Drug discovery R&D productivity, a call for action?
I’m working on an analysis of drug discovery productivity and I’d like some feedback and possible explanations for a phenomenon that seems quite clear and disturbing.
1. From 1963-1980, an average of 13 new small molecule drugs that would eventually reach the marketplace, not reformulations or line extensions, were discovered - not launched, discovered - worldwide per year per billion dollars R&D investment (year 2000 dollars).
2. Starting in 1980 and continuing through 2000, the average number of new small molecule drugs discovered that would eventually reach the marketplace steadily declined to approximately 1 drug per billion per year (year 2000 dollars).
In other words, since 1980, a period hailed as one of unprecedented discoveries in drug discovery and development, innovative productivity has actually dropped more than 10-fold.
More recent data indicate that this trend is continuing, with current estimates being about 0.8 drug (new or otherwise) per billion dollars spent.
I know various sources, including the TCSDD, have described a decline, but I hadn't anticipated something of this magnitude, especially as it coincides precisely with the advent of so many discoveries and technologies that promised to make drug discovery better, cheaper and faster. My sense is that the industry's public explanation has been, "that's just how much it costs to discover a drug these days". That doesn't seem like a very deep or empowering analysis of the situation and it begs the question "WHY does this cost increase so dramatically year after year?"
Is this analysis consistent with what you have heard or read? If so, what do you think could be the fundamental drivers of this remarkable decline and what can be done to prevent and reverse it? All ideas and criticisms would be appreciated.
Good Answers (9)
Del F.
Principal Scientist at API Purifications LLC
Best Answers in: Biotech (11), Venture Capital and Private Equity (1), Government Policy (1), Green Products (1)
Hi Richard,
Your analysis is consistent with my understanding.
The problem you're addressing is not technical in nature. There is no scientific solution to this issue.
The problem is cultural, and the solution will be realised only upon the institution of changes in the spheres of politics and corporate structure.
The decrease in efficiency is due to the imposition of antiquated economic models to the mechanism of discovery. The model says specialisation of labour, labour reduction, and intensive capitalisation lead to greatest efficiency. A $125,000 liquid handler is worth more than a $125,000 senior researcher because the liquid handler can perform thousands of operations per day while the researcher just drinks coffee and adds to overhead. Discovery can be outsourced—preferably overseas—and this will decrease expenses while also encouraging specialisation, which will add to efficiency. Greatest dividends are obtained by exploiting areas as yet undeveloped by competitors. Thus the "magic bullets" that others in this forum have brought up.
All of this is just so much capitalist nonsense.
What we are seeing here is the 21st century scientific parallel to the Great Depression of 1929. The Great Depression was brought about by the capitalistic over-exuberance of the 1920s. Laissez-faire simply does not work, even at the national scale. We see now the inevitable result of applying this broken economic model to our industry.
Creativity and discovery/development/clinical evaluation pipeline integrity and robustness require not compartmentalisation, but cross-functional expertise. This is achieved by having in place talented scientific shepherds at least from earliest development (and better yet from earliest discovery) through product launch. This is no longer the norm, even at big pharma. No benefit to efficiency is realised by compartmentalisation of the effort.
As I noted in response to another of your questions, the expectation of increased efficiency through volume expansion in discovery is a fatally flawed idea. But golly, we sure impress the CEO, don't we? After all, we did all these in-silico SARs--millions of 'em--even needed to borrow the university supercomputer for a couple months. Then we ran three hundred thousand combinatorial reactions. We don't have anything yet, but golly, Mr. CEO, you can see for yourself, it's not for lack of trying...
The problem is that Mr. CEO tied our hands. He laid off two thirds of the scientific staff. We have all these great robots now, producing nothing. We lost a good chunk of chemical and biological expertise. Our expertise is spread out among computer programmers (for the in-silico work and the robots and the LIMS and so on), project managers, and other specialties that have little to do with the business of evaluating a drug's potential.
The solution is cultural and political. We need to terminate the employment of Mr. CEO, and find a leader who understands the requirements of discovery. In silico and combinatorial work have limited utility, and efficiency is a function of scientific expertise, not capital investment. That means having a full stable of scientists, not cold rooms full of tens of thousands of 96-well plates. The next crop of pharma companies in Indonesia, India, Singapore, and China, care not a whit for our silly 18th century economic ideas; they're going to apply 21st century ideas--and soundly beat us every step of the way.
