Skip to Content

Eroom’s Law

There’s another “Troubles of Drug Discovery” piece in Nature Reviews Drug Discovery, but it’s a good one. It introduces the concept of “Eroom’s Law”, and if you haven’t had your coffee yet (don’t drink it, myself, actually), that’s “Moore’s Law” spelled backwards. It refers, as you’d fear, to processes that are getting steadily slower and more difficult with time. You know, like getting drugs to market seems to be.

Eroom’s Law indicates that powerful forces have outweighed scientific, technical and managerial improvements over the past 60 years, and/or that some of the improvements have been less ‘improving’ than commonly thought. The more positive anyone is about the past several decades of progress, the more negative they should be about the strength of countervailing forces. If someone is optimistic about the prospects for R&D today, they presumably believe the countervailing forces — whatever they are — are starting to abate, or that there has been a sudden and unprecedented acceleration in scientific, technological or managerial progress that will soon become visible in new drug approvals.

Here’s the ugly trend (dollars are inflation-adjusted:
R&D trend
I particularly enjoyed, in a grim way, this part:

However, readers of much of what has been written about R&D productivity in the drug industry might be left with the impression that Eroom’s Law can simply be reversed by strategies such as greater management attention to factors such as project costs and speed of implementation, by reorganizing R&D structures into smaller focused units in some cases or larger units with superior economies of scale in others, by outsourcing to lower-cost countries, by adjusting management metrics and introducing R&D ‘performance scorecards’, or by somehow making scientists more ‘entrepreneurial’. In our view, these changes might help at the margins but it feels as though most are not addressing the core of the productivity problem.

In the original paper, each of those comma-separated phrases is referenced to the papers that have proposed them, which is being rather scrupulously cruel. But I don’t blame the authors, and I don’t really disagree with their analysis, either. As they go on to say, investors don’t seem to disagree, either. The cost-cutting that we’re seeing everywhere, particularly cutbacks in research (see all that Sanofi stuff the other day!) are the clearest indicator. People are acting as if the return on pharmaceutical R&D is insufficient compared to the cost of capital, and if you think differently, well, now’s a heck of a time to clean up as a contrarian.
Now, the companies (and CEOS) involved in this generally talk about how they’re going to turn things around, how cutting their own research will put things on a better footing, how doing external deals will more than make up for it, and so on. But it’s getting increasingly hard to believe that. We are heading, at speed, for a world in which fewer and fewer useful medicines are discovered, while more and more people want them.
The authors have four factors that they highlight which have gotten us into this fix, and all four of them are worth discussing (although not all in one post!) The first is what they call the “Better Than the Beatles” effect. That’s what we face as we continue to compete against our greatest hits of the past. Take generic Lipitor, as a recent example. It’s cheap, and it certainly seems to do the job it’s prescribed for (lowering LDL). Between it and the other generic statins, you’re going to have a rocky uphill climb if you want to bring a new LDL-lowering therapy to market (which is why not many people are trying to do that).
I think that this is insufficiently appreciated outside of the drug business. Nothing goes away unless it’s well and truly superseded. Aspirin is still with is. Ibuprofen still sells like crazy. Blood pressure medicines are, in many cases, cheap as dirt, and the later types are inexorably headed that way. Every single drug that we discover is headed that way; patents are wasting assets, even patents on biologics, although those have been wasting more slowly (with the pace set to pick up). As this paper points out, very few other industries have this problem, or to this degree. (Even the entertainment industry, whose past productions do form a back catalog, has the desire for novelty on its side). But we’re in the position of someone trying to come up with a better comb.
More on their other reasons in the next posts – there are some particularly good topics in there, and I don’t want to mix everything together. . .

46 comments on “Eroom’s Law”

  1. Biotrekker says:

    Great piece. The graph pretty much says it all. our point about “nothing going away” is really not well understood, generally.

