About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« DTC Advertising: Is That Where Things Fell Apart? | Main | For Responsible Stem Cell Reporting. . . »

June 28, 2011

Drug R&D Spending Now Down (But Look at the History)

Email This Entry

Posted by Derek

I hate to be such a shining beacon of happiness today, but this news can't very well be ignored, can it? For the first time ever, total drug R&D spending seems to have declined:

The global drug industry cut its research spending for the first time ever in 2010, after decades of relentless increases, and the pace of decline looks set to quicken this year.

Overall expenditure on discovering and developing new medicines amounted to an estimated $68 billion last year, down nearly 3 percent on the $70 billion spent in both 2008 and 2009, according to Thomson Reuters data released on Monday.

The fall reflects a growing disillusionment with poor returns on pharmaceutical R&D. Disappointing research productivity is arguably the biggest single factor behind the declining valuations of the sector over the past decade.

This is not good - although, to be sure, we've had plenty of warning that this day would be coming. But looking at it from another perspective, you might wonder what's taken so long. Matthew Herper has a piece up highlighting the chart below, from the Boston Consulting Group. It plots new drugs versus R&D spending in constant dollars, and if you're wondering what the Good Old Days looked like, here they are. Or were:
What's most intriguing to me about this graph is the way it seems to validate the "low-hanging fruit" argument. This looks like the course of an industry that has, from the very beginning of its modern era, been finding it steadily, relentlessly harder to mine the ore that it runs on. But that analogy leaves out another key factor that makes that line go down: good drugs don't go away. They just go generic, and get cheaper than ever. You can also interpret this graph as showing the gradual buildup of cheap, effective generics for a number of major conditions (cardiovascular, in particular).

There's one other factor that ties in with those thoughts - the therapeutic areas that we've been able to address. Look at that spike in the 1990s, labeled PDUFA and HIV. Part of that jump is, as a colleague theorized with me just this morning, the fact that a completely new disease appeared. And it was one that, in the end, we could do something about - as opposed to, say, Alzheimer's. So if you want to be completely evil about it, then the Huey Lewis model of fixing pharma has it wrong: we don't need a new drug. We need a new disease. Or several.

Well, that's clearly not the way to look at it. I don't actually think that we need to add to the list of human ailments; it's long enough already. But given all the factors listed (and the ever-tightening regulatory/safety environment, on top of them), another colleague of mine looked at this chart and asked if we ever could have expected it to look any different. Could that line go anywhere else but down? The promise of things like the genomics frenzy was, I think, that it would turn things around (and that hope still lives on in the heart of Francis Collins), even though some people argue that it did the reverse.

Comments (60) + TrackBacks (0) | Category: Business and Markets | Drug Development | Drug Industry History


1. Rick on June 28, 2011 11:11 AM writes...

Below is a repost of what I added to Matt's blog. The policy implications of the two different observations are different. "SCB" refers to the Sanford C Bernstein report from which the graph above was taken.

A few years ago, I published a similar analysis of new drug discovery productivity in a report for the Project Management Institute’s Pharmaceutical Specific Interest Group Newsletter (December 2008), correcting for the time lag between discovery of a drug (the time during which the R&D was being done) and launch (after which all the R&D money had been spent). The conclusion was the same for drug discovery after the mid-19980s: an exponential decay curve with a slope indicating we were discovering approximately 1 fewer new drugs per year.

However, the reports differ in the behavior of the graph prior to 1980. The SCB analysis indicates that there was a steady decline going all the way back to the beginning of the modern pharmaceutical industry, parallel to what we see today (albeit with a higher quotient), whereas my analysis found that the trend was essentially random prior to 1980, giving a zero-low slope line (almost zero decline) centering on roughly 12 new drugs per billion dollars R&D annually, adjusted to Y2K dollars using CPI. Adjustment using the higher biomedical price index yields 9-10 new drugs per billion dollars. Another analysis, published by the US GAO in 2006, found something akin to the latter result.