The Solution:
1. Apply useful economic models to pharma.
2. Apply cross-functional scientific understanding to the process of discovery.
3. Exploit natural biochemical diversity, not in silico SAR-based models
It’s not likely to happen—not until Asia is the major producer of NMEs. And by then it may be too late for us old fossils in North America. But it’s what we need to do.
All the best,
Del Freeberg
API Purifications Inc.
I started my career in the pharmaceutical industry as a synthetic/medicinal chemist. Although I have not been part of "big pharma" for quite awhile, I think several things have changed.
The industry moved in two opposite directions at the same time. First, there was and continues to be a fascination with biomolecule induced design of new small molecule drug candidates. The genome project exacerbated this trend with its promise of new drugs and eradication of various genetic caused diseases. All of this led to a need to have carefully predicted and designed mechanisms of action for every drug candidate. Broad screening approaches were virtually eliminated, replaced with combinatorial chemistry which in fact is not broad screening.
On the other hand, the pharmaceutical companies began focusing later and later in the drug development process. Instead of research scientists, the companies began staffing with more and more MD's. No disrespect to medical doctors, but they are not trained to manage nor lead innovative drug design programs.
The overall result is a limited drug design capability due to strict "receptor driven" constraints and a decrease in acceptable disease targets based on drug markets and "acceptable" mechanisms of action.
Active small molecule design is not dead. In my private laboratory, I have designed and synthesized two new families of extremely potent dual targeted HIV/AIDS drugs, a new family of potent antibiotics for multi-resistant TB and quite recently, a "XXXXX" activated, non-toxic oncolytic agents. Based on the numbers in the literature, I should have a spare $3 billion lying around!!!
It used to be that synthetic chemistry was used to design the drug receptor site. Now the receptor is being predetermined by computer modeling, etc. I'm sure that was supposed to save money on synthetic chemistry.
Very interesting information. I believe people forget that the drugs released since 1980 actually have been in the pipeline for a long time before coming to market. I am afraid that, if your data are accurate, we will experience a distinct shortage of new drugs in the years ahead. Pharmaceutical companies already talk about "empty pipelines" and the panic at Pfizer caused by the expiration of the Lipitor patent also is an anecdotal indication that there isn't much in the works that will replace today's medications.
I suspect that this situation is due to two factors. One may be the lack of intensity on the part of the pharmaceutical companies. The pipelines were full, money was coming in and there was no reason to be especially concerned about the future. That is not to say that these companies did not invest in research, on the contrary, they invested a great deal. However, those investments might not have been spent in the most efficient manner.
The other factor might be increased complexity of the bio-medical science. Many more data now are available and every increase in the quantity of data geometrically increases the complexity of the information.
I am hoping that this is a temporary situation which will resolve itself in the years ahead. Pharmaceutical companies are in trouble and they are refocusing their discovery and research efforts and new tools for managing the complexity of the information are becoming available.
BTW, if possible, I would greatly appreciate if you could send me a copy of these data.
I would say that the analysis you cite is within the ball park of other analyses I've read on the topic of R&D productivity. As to your second question, I'd like to offer a few (certainly not a complete set) ideas about what drivers got us into this situation.
1) It should not come as a suprise that finding effective and safe treatments is becoming harder. As the pathophysiology of diseases was uncovered and potential treatments developed, the low-hanging fruit was taken pretty fast. What we are dealing with now are complex, chronic, poly-genic, multi-factorial diseases that are not likely to be amenable to simple pharmacological interventions. In addition, we've had some success with symptomatic treatments, but now the hurdle is to find disease modifiers!
2) Advances in molecular biology and molecular genetics made it pretty simple to isolate and characterize novel receptors, peptides etc. These technical advances were coupled with a belief that if you dissected a disease deep enough, we'd find a single root cause..some defective receptor or mis-folded protein. This led us to neglect cell physiology or as some call it now "systems biology" and become almost unable to deal with disorders that involve physiological systems and cascades.
3) The explosion of molecular and genetic biology then led to the notion that if specific receptors were at the root cause of many disorders, then selective compounds targeted at these receptors should be effective and safe treatments. Since you can create in-vitro systems to test for affinity to many receptors etc, this seduced pharma into thinking we could industrialize the process of identifying potent and selective compounds.
So what have we learned from these drivers...the diseases that we are most concerned about now (cancer, Alzheimers, CAD) are not simple disorders. Basic research in these areas continues to point out that there are no "single" causes, therefore the notion that we will identify some root cause starts to seem naive. Since the kinds of drugs that are likely to work in these disorders cannot be identified through high-throughput screening of massive libraries against specific targets, we have ended up with massive costs in drug discovery, minimally effective drugs being taken into clinical trials and safety issues emerging once compounds are approved because we never really understood the systems we were trying to interfere with.
Just some thoughts...
I also would add that the genomics/proteomics revolution has not paid the expected dividends and that there was not enough foresight in the early days of combi-chem which accompanied this revolution.
When NME's are considered from 83-2004, it is constant (give or take a few) with the exception of 1991 (1992?) and 1996 which are spiked. When one looks at this in terms of NME's/$billion spent it is >1NME/$billion until 1992...exception again is 1991 and 1996 where we are at something like 3-6/$billion. Possibly representing a healthy backfilled pipeline. looking at 2004, we are down to 0.6-0.7 per billion.
This hints to a lot of focus early on...less targets, low hanging fruit, less tools. After diversification in the 1990's, spending exploded. Unfortunateky the biological validation is still far behind and now the pipelines are dry.
this is compounded with pantent expiration and much tougher regulatory considerations (FDA being much more risk averse)--the regulatory issue can be seen with the number of compounds entering phase I and compare that with what enters Phase III--60% drop off.
This is a complicated problem with no one answer or solution, but it is clear with the direction that pharma is taking that they are looking outside of their internal R&D efforts to solve it.
Clarification added September 11, 2008:
I would also add, that this lack of success is not really a lack of innovation...it is just a gap. there are ~1500 compds per yr going into phase1 and this increases ~10% each year since ~2004. Also, there are several thousand (at least) compounds in development.
the 1990's was also not wasted, it just hasn't filled the pipelines yet. if you look at the top selling drugs of 2007, only ~20 small molecules were approved since 2000, the remaining ones (SM's) were produced in the 1990's; so the 90's did not fill the pipeline really for the 2000's.
the answer concerning disease complexity is certainly accurate, and this was underestimated. but i think it is not fair to catagorize this is a true lack of innovation.
The entire industry has been changing and will continue to change and we need to catch up with respect to regulatory and other issues. if one takes a close look at this, it is not all bad.
Ashley P.
Director of Regulatory Affairs at Takeda Global Research & Development Center (Asia) Pte Ltd
Best Answers in: Government Policy (1), Biotech (1)
Hi Richard,
A very interesting question. I believe the scenario you describe is the result of a number of contributing factors, including:
1) Increased regulation of pharmaceutical/biotech products - Since the passage of the Food and Drug Administration Act in 1988, there has been an ever-increasing level of regulatory scrutiny applied by FDA to pharmaceutical and biotech products. A similar trend concurrently occurred in Europe, Japan and most western markets. Whilst this increased regulatory scrutiny is of course necessary to ensure patient safety, the increased level of safety, efficacy and quality evidence required for product development and marketing translates directly into increased cost to the pharma/biotech companies.
2) Increased competition – Since 1980, and particularly during the ‘biotech boom’, there are many more pharma/biotech companies of all sizes, all competing for IP, all developing new, improved products in most therapeutic areas. Improved products in any given therapeutic area effectively ‘raises the bar’ for any new product coming through development. This is particularly evident in the treatment of HIV, where the original anti-HIV drugs were effective, but quite toxic in the medium- to long-term. At that time however, the risk:benefit argument for patients was clear, given that without treatment their life expectancy was measured in months. As the efficacy and safety of new anti-HIV drugs increased over time, HIV shifted from an acutely life-threatening disease, to a chronic disease, with patients’ average life expectancy now around 25-40 years post-infection. As a result, regulators expect new HIV drugs to have fewer long-term side-effects, since patients will be using the drugs for decades.
As a result of the increased cost of development of new products, many pharma companies are now moving to establish collaborative agreements with universities and other academic institutions in an attempt to streamline the drug discovery process, by leveraging the strength of universities in the early-stage/discovery research arena. I believe this trend will continue, and eventually we will see far less discovery research conducted ‘privately’ by pharma/biotech companies, who will focus on clinical development of lead compounds derived from academic research institutions.
Moreover, the most effective way to reduce the cost of drug development is to ‘work smarter’. Have a sound development plan and an appropriate regulatory strategy.
Hi Richard,
Several people have already posted well supported comments on the cause of our productivity decline in drug discovery, but I don't believe that we have yet addressed your question on what can be done to prevent or reverse the situation.
Large pharma's R&D teams have been shrinking for several years in response to cost controls, generic erosion and the difficulty of bringing blockbuster medicines to market. In order to support the pipeline, pharma has responded by increasing spend on in-licensing – AZ, GSK and Pfizer have all publicly commented that 50% their pipelines will be sourced externally by 2011. The savings realised from reduced R&D operations have enabled this increased spend, but in-licensing reduces the value that can be realised by the corporation that brings the drug to market.
The future for drug discovery lies in sourcing scientific services externally. We already do this extensively for clinical trials and some areas of pre-clinical. By accessing capability, innovation and knowledge in discovery sciences from external sources, we will create agile and focussed research groups that can rapidly respond to market changes and they will be more productive because they are rewarded according to success. GSK has already started to adopt this approach with their move to the Discovery Performance Units that access finances through a board of internal and external reviewers similar to a Venture Capital model.
You'll be glad to hear that innovation in drug discovery is alive and well, we've just not been accessing it in ways that maximise its value. We need to think about how we operate in a different way, understand the impact of the emerging economies and how we access their talent. And we need to develop new processes that deliver innovation and productivity to a sector that is always facing new challenges.
I very much welcome further discussions on how we can adapt to this new model.
Best Regards
David
Roberta W.
Group Director at Teva Pharmaceuticals
Best Answers in: Biotech (13), Ethics (3), Venture Capital and Private Equity (1), Compensation and Benefits (1), Change Management (1), Social Enterpreneurship (1), Quality Management and Standards (1), Market Research and Definition (1), Engineering (1), Career Management (1), Wireless (1)
Hi Richard:
Here are my two cents about the slowdown in pharma approvals (and also potential issues with drugs once they get to market)
1) Drug development really has become more expensive and it really does take longer to get a drug to market. The FDA and EMEA are requiring larger trials, asking for additional types of testing to be done and requiring more post-marketing testing than ever before. National agencies and physician organizations are asking for head-to-head trials to improve information for choosing drugs. While this may ensure that marketed drugs are safer once they reach market, it has added to the expense and time of development.
2) Many patients don't want to participate in drug trials. I can understand this; being in a trial (I've done it) takes a ridiculous amount of time and schedules are never arranged around normal working life. Since many trials depend on "professional" volunteers, I think we often don't get the most realistic clinical data in pharma trials. I'm not sure if this impacts on what gets to market and how the drugs work once they get to market but it may. Since this also makes the pool of subjects for a trials smaller, it extends the study time which makes the whole process more expensive.I think that the public has to take an interest in and support research on drugs that are supposed to help them. If you choose not to ever be a test subject, you are choosing not to support development of drugs that may help you.
3) Government spending for scientific discovery has been reduced over the past 20 years. Our current administration seems bent on turning the US into a third world nation with respect to science education and research funding. Industry research has always been more narrow in scope than academic research and the two are complimentary. If we continue to slash government funding for academic research, we should not be surprised that the feed-in to industry is reduced and that talent perceives that there are better places to earn a living than in scientific/medical research.
4) We need to understand disease processes better. We now know that these are much more complex than we thought they would be after some of the big scientific breakthroughs in the 1990's. We have learned that animal models don't replicate human disease very well, that diseases behave differently in real life than in cultured cells and test tubes, that outbred humans are different than inbred laboratory animals, and that certain receptor/ligand interactions may be very important to system functioning,but they are not the absolute requirements we thought they were or there are redundancies that can "pinch hit" for them in certain cases. We were being too reductionist and human systems that are clever enough to have survived onslaughts for centuries are much more complicated than we thought. Unfortunately so are the disease processes that affect us.
5) Dave's comment about the genomics/proteomics revolution not panning out yet is true, but I think this goes back to #3 and 4. In oncology there have been major breakthroughs and drugs approved around tumor markers (for example her2); biomarkers have not yet been as productive in immunology (my area of research). I think we may still be looking at the wrong things in some fields of medicine because we simply don't know what to look at yet.
6) We need to control our own behavior as it affects our health. People eat too much of the wrong foods, don't exercise enough, smoke, and have stress producing lifestyles and expect pharmaceutical research to come up with pills to correct all of this (without side effects, of course). Some of the biggest epidemics (obesity, type II diabetes, COPD) are self-created. Patient non-compliance is an issue with control of many diseases. Why do we ask the medical/pharma industries ito keep us healthy instead of taking responsibility for factors we can control ourselves?
Roberta
Hi Richard,
There is no reason to believe that the number of discovered new chemical entities (NCE) is small.
According to your data starting in 1980 and continuing through 2000 the average number of new small molecule drugs discovered that would eventually reach the marketplace steadily declined.
There are several reasons for that. One big reason: In the 80s and 90s the trend in the industry was to consider the development of an NCE molecule only if it is at least a billion dollar drug, though that trend is markedly declining lately.
Criteria for acceptable safety profile has changed dramatically for the drugs during this period and still changing. There are a couple of examples here;
Cardiovascular safety evaluation of an NCE (potential to prolong QT interval) was not a requirement then. Now it is. This is an added cost and the cost of these evaluations is high. This killed several drugs, most of them in early human trials, some in late stages of development, and some post approval as well.
The other examples: they need to be tested for potential for abuse, potential for sleepiness, long term safety (now it is necessary to study them for longer periods, in some cases 2 to 5 years)
The other reasons the cost is going up:
Companies want to do the trials fast; time is money. This approach right away increases the cost.
We cannot overlook the obvious, inflation. Needs I go on?
The pharmaceutical/biotech companies and the regulators are between a rock and a hard place now. They are expected to get the best drugs into the market without any adverse safety profile very quickly with the least expenditure.
The data you have is well put together, but they obviously did not include a lot of other contributing factors and trends in that.
As far as the “discoveries and technologies that boast to make drug discovery better, cheaper and faster” go, majority of them are helpful in weeding the NCEs at an early stage.
More Answers (5)
Russell W.
Guest lecturer - Macroeconomics. IT Management for the Enterprise at University of Kentucky
Best Answers in: Enterprise Software (2), Using LinkedIn (2), Ethics (1), Computers and Software (1), Software Development (1)
How about some drugs that actually cure something?
Roughly three months ago I was wondering about the efficacy and the economy to continuously develop "new" cell lines for R&D purposes. Your analysis apparently confirms my naive feelings.
Links:
Basicaly the whole problem lies in the testing n clinical trials which take a long period of time.. just imagine what a person will undrergo who has devoted the best decade of his life for the research on a drug n it takes 6-7 years to testify with a succes rate of less than 2% n even then it "may" b launched in d mkt. depending upon the mkt forces n the will of investors..
all this effects the interest of the scientist at the root level n they prefer to go to other fields instead. so the talent deserved in this field is distributed in othr fields..also if the drug is against an infection then such a log time may result into ineffectiveness of the drug due to constant mutations in the target organism.
THE POSSIBLE SOLUTION for this z to make the testing procedure more friendly n convinient.. this may b achieved by using computer based testing procedure (which z in developmnt phase right now) We can call this a virtual biosystem this would include the information of the human genome(s) n the protiens coded (and other biomolecules) in humans (it needs to be versatile in this respect) by this the drug structure could be tested against various biomolecules present in human body to check its rxn with them and thus to predict the type and extent of side effects (if any) n also it could be tested in wet lab aganst the varios bioelements present in human body to further analyse n confirm the results. this would certainly reduce the time period n moreover the scientists could easily check the feasiblity of the drug they have discovered n also they may change the structure in accordance with the requirement of biological systems n thereby increasing the success rate manyfold!!
--Inputs would b high amounts of research at gene level n enormous database would have to be managed
--Output would be better n efficient drugs in less time with less cost..
