthroughs. For a few decades we’ve been throwing serious research and money at the cancer wall. But look how far we have come! There is a deep understanding of pathways, genes, and key molecular interactions for many types of cancer. That stuff didn’t exist even 10 years ago. The role of the immune system and a little about how to modulate that is being fleshed out. It’s crazy to think about, but I truly believe most cancers will be curable within my lifetime.

    Here is another one: HIV. It wasn’t until Bush Jr. that there was a real initiative to cure HIV. Within 20 years there are vaccine trials. My understanding (if there is an expert in the house please correct me) that efficacy is low, around 30%. But, hell, that’s infinitely better than 0%.

    Here is my point. The show will go on! Healthcare and science and research will ALWAYS progress. If you step away from the trees for a minute, a case can be made that we are on the verge of some serious breakthroughs. Great strides have been made in the last 25 years. What might seems scary is that the role of small molecules is changing.

    Perhaps the analysis simply (unknowingly?) Tracks the rise and fall of the small molecule as the (first, primary, only?) Pharmaceutical.

    Maybe in 200 years, people will look back on our lifetimes and chuckle that we attempted to treat everything with small molecules.

  24. PI says:

    Funny how there are never Pharma layoffs in the executive division….you know, the people that sit in meetings all day, never work weekends, and couldnt descibe the basic biology of drug action to save their lives.

Comments are closed.