Skip to main content

Drug Development

GlaxoSmithKline Reviews the Troops

Several readers sent along this article from the Times of London (via the Ottawa Citizen) on GlaxoSmithKline’s current research setup. You can tell that the company is trying to get press for this effort, because otherwise these are the sorts of internal arrangements that would never be in the newspapers. (The direct quotes from the various people in the article are also a clear sign that GSK wants the publicity).
The piece details the three-year cycle of the company’s Drug Performance Units (DPUs), which have to come and justify their existence at those intervals. We’re just now hitting the first three-year review, and as the article says, not all the DPUs are expected to make it through:

In 2008, the company organized its scientists into small teams, some with just a handful of staff, and set them to work on different diseases. At the time, every one of these drug performance units (DPUs) had to plead its case for a slice of Glaxo’s four-billion-pound research and development budget. Three years on and each of the 38 DPUs is having to plead its case for another dollop of funding to 2014. . .
. . .Such a far-reaching overhaul of a fundamental part of the business has proved painful to achieve. Witty said: “If you look across research and development at Glaxo, I would say we are night-and-day different from where we were three, four, five years ago. It has been a tough period of change and challenge for people in the company. When you go through that period, of course there are moments when morale is challenged and people are worried about what will happen.”
But he said it has been worth the upheaval: “The research and development organization has never been healthier in terms of its performance and in terms of its potential.”

I’m not in a position to say whether he’s right or not. One problem (mentioned by an executive in the story) is that three years isn’t really long enough to say whether things are working out or not. That might give you a read on the number of preclinical projects, whether that seems to be increasing or not. But that number is notoriously easy to jigger around – just lower the bar a bit, and your productivity problem is solved, on paper. The big question is the quality of those compounds and projects, and that takes a lot more time to evaluate. And then there’s the problem that the extent that you can actually improve that quality may still not be enough to really affect your clinical failure rates much, anyway, depending on the therapeutic area.
Is this a sound idea, though? It could be – asking projects and therapeutic areas to justify their existence every so often could keep them from going off the rails and motivate them to produce results. Or, on the other hand, it could motivate them to tell management exactly what they want to hear, whether that corresponds to reality or not. All of these tools can cut in both directions, and I’ve no idea which way the blades are moving at GSK.
There’s another consideration that applies to any new management scheme. How long will GSK give this system? How many three-year cycles will be needed to really say if it’s effective, and how many will actually be run? Has any big drug company kept its R&D arrangements stable for as long as nine years, say, in recent history?

35 comments on “GlaxoSmithKline Reviews the Troops”

  1. wwjd says:

    The biggest flaw in the plan is that Moncef and Patrick are the ones who evaluate the progress of the DPUs.

  2. Moliere says:

    See comment by TwoPointZero here:
    The whole thing sounds like deck-shifting on the Titanic as far as R&D @GSK is concerned. Most believe that their medium- to long-term strategy is to “externalize” R&D, if you don’t be;lieve me check out this gem on YouTube:

  3. Hap says:

    I wonder if their management could abide a similar system. I suspect then that we’d “need the dependability of long-term funding” to ensure the success and loyalty of management.
    Comment 1 also makes sense – while having better and nearer-term signposts to know what’s going on with your (expensive) research is a good idea, it doesn’t really matter if your driver is drunk, or can’t read.

  4. PharmaHeretic says:

    The Romans did something similar by throwing slaves, criminals and gladiators in arenas for fight-to-death spectacles. I am glad to see that GSK has revived this ancient roman tradition.

  5. exGlaxoid says:

    Having worked for GSK for 20 years I can say without a doubt that the new approach is yet another fad that will also fail to fix the problems with big Pharma today. They have merged, reorg’ed, rightsized, automated, six-sigma’d, lean’orged, and every other buzzword possible, but everytime the scientists get a project moving well, the organization changes and it loses its momentum.
    Over the years top managers have pushed their pet projects; in-licensed magic beans from Affymax, Sirtris, and a host of other; and spent millions on pipe-dream ideas which were discarded a few years later.
    Now the idea is to take the researchers in small groups and pit them against each other like a tawdry reality show, which creates little fiefdoms rather than encouraging real collaborations and wise sharing of people and resources. This will succeed in duplicating the one thing that most biotech’s seem good at, exaggerating their results. So each DPU will claim that they have the best projects with the best drugs rather than kill projects before wasting a lot of time and money on them, as they have to have something to parade at the dog and pony show.
    Best of luck to the DPUs.