The difference between the two depictions has important strategic and policy implications, particularly as it relates to finding and analyzing root causes. The SCB report indicates that what we see today is part of a trend that’s been going on since the dawn of the modern pharmaceutical industry, so we might want to study factors inherent in drug discovery. My analysis and the US GAO report suggest that the productivity of the industry is biphasic (steady up until 1980s and exponentially declining thereafter), suggesting that something(s) happened midway in the history of the industry (1940s to today) that negatively influenced new drug discovery productivity.

I hope that the SCB report is only the beginning of a more careful analysis of historical trends in drug discovery productivity, not the last word. Settling for the conclusion “that’s just the way it’s always been”, doesn’t offer much hope.

Permalink to Comment

2. Rick on June 28, 2011 11:21 AM writes...

FYI, according to the FDA and a 2006 US GAO report on drug discovery productivity, that spike in 1996, which you mentioned and the report partly attributes to new HIV treatments, was actually entirely explained by cleaning up a backlog of approvals due to PDUFA. In essence, it was a change in the rules, not a one-disease-one-year break in the decline of new drug discovery that yielded the spike.

Permalink to Comment

3. Frank Adrian on June 28, 2011 12:43 PM writes...

It's not just pharmacological R&D. It's happening in computers, in physics, in materials science...

The issue is that we're reaching the limits of our abilities to control the world once again. Silicon chips are reaching limits in their heat dissipation, physics experiments become ever more expensive and pushing into realms that we have no way to influence, the main thing that genomics has taught us is that biochemical pathways really aren't suited for simple control, etc. Nanotech might push things forward a bit more, but it's still more a lab curiosity than a technology.

The only good thing coming out of this is that one still gets to call Ray Kurzweil a nut... How's that singularity coming, Ray?

Permalink to Comment

4. Mike W on June 28, 2011 12:48 PM writes...

5, or 10, or 15 drugs per billion dollars? If EVERY ONE of our programs were successful, we couldn't come close to that today.

Permalink to Comment

5. bbooooooya on June 28, 2011 1:12 PM writes...

No, this graph must be incorrect. According to an article in the Wall Street Journal ("Human Element: Drug Industry's Big Push Into Technology Falls Short; Testing Machines Were Built To Streamline Research -- But May Be Stifling It; Officials See Payoff After 2010", Peter Landers. Wall Street Journal. (Eastern edition). New York, N.Y.: Feb 24, 2004. pg. A.1).

According to a Martin Mackay (a VP at PFE, I'll bet he's done great things): "Martin Mackay, a senior vice president at Pfizer's research labs, said recent data show a higher percentage of compounds at Pfizer are making it through each stage of testing, suggesting that the effort to improve the technology is working. "The proof of the pudding will be in 10 years' time," he said. "We're very confident.""

To be fair, this type of scale back in spending after no real benefit (for a, dated but still relevant, example:, is not surprising.

Partly, as mentioned, easiest diseases have drugs (though cancer remains a scourge) and partly investors will only shovel so much money into projects with no return before taking their capital elsewhere.

One can blame bad management (or lazy workers or those durned foreigners), but it could just be decline of a mature industry. To be fair, when I worked in biotech I saw lots of bench-level chemists jumping to combi chem as though it were manna from heaven. It wasn't.

Permalink to Comment

6. TJMC on June 28, 2011 1:31 PM writes...

Nice graph, but it apparently misses out one of the key confounders of such analyses. That is, the use of NME as "New Molecular Entity" approved (in this case adding in all the non-innovative follow-on drugs that are extremely similar in structure.) I instead call this "nme", for the reasons below.

"nme's" are different from what I and others before me have called "New Medical Entities", intending to describe the first in a novel drug family.

Why is this important? Well, the innovator that finds a new target, develops a NME (Medical) therapeutic, and gets is approved, is truly innovative and hopefully reaps the benefits of all that thinking and work. Other firms watching the innovator may improve on the NME incrementally, and hopefully benefit in proportion to the value the improved outcomes.

I would as Rick mentions, welcome having BCG improve the analysis to focus on real innovation vs. Y2K monies over the time period. I have been looking at the effects of mergers, operational excellence transforms and strategic shifts on NME productivity. Hopefully looking forward instead of continually backward. Do we really think that how R&D was done 20, even 10 years ago is what we do or need now? What we need now are indicators of what is working and creating the innovation and productivity our reality demands now. (see )

Retrospectives do not paint a better picture than the above chart, but the eternal optimist in me follows Faust's core principle; (sic) "Those that keep striving are not damned, but will see redemption". Perhaps this drop in R&D spending is instead a auger of improved productivity that will appear later (maybe after the 5-9 year R&D development cycle)? One can hope...

Permalink to Comment

7. Rick on June 28, 2011 1:46 PM writes...

You're comment about chemists jumping to combichem could be applied to similar jumps to other hot new technologies. In addition to jumping TO things of unproven (even today for some of them), they alsp jumped AWAY FROM techniques that still worked, but were declared prematurely dead by bad management and merchants of newtech.

Permalink to Comment

8. Mike Hamilton on June 28, 2011 1:50 PM writes...


I think your analysis is missing an important factor: regulatory pressures.

Could you imagine in the 1970's that a drug that is a combination of two approved drugs (I'm thinking Contrave/Qnexa here) would have been rejected by the FDA on safety concerns?

I don't disagree with your assessment that new drugs have to compete with existing generics, but I don't think you can ignore the incredible hurdles that the FDA has setup for new drug approval. These are hurdles that were never faced 20+ years ago.


Permalink to Comment

9. Chemjobber on June 28, 2011 2:17 PM writes...

Excellent 80's rock reference, btw.

Permalink to Comment

10. TJMC on June 28, 2011 2:18 PM writes...

After my last comment, I was curious to get up from the details and look down from the 50,000 foot level (like BCG was doing), but at NME's (and not "nme's") Surprising to me was that there might be an inflection moving in the positive direction in NME/$B spent. However, I realized it could also be an artifact from not applying "lag".

I like to lag such analyses so that the output of NME reflects the spending of 5-7 years before approval. Unfortunately, my data for firms' R&D spending prior to 2000 is a bit spotty. Happy to post results but would really prefer to use a uniform spend in R&D from some "reputable" source, IF READERS CAN POINT TO ONE. Something like what BCG used to go back to ~ 1992-4 would work.

Finally, my data winnows out the NME's from the follow-ons, seeing perhaps a differing pattern appearing. Possibly the truly innovative R&D metrics are hidden by that "chaff". In other words, over the past decade, the FDA became loath to approve marginally better drugs with lesser use/safety profiles. So part of the reduction of "nme's" in the past decade could be caused by that bias (which in the past would allow for a larger proportion of such follow-ons.)

Permalink to Comment

11. TJMC on June 28, 2011 2:25 PM writes...

My post # 9 makes more sense if my prior post was shown. In short, BCG and the chart describe "New Molecular Entities" approved. Less useful - especially in this decade.

I use the lower case "nme" for these instead of "New Medical Entities" ("NME's for real) that focus on first in a chemical class or family that is novel for a disease target. These innovators (hopefully) reap far more value than the "nme" follow-ons (which do serve a purpose if efficacy/safety is truly improved.)

Permalink to Comment

12. And D on June 28, 2011 2:28 PM writes...

Some analysis of why new drugs fail to make it to the market compared to the 1970s/1980s might be useful - are we really seeing more failures for "regulatory" reasons, or are there other factors which dominate?

Permalink to Comment

13. bbooooooya on June 28, 2011 3:00 PM writes...

"Could you imagine in the 1970's that a drug that is a combination of two approved drugs (I'm thinking Contrave/Qnexa here)"

Was not in dd 70s so don't know, but I hope (Contrave for sure) would have been viewed as a BS application hail mary by a bunch of execs wanting to perpetuate their salaries for a few more years even back then.

Nothing is stopping MDs from prescribing these combinations currently for weight loss.

Permalink to Comment

14. J. Peterson on June 28, 2011 3:12 PM writes...

...We don't need a new drug. We need a new disease.

Wow. That's the scariest thing I've read in a blog for a long time. What's to stop an evil drug company from secretly releasing some sort of pathogen that only their patented cure effectively treats?

If nothing else, you've got the plotline for a great sci-fi thriller.

Permalink to Comment

15. TJMC on June 28, 2011 3:15 PM writes...

bbooooya #13:

I agree that some of those combinations are a "hail mary pass" from the minds of some marketing executives. However, if there is rigor in the clinical trials in efficacy and safety to get such approvals, then that is better for society.

We have far too many drug-drug interactions that cannot be discerned from standard postapproval monitoring. Especially with longer term effects on heart tissue, etc. that require specified expensive testing.

Permalink to Comment

16. NoDrugsNoJobs on June 28, 2011 3:27 PM writes...

Geez, and I thought the declining productivity was due to direct to consumer marketing and Vioxx, what a shock that other factors might be at work!

Permalink to Comment

17. NoDrugsNoJobs on June 28, 2011 3:29 PM writes...

Geez, and I thought the declining productivity was due to direct to consumer marketing and Vioxx, what a shock that other factors might be at work!

Permalink to Comment

18. bbooooooya on June 28, 2011 3:31 PM writes...

"Could you imagine in the 1970's that a drug that is a combination of two approved drugs (I'm thinking Contrave/Qnexa here)"

Was not in dd 70s so don't know, but I hope (Contrave for sure) would have been viewed as a BS application hail mary by a bunch of execs wanting to perpetuate their salaries for a few more years even back then.

Nothing is stopping MDs from prescribing these combinations currently for weight loss.

Permalink to Comment

19. Anonymous on June 28, 2011 3:34 PM writes...

I don't particularly agree with the term "disappointing research productivity". Most people in the labs are working their asses off and producing as fast as they can.
Sooo, if it's not the researchers...

Permalink to Comment

20. You're Pfizered on June 28, 2011 3:49 PM writes...

I don't particularly agree with the term "disappointing research productivity". Most people in the labs are working their asses off and producing as fast as they can.

Yes, the terminology used to describe the difficulty of finding suitable drug-like compounds is incredibly irritating. It sure makes it sound as if pharma researchers are sitting around collecting paychecks and not working very hard. Given the current environment, diminished staffing due to layoffs and mounting pressures to do more with less so you can keep your job, it's nowhere near the truth.

What we do is some of the most difficult work there is, and yet all we ever hear about is 'lack of productivity' as if we're satisfied with failing most of the time. Very frustrating.

Permalink to Comment

21. Matthew Herper on June 28, 2011 5:12 PM writes...

@19 -- Nobody's disappointed in the bench scientists, or at least, I'm not. I think the bench scientists are often heroic.

But the productivity is disappointing. How should I say it?

Permalink to Comment

22. hibob on June 28, 2011 5:20 PM writes...

Does anyone have anything like a chart of compounds tested vs NMEs approved for the past few decades?

Permalink to Comment

23. Rick on June 28, 2011 5:53 PM writes...

hibob (#22), I created a chart like that a few years ago, using data from the Society for Biomolecular Screening (which is named something else now), but I can't find it now (probably lost a couple of computer crashes ago). Suffice it to say that if you think the decline in NMEs per billion dollars spent per year is depressing, looking at the number of NMEs per billions of compounds screened per year might drive one to suicide.

But I'll do you one better, you can create lots of charts, based on publicly available data, looking at new drugs discovered per new XYZ per year (e.g. per genome sequenced, per crystal structure solved, per computational chemistry software package...) - I kind of made it a neurotic hobby the past few years - and they generally point to the same stunningly counterintuitive conclusion: the harder we try or the more we spend, the worse we get at discovering new drugs. That's right, output correlates INVERSELY with effort/money! It seems bizarre unless you consider one of two options: the problems are just that refractile to solutions or the efforts are misdirected.

Permalink to Comment

24. Rick on June 28, 2011 5:59 PM writes...

TJMC (#s 6, 10, 11) You are spot on that it makes a big difference how you define the numerator in the formula (NMEs, nmes, novel drugs). In the analysis I mentioned in my post #1, the focus was on truly novel molecules, based on a published analysis of the IDDB by a medicinal chemist, not the FDA's terminology. I think that's one reason why I saw the number of novel drugs discovered per year per billion dollars was actually relatively stable between the late 1950s and the mid-1980s at around 12 drugs per billion per year, whereas the chart in this blog suggests a steady decline. It's an important distinction if you're really interested in analyzing innovation.

Permalink to Comment

25. Curious Wavefunction on June 28, 2011 6:09 PM writes...

-We don't need a new drug. We need a new disease. Or several.

I am sure that with the rise of antibiotic-resistant bugs and the neglect of basic antibacterial drug discovery, our wishes will unfortunately soon be fulfilled.

Permalink to Comment

26. bbooooooya on June 28, 2011 7:22 PM writes...

"We don't need a new drug. We need a new disease. Or several"

Restless leg syndrome (RLS)

Erectile Dysfunction (ED)

Gastroesphogial reflux disease (GERD)

Low Testosterone (Low-T)

Attention deficit hyperactivity disorder (ADHD)

Dysphoric Social Attention Consumption Deficit Anxiety Disorder (DSACDAD)

The marketing guys have been pretty productive, actually.....

Permalink to Comment

27. Anonymous on June 28, 2011 7:49 PM writes...

is outsourcing to China/India a big factor?

Permalink to Comment

28. drug_hunter on June 28, 2011 8:47 PM writes...

Guys, guys, c'mon, its just (a) generics raise the bar -- new drugs have to keep getting better and (b) tougher regulatory & payer environment. We're not getting more stupid; our jobs are tougher now. Yes, mergers and other idiocy makes life (much) tougher; but IMHO even Paul Janssen would be struggling in the current climate.

Permalink to Comment

29. TJMC on June 28, 2011 9:40 PM writes...

Rick - Your post #23 cites publicly available data. Can you point to a reliable one for overall R&D spending for past 15-20 years? As I mentioned in #10, looking at truly innovative NME's may actually show a upturn in NME/$B - starting near the 2008 date of your last analysis. I mostly look at individual company yearly reports to try and parse out the marketing spends that "hide" in R&D budgets. - Terry

Permalink to Comment

30. Anonymous on June 28, 2011 10:53 PM writes...

@21 How should you say it? How about not with the framing of the business community? I get the distinct impression that reporters and columnists are not talking to the people who actually work in the labs. I'm not sure why that is.
For the past 15 years or so, the business and consultant class have been getting it disastrously wrong. Oddly enough, beating up on research does not improve morale nor does it appear to be improving productivity.
OTOH, I doubt that you're going to hear from people who are still in the trenches. They're all scared to death that they're next. And yes, they're stressed out and overworked beyond belief.

Permalink to Comment

31. CMCguy on June 28, 2011 11:14 PM writes...

#21 Matt Herper how about New Age PC Spin "less success in overcoming huge obstacles to introduction of a medicinal products"? Majority of Science/Research is about dealing with failure so if view only as "numbers" as has been tendency it can be disappointing where the rare successes get lost.

Permalink to Comment

32. Hibob on June 29, 2011 12:21 AM writes...

@Rick #23: Thanks Rick - that's what I would expect to see for all of the data series that start to explain why the dollars per NME plot is so dismal: molecules screened per NME, clinical trial patient years per NME, researcher man hours per NME, etc. The science didn't get stupider - the problems got harder. Do you have any suggestions for data sources?

@ Matt #21: How to say it: "Our scientists are more productive and have keener insight now than ever before. They can now say "Oh No. That's not gonna work at all" hundreds of times more often per year than they could in the past. " Also, Agile-Lean-Taylorization improvements have put the lyophilizer 23 feet closer to the balance,

Permalink to Comment

33. processchemist on June 29, 2011 1:20 AM writes...

"IMHO even Paul Janssen would be struggling in the current climate"

In the current climate aspirin wouldn't have been developed.
I suspect that the current model trashes early many compounds of potential interest (CACO-2, in vitro herg and so on).
In a project where we was involved a whole class of compounds demonstrated some cardiac tox issues in vitro. But cardiac tox in vivo (dog) demonstrated a safe profile... In other contexts the class would have been dropped, but our customer had an old school approach (in vivo SAR and so on). The candidate is currently in Phase II.

Permalink to Comment