  2. Rick Wobbe says:

    A few years ago I published a similar graph in an article for the newsletter for PMI’s Pharma Specific Interest Group covering the productivity of the discovery of novel drugs (not all drugs) over the same time period. You can find it here:
    There was one interesting difference I found: there was a pronounced bi-phasic behavior to the curved centered with a break occurring in the mid-1980s. Pre-1980s the number of novel drugs discovered per $ billion bounced randomly around 12 novel drugs/$billion (Y2K dollars inflation adjusted). After that, an exponential decay started similar to that depicted in the graph you’ve shown with a half-life of about 1 fewer novel drugs discovered per $ billion. Interestingly (and perhaps informatively???), the break point coincided exactly with what is labeled as the “first wave of biotechnology-derived therapies”. Because we were all so convinced that the dawn of the biotechnology era was supposed to usher in faster, cheaper, better drug discovery, and yet the trend persists today, I believe this contrary result bespeaks something profound that we have yet to see clearly.

  3. Anonymous says:

    Not sure if this is useful to understand their rigor vs. motives, but NRDD’s disclaimer notes that three of the four authors are employees of a “sell side investment firm”. From Wikipedia: “Sell side … is a general term that indicates a firm that sells investment services to asset management firms…”

  4. John says:

    I don’t think anyone is arguing that the FDA safety regime hasn’t shifted over into a net negative. They’re effectively valuing each human life at $100m or more. The EPA uses $7m, by contrast.
    How do we fix this? Exporting the last few pharma jobs to less squeamish countries seems like a bad strategy.

  5. David P says:

    I’d love to see the same graph plotted against CEO salary.

  6. johnnyboy says:

    The scariest thing about this graph is that the Y axis is logarithmic… Like I asked before, how long before we all have to pack it in an go fishing ?
    On the ‘better than the beatles’ problem, my usual analogy for the layman is to imagine that you are a car-making company, in a world in which most people already have a car, because cars never get damaged or break down. And there’s a government agency that ensures that every new car model you try to sell is really significantly better than what is already on the market. How long would your business last ?

  7. PharmaHeretic says:

    Did anybody notice that there is not much of a slope between 1970 and mid-1990s? The hacks (“esteemed” scientists) are lying as usual.
    I detest subhuman scum who try to mislead by manipulating data.

  8. luysii says:

    All too true, but the real reason is that the low hanging fruit has been picked, and we are attempting to modify a system whose internal workings we only dimly understand. I’ve posted 21 reasons why this is the case — the latest being A list of all 21 is available at

  9. Rick Wobbe says:

    I wonder how the situation would look if you measured the size of the available drug space left (a kind of surrogate for residual need) in terms of: a, net profit vs. b, fraction of diseases that are truly poorly treated or untreated. I haven’t seen such an analysis, but I suspect it would look like this: much/most of the space has been adequately filled when measured as net profit (case a); and most of the space remains to be filled when measured as under- or un-treated diseases (case b).
    Case a would be what investors are concerned about. Who would care about case b?

  10. Anonymous says:

    @7 PharmaHeretic
    Indeed. I wonder what the linear graph would look like. Would you see stagnation there, from 1968’ish to 198’ish.

  11. Rick Wobbe says:

    @ PharmaHeretic, #7,
    Sorry to be hogging space, but I’m pretty passionate about this. I will restrain myself after this. PharmaHeretic, you are correct about the relatively level segment centered around the 70s. If you look at the data on a linear plot and line it up against a running average rather than a linear regression line, your observation is even more apparent. In my opinion, a regression line is inappropriate because it implicitly assumes that there are no “externalities” (i.e. new variables) that might enter/leave over time. With great apologies for looking like a self-promoter, look here ( for a linear plot.
    One more subtle, but important point. Note that the X-axis refers to the year in which drugs reached market, which occurs years AFTER the R&D dollars on the Y-axis were spent on that drug, so there should be an offset if you want to compare R&D spending with it’s impact on the discovery and development of drugs in a relevant time-frame.

  12. Hap says:

    From 1968-1985 looks to fit in a pretty flat band – the trends for the years before and after seem to be consistently downward-sloped, In the worst case (1968-1983), the line slopes downward more rapidly than the whole line, but other lines through that segment go upward. The fit line is above the approvals for that segment early and below the approvals for that segment late – the errors are not randomly distributed within that segment.

  13. fredo says:

    What about calling it ‘Eroom’s Wall’ – because if you believe the hypothesis drug discovery is hitting an insurmountable obstacle and may soon be substantially uneconomic…

  14. SP says:

    And unlike the forward-spelled law, there’s no hypothetical physical limits on this graph- it’s entirely possible to go to infinite cost per drug (ie, 0 drugs produced in a year.)

  15. David P says:

    @8 Luysii, did you see John LaMattina’s take on the “low hanging fruit”? On his Forbes blog.

  16. Jonny says:

    @8…agree with you on this.
    A related chart dealing with the complexity of the issues remaining to be solved over time (after the relatively easy fruit such as hypertension have largely been picked) would probably be enantiomeric to the chart shown above, implying that the
    deteriorating “productivity” of R&D is entirely rational.

  17. anon the II says:

    Over my long storied career, I worked at two major pharmaceutical companies and an assortment of biotechs. One thing has always bothered me. The cultures at the two big pharmas were radically different. One was miserly, the other spent money like crazy. One was humble, the other arrogant. One was east coast, the other mid-western. Everything was different. And yet, the result is the same. Sort of argues for an external force. I always blamed it on DTC advertising and the mentality that put in the board room. But maybe it was bigger than that. Maybe marketing was just one reaction, amongst many, to that slow inevitable decay.

  18. NoDrugsNoJobs says:

    A society has to ultimately decide what it wishes to discourage and encourage. If pharma research is getting tougher to see a return on investment such that research is not paying off, do we wish to change the equation or do we believe that pharma research is not important enough to encourage. Current drug exclusivity times are largely based on patents and are not as long as for other areas such as computer manufacturing. The average drug will receive something like 10-11 years of exclusivity from patents even though patents provide 20-21 years of exclusivity for many other technologies. The reasons are primarily due to extraordinarily long development times and Hatch Waxman which actually encourages generic companys to bust the innovators patents.
    Normally, we think patents exist to incentivize risk taking in R&D but in what is probably the riskiest and most expensive of all commercial industries we see a paradoxically inverse reward of exclusivity. Patent terms are not carved in stone….copyrights are way longer and trademarks can be forever. The question is what do we wish to encourage? What are our priorities?
    This could all be fixed quite easily really. In Europe, there is 10-11 years of guaranteed exclusivity absent patents and oddly enough, the new biologics legislation passed in the US gives 12 years of guaranteed exclusivity to proteins like antibodies although the cost of bringing a small drug to market has been estimated to be higher than for a biological. Small molecules hold the most promise for answering so many of the pressing health concerns around the world from treating HepC to malaria to you name it. Orally available, small molecule drugs that can ultimately be cheaply produced are what it seems to me, we would wish to incentivize but what do I know, I’m just a medchemist.

  19. anon 3 says:

    “And there’s a government agency that ensures that every new car model you try to sell is really significantly better than what is already on the market”
    This is a common misconception. The FDA does NOT ensure drugs are superior to previous drugs; they have to be safe, effective, and consistent based on the latest science. The issue is that sometimes the science suggests that the innovator, or fisrt in class, wasn’t really that safe or effective.

  20. barry says:

    Before we despair at the prospect of improving on Lipitor or ibuprofen, we must remember that there remain real unmet medical needs in which we can be first, or at least best. Most cancers (and we still don’t know how many diseases “cancer” is) are in this category.
    And for the foreseeable future, we’ll always be needing novel antibiotics. The golden oldies in that field are not stealing potential market share from the new ones (if there were new ones) because they rapidly become inefficacious.

  21. BOB says:

    I guess it’s a small point, but shouldn’t the y-axis be adjusted for inflation? And perhaps for the cost of drugs as well. Profits once a drug is approved are far higher now compared to 1950.
    I feel like the ratio of cost of a drug to the profit realized per drug would be a better metric. However, I’m pretty sure that graph wouldn’t look too pretty either.

  22. JasonP says:

    I for one find it highly offensive that you do not drink coffee Derek. Obviously you are not working hard enough.

  23. Drsnowboard says:

    and I find it incredible that Nature charges $32 dollars for an article like this. Roll on open access publishing, after all – the review process for this one can’t have been incredibly demanding..

  24. wwjd says:

    The graph seems to be the inverse of global temperatures. Maybe global warming is causing the decline of pharma productivity.

  25. LeeH says:

    Why the discrepency between the cost/drug in this analysis (~$1.5B) and that from the previous blog “The Terrifying Cost of New Drugs”, which put the cost at $6-11B?

  26. TJMC says:

    First, I think the authors have done an exceptional job in breadth and depth on this issue. Too often folks in the industry or external consultants focus on one magic bullet or singular perspective based on their agenda or experience. This and the references are a great resource for those that want to really understand the issues. Or at least a good starting point. I have a number of critiques, but want to first say “well done”.
    The first of two main issues I have is that the analysis embraces too wide a timeframe. While the authors’ make a point of comparing R&D approaches in the two main periods (“phenotypic vs. targeted”), the need is to fix what we have now in this decade. Waaay too far to back up 30 years.
    Second, and more importantly, they fail to measure apples to apples, when measuring new drugs approved vs. dollars invested. The “many drugs approved” in the “golden age” probably did not garner anywhere near the median profits of 2000-2012 drugs. This would lessen the first chart’s gloom considerably. NOT reverse it, but let’s measure apples to apples, dollars (invested) to dollars (euros, yen…) returned. I have been focusing on the overall process and returns for the industry in this way for more than 20 years. Despite the gloom, I still see value to society and hope going forward, but there are a lot of distractions from the right focus and measures. NME per $Billion is not the right focus or measure.
    For instance, the talk of a “Eroom wall” echos a similar “ceiling” Lehman publicized in 2000 (so is there anything new under the sun?);

  27. Anonymous says:

    #23 – Yes, but the authors are in the business of being paid for that kind of analysis.
    Unless the “follow-on value” is more than the glory of being published and they purpose instead to lay the groundwork for their credentials in the role they describe as “CDDO” (Chief Dead Drug Officer). Actuallly an interesting idea and need right up there with “who bells the cat”?

  28. g says:

    I’m a bit pessimistic about the drug industry in the next 20-40 years. They’ve done a fantastic job getting drugs for many major indications that do pretty good. With current prescribing habits/reimbursement models, it is going to be hard to continue improving without having the benefit of incremental improvements (not getting ROI for me-too drugs).
    When I look at the current problems facing healthcare, spending too little money on drugs doesn’t seem to be a problem. The bigger problems don’t even have anything to do with a lack of medications. We can’t pay for our current level of care let alone adding on new treatments, personalized medicine, etc. The biggest changes to healthcare will be on preventative medicine (often using generics) and lifestyle modifications to prevent chronic illnesses associated with living longer and a Western lifestyle. Neither of these requires new drugs, but would save $ trillions. My guess is that pharma is headed for a long slump until some drug-resistant epidemic spreads to the developed world.

  29. TJMC says:

    #18 – I think you raise a key idea for resolution to the external issues raised in the article. (as opposed to the CDO role mentioned by #27)
    That is, if society puts a greater burden (safety, etc.), yet high value on remaining unmet needs, they need to incentivize that investment by offsetting the time to market vs. patent length issue. While operational efficiencies and improving the R&D process can go on and on, the external factors may be beyond such internal improvement efforts.

  30. Jeffrey Soreff says:

    9. Rick Wobbe
    Re the finite space:
    Note that an exponential decline in drugs/effort
    implies that, even with no cut in spending,
    the number of drugs remaining to be found, _ever_,
    is roughly the number that were found during the
    last halving of drugs/effort, roughly the last
    decade. Again, that’s _without_ disinvestment.

  31. fishlips says:

    The problem with Pharma today is I spend my day doing no bench research, repetitively revise my yearly objectives and at the end of the day, grossly embellish my program direct time card. I make every attempt to get along with everyone I hate, laugh at my ignorant boss’s jokes and sit in the front row chanting like a parishoner at a bible belt megachurch while our R&D head does his bobblehead routine. Multiply my day by 20 years and 67,000 employees.

  32. Rich Rostrom says:

    I think this graph understates the problem. The plotted value is US drug approvals/US research spending.
    But US research is not the only source of US approvals. Have there been no drugs developed elsewhere and brought to the US?
    Over the last 60 years, drug research spending outside the US has expanded considerably (or so ISTM). Europe and Asia have both convered with the US in wealth and development.
    Also, US-based companies have expanded research overseas. Are those expenditures included? (And OTOH, what about in-US expenditures by foreign-based corporations, such as GlaxoSmithKline?)
    Surely the ratio should be US drug approvals to global drug research spending.

  33. dearieme says:

    What will happen when it’s widely realised that (i) For most categories of patients, statins don’t extend life, but (ii) taken for years they cause seriously nasty side effects? Exit an industry?

  34. cirrus says:

    Where is the beef? With all these smart discussions, someone should be able to find a way to do it right or make money out of it like the article’ authors.

  35. cynical1 says:

    @ fishlips…….
    that was classic………but not part of their equation……

  36. Mike says:

    Mostly, I think big pharma is getting tired of being high-stakes gamblers. The whole idea of large-scale drug research was that you bet on as many mechanisms as possible, losing your entire bet on almost all of them, but on the very, very few that worked out, making enough to pay for another round of betting on everything (and a tidy profit).
    Now big pharma is betting on almost nothing. Instead, they expect a lot of little companies to each bet their shirts on a single mechanism, and almost all of them will go broke. But every once in a while, one or two will win the lottery and come up with that multi-billion dollar idea. Then big pharma steps in and buys the lottery ticket from them.
    This will make it look like the cost of drug discovery is going down, because the costs of trying and failing on a mechanism will disappear into a large number of failed companies rather than being collected in a single large research unit that expects to be around for decades. The one or two companies that do come up with something good don’t have to justify the costs of a whole mess of unsuccessful programs in the selling price of their successful ones, and big pharma doesn’t have to account for research costs at all.

  37. Old Chemist says:

    The real underlying reason for the ‘decline in productivity” is that two of the major remaining unmet disease areas, cancer & neuroscience, have incredibly complex biology systems and that simple, hit-a-target-get-the-response-you-want strategies no longer work. I can’t address neuroscience, but in cancer, the idea of single-agent treatment is manifestly unreasonable, and determining the best combination for treating a particular cancer is probably going to ultimately come from an analysis of existing clinical data. Unfortunately, these data are expensive and slow to come by, but right now, anyway, there is no good alternative. It’s as if we are slowly assembling a toolbox, the best use of which will take a long time and a lot of effort to figure out.
    There are no business or R & D-streamlining or sigma-6 solutions to this problem, contrary to what analysts & CEO wish.

  38. TJMC says:

    I mentioned above that I still have optimism that some R&D strategies will pan out, and that there will be some winners.
    One reason: I disagree with all but one of their root-causes, but Tollman, et al in Nature Reviews/ Drug Discovery, do a great job of comparing the value R&D brings to 10 different firms. There are substantial differences between firms, so maybe some of the “losers” (all larger firms) are dragging the industry down. If you have access, check out their charts for some surprises. Not sure if they can be reposted without permission.
    “Identifying R&D Outliers” NRDD 10, p653-654 (2011)
    I intend to compare their approach to similar analyses on what is creating better returns on R&D now. See if there are marked differences between firms focusing on NME’s vs. “nme’s”. Will post that in coming week.

  39. Jack Scannell says:

    One of the posts mentioned the high cost of buying the article from Nature Reviews for those who do not have a subscription. As one of the authours, I have reprints (or, rather, “eprints”)and can email them to those who request one. jack.scannell [but without the space]

  40. Legacy Merck Guy says:

    @ #5 The chart of the investment/drug ratio is linear. The one with the executive pay is exponential.

  41. Chemist turned Hedgie says:

    @3 and others
    I understand the intake of breath on seeing that the authors work for an investment bank, but rest assured these guys are some of the smartest and least biased in the industry. Most of their competitors would struggle to get published in their parish magazine, let alone a Nature journal. I know because I am one of their clients…

  42. David Booth says:

    The problem is not in the trend that this graph exposes, but in the authors’ and readers’ expectations!
    OF COURSE the biggest gains in a newly discovered market are at the beginning. There simply cannot be an unlimited number of blockbuster drugs. Why would anyone expect anything BUT a diminishing ROI as the prospects are increasingly picked over and drugs become increasingly specialized? This graph should not be a surprise to anyone.

  43. Deyan Mehandjiev says:

    I’d love to see the same graph plotted against CEO salary.

  44. Deyan Mehandjiev says:

    I’d love to see the same graph plotted against CEO salary.

Comments are closed.