  6. Wiz says:

    My opinion is somewhat biased because GSK and I agreed to see other people at the end of 2008 (around the time of the introduction of DPUs)but I doubt the DPUs have had much impact at all. They kept a lot of the same management (and projects) and slotted them into the DPUs and it seemed like it was going to be a pretty superficial change (like when they introduced the CEDDs a couple years before). I think that the idea/concept of the DPUs could offer some advantages, but with the same people running the show I doubt that things have changed that much in the labs. Time will tell I guess.

  7. drug_hunter says:

    exGlaxoid has it exactly correct: separate DPUs will absolutely STIFLE interactions across diseases. I just came out of a discussion where the same target makes sense in both inflam and neuro disorders. We have experts from both areas (and 3 sites) involved in the discussion. Biology and disease are all about networks, and the same targets are connected through multiple networks to multiple diseases.

  8. youwish says:

    I actually thought the DPU idea was a good one. Get everyone you need to progress the project in a small unit to foster communication, make it clear that the project is the only game in towm so there’s no moonlighting, and ensure that the unit is accountable for performance. That really is the essence of the biotech experience.
    As everyone has alluded to, the difficulty is in evaluating the progress. If you don’t make tough decisions then it defeats the purpose of the model, but if you cut prematurely then you risk overlooking an asset. I actually think the decision would best be made by some kind of panel with significant proportion of outsiders (academics, clinicians, and, yes, scientific consultants). That way the political games are minimized.
    I also doubt that many people in the failed DPUs will get canned. The successful DPus will absorb the resources as their projects grow.
    The DPU heads are actually pretty close to the science, not just desk jockeys, so they should find a place after the whittling. Hopefully management will remember that some DPUs will fail not because of poor execution but because of an incorrect founding hypothesis. Showing a hypothesis is erroneous can be just as valuable as progressing a successful one.

  9. Mike says:

    Check out the Pharma Giles blog for a slightly different spin on this topic:

  10. biotechbaumer says:

    This strategy sounds eerily similar to what professors have to go through in academia. In fact, the whole grant writing/renewal process is a big reason why many scientists opt towards the industry career track! Hopefully all of industry doesn’t pick up on this DPU concept, otherwise, this may negatively impact the future crop of scientists.

  11. AlChemist says:

    ask Roche Nutley about the Dennis Loh experiment- remember the Rubicon

  12. pete says:

    random thoughts about DPU concept:
    1) Is it really so novel? [nod to exGlaxoid @5] The concept of including biologists + chemists together in defending their common project to executive & SAB review is nothing new, in my experience.
    2) Balkanization/Lack of cross-fertilization between DPUs [drug_hunter @7] . You’d like to think that important overlap/synergy between promising DPUs would be recognized and highlighted — that is, unless the executive + SAB are collectively drunk and illiterate [Hap @3].
    3) 38 DPUs: IMHO, there may be more value in picking fewer disease areas and *more diversity* in screening approaches. For example, it’s been a mug’s game trying to predict relative value of drug screens driven by *target* vs *disease phenotype*: think of all the once-excellent drug targets that became radioactive; all the once-excellent animal models of disease that were discarded like a bad memory.

  13. HelicalZz says:

    I like the idea of a DPU. If the program gets cut or fails to get the adequate attention of management, it is already organized to go out and get funding on its own (spin itself out, seek venture capital, etc.). Whether it can or not is part of both the problem and solution.

  14. HelicalZz says:

    #12 Pete
    1) Is it really so novel? [nod to exGlaxoid @5] The concept of including biologists + chemists together in defending their common project to executive & SAB review is nothing new, in my experience.
    No, that isn’t novel at all. What would be is adding the MBA and clinical strategy team i.e. like a small company. Chemists and biologists can argue scientific merits all day. How many appreciate market size and potential, competitive landscape and areas of unmet need, clinical path to approval such as how many patients would be required in how many arms of what type of pivotal study, how solid is the IP position, etc. etc.
    A drug that will likely generate only $200M per year, at peak revenue, has challengable IP, and requires a 5K person phase III trial to demonstrate efficacy is a non-starter no matter how responsive the cells are or how low the binding constant is.
    If you are part of a small company, you are aware of all this stuff, but inside a large one it can be easy to get insulated from most of it. Yet it is all important when considering a drug program, especially in comparison to another that is also looking for funding (that is the crux of opportunity costs). If a DPU is a formation that puts together a team with all these pieces and gets them all educated on each of them, then it should serve the organization well.
    Yes, this is again the old ‘form separate biotechs within the bigger company’ paradigm that gets a lot of lip service but doesn’t seem to ever fully form. It needs to. M&A is just R&D with different accounting after all (capital expenditure vs. expense).

  15. Anonymous says:

    #8 I would not gurantee that other DPU’s will absorb the resources of those that “fail”. Remember GSK canned 5 DPU’s in Harlow/Verona in 2010 before they had a chance of making it to the end of the current 3 year cycle. Their scientists were not “reabsorbed”

  16. milkshake says:

    @helical: a smaller units that have direct stake in the success of their project is an idea based on emulating startups. The problem is that it was the management that destroyed the spirit of discovery to begin with, and the moronic management is not going away and breaking the research department into DPU only invites more interference and power plays from management. Establishing the DPU and the evaluation will be by necessity a political process. For example, if I was a department head and there was my protege in charge of one of these units, I would make sure he gets the best projects and the best people assigned to him so that 3 years down the road he is not only going to survive but also could be promoted. (Its better to have my protege promoted than some other guy who might be a golden boy sponsored of one of my rivals.)

  17. Anonymous says:

    @7, totally agree, small point though – there is only one network. The misguided vast majority talk about ‘sub-networks’, but they don’t exist either 🙂

  18. pete says:

    @14 HelicalZz
    The scenario you suggest of DPUs as semi-independent, holistic research operations is compelling — closer to the ‘Biotech w/in Pharma’ idea that’s being so loudly touted these days.
    But my read of the Times article is that these DPUs consist primarily of discovery scientists, not clinical-, market intelligence-, or scale up development-types.

  19. Anonymous says:

    @10, I said the same thing!

  20. exGlaxoid says:

    14. HelicalZz on September 1, 2011 2:15 PM writes…
    “No, that isn’t novel at all. What would be is adding the MBA and clinical strategy team i.e. like a small company. Chemists and biologists can argue scientific merits all day. How many appreciate market size and potential, competitive landscape and areas of unmet need, clinical path to approval such as how many patients would be required in how many arms of what type of pivotal study, how solid is the IP position, etc. etc.
    A drug that will likely generate only $200M per year, at peak revenue, has challengable IP, and requires a 5K person phase III trial to demonstrate efficacy is a non-starter no matter how responsive the cells are or how low the binding constant is.”
    That all sounds good except for two problems that I see. First, most projects that were really bad from the revenue, IP, and market share concepts were started at the TOP of management and pushed down. These went on for years, got a “lead” and “candidate” after years of research and then were killed after spending a lot of time and money. But the issues were apparent, to most, early on, but ignored at the highest levels of management, like VPs and up. One example was a drug target that had a generic so cheap that even though we meet every criteria laid out for us, we could not manufacture ANY new product for less than the generic sold for, as it was so cheap. So no matter how good our drug was, the cost would always be so high that it would never make a formulary.
    Secondly, much of the “value” and IP of a target are hard to really measure. GSK valued Cialis at nearly nothing before giving it to ICOS, which became Lilly, who found it worth billions. And Merck had a great patent on Proscar except for a tiny little exception called Avodart. I can think of at least 10 cases where the science lead to a compound that marketing or commercial had never even considered useful. If you are going to let “science” drive the project, then you should start with viable targets and then let the science lead the project.
    I am sure that if I had gone to the VP with a proposal to invent a drug to lengthen eyelashes I would have been fired, but that does not mean it can’t exist or won’t sell at a profit.

  21. Hap says:

    12: I don’t know enough about GSK’s management to know whether they’re competent or not. If they aren’t competent or don’t know what to do, then rearranging the underlings will not make their company successful.

  22. WWJD says:

    Hap, let’s just say GSK management paid $720 million for resveratol and leave at that.

  23. Anonymous says:

    @#11 you are right. Unfortunately, Roche has a bunch of ex-glaxo management so they will mirror GSK’s demise. Especially if they keep giving rubber chickens to their employees!!! LOL Is this an exercise in demoralization or simply immature non-sense??

  24. Cytirps says:

    Why not have the idiot CEO and R&D head defend their stupid reorg or have fight for their existence?

  25. Anonymous says:

    I don’t understand how you can make someone defend their job when they are given the target/drug. People with a choice (talent) will quickly leave. And like others have pointed out, you aren’t selecting for the best target/drug with this…just the best BSer. Re-molding a business unit doesn’t change the science (and the decisions made by those higher up)…I don’t understand. This method is so aggressive I wonder if it’s a last ditch effort.

  26. MydreamjobdiedatGSK says:

    Look, the DPU’s are a sham. The CEDD’s already divided the company into therapeutic areas….which is fine (get biologists, chemists, DMPK, clinical, etc aligned behind a common goal). Breaking those CEDD’s down further into DPU’s only does one thing – make it easier to hand out pink slips to a smaller group of people. Layoffs are easy – just dump a DPU that you don’t like instead of the hassle of cutting 10% of a CEDD and all the reinterviewing/reorganizing/realigning that goes on.

  27. Innovorich says:

    The idea of “biotechs within pharma” sounds good (since the per $ productivity of biotech is higher), but are GSK and others confusing what the actual incentive/driver for this phenomenon actually is? i.e. they’re putting this 3-yr pass-or-fail date in place as if that’s the only change required. What about the types of people employed? What about upside potential for them? (shares/royalties in product). What about management style and (as discussed by HelicalZz) management ability and knowledge?
    Anyway – 3 years is actually a lifetime for most biotechs – 3 months would be more accurate!

  28. And D says:

    Given time and not too much meddling, the DPU model may well work. Not sure if this observation is true of other big pharma, but it seems that every time there is a major re-arrangement of research or merger it takes about 3-5 years for whole process to settle down and become productive. Minor changes as a result of the 3 year review may therefore be OK, but mass culling/major re-arranging of DPUs would risk R&D productivity further.

  29. A-NON says:

    It was cynically well recognized in the hallways of GSK R&D when the 3 year DPU model was begun that the in this newest organizational rendition “would be declared a success” after the first 3-year iteration, as long as the same management structure remained in place. The same upper management is well entrenched, with the model being declared a great success (as there always can be improvements with the inevitable ongoing tweaking to any organizational model), even as victory is declared within their own image.

  30. Another Anon says:

    Year 1: Settling on Staff as part of “rightsizing”, figuring out who is left in other parts of the organization to help do techinical activities, focussing on he revised priority R&D efforts.
    Year 2: Trying to do some work on efforts, while avoiding too many upper managment reviews (and trying to keep CD from making ridiculous comments, providing irrelevent “literature” information that one then has to address)
    Year 3: Looking at what was done in years 1 & 2 to prepare summaries in defending progress of the 3 year cycle, refocussing in areas that have not made sufficient progress and redeploying staff, preparing a new three year plan and contingency budgegts for the next cycle.
    Yup, a lot of progress is made in three years.

  31. Nuclear Option says:

    As an onlooker, my sense is that GSK is not trying something new, just admittedly repackaging R&D using the discovery strategy of nimble collaborating academic labs or the small biotechs they have valued (or over-valued but that is a subject best left to others here). I think it can work, and may be working. To be specific, I have been impressed with EpiNova. It seems they took ownership over an emerging cardiovascular target at GSK, identified an inflammatory indication then developed deep expertise with a likelihood of innovative drugs emerging from their efforts.

  32. wayofthedinosaur says:

    I have found GSK to be a very unusual R&D environment in the industry. R&D seems hyper-focused on the business, cost and efficiency rather than program decision points for programs, which utlimately costs more due to the excessive focus on those parameters detracting from research. Everything gets changed so determination of success or failure of any individual change cannot be evaluated appropriately. Mgmt re-defines successful change and pat themselves on the back. Sirtris was, and is, a monumental cluster and has no probablity for success. If corporate nature doesnt select against the mgmt who coordinated this, expect more of the same.

  33. tuky tuky says:

    A similar “DPU”approach is also followed by certain funding bodies/charities with academic groups in the UK: Professor X gets awarded X amount of money over 5 years to tackle target X. Every year there is a review panel from industry that assesses the progress of the group and decides if the funding continues during the next year. If results don’t meet the mile stones, they shut it down. This same situation is repeated in different DPU’s / small drug discovery groups from all over the country.

  34. KW's ex says:

    There needs to be some deep cuts in oncology especially in the protected C-Band class. What a bunch of political hacks. Get back in the lab or be gone.

  35. Pharmadude says:

    Those who championed the 3-year plan and implemented it will declare it a great success. Doesn’t matter if it actually is or not.

Comments are closed.