Corante

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: derekb.lowe@gmail.com Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Emolecules
ChemSpider
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
PubChem
Not Voodoo
DailyMed
Druglib
Clinicaltrials.gov

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
Kilomentor
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
ChemBark
Realizations in Biostatistics
Chemjobber
Pharmalot
ChemSpider Blog
Pharmagossip
Med-Chemist
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
SimBioSys
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Business|Bytes|Genes|Molecules
Eye on FDA
Chemical Forums
Depth-First
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
FuturePundit
Aetiology
Gene Expression (I)
Gene Expression (II)
Sciencebase
Pharyngula
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net


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


Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem


Politics / Current Events
Virginia Postrel
Instapundit
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

« Reproducing Scientific Results - On Purpose | Main | More On Reproducing Scientific Results: Organic Chemistry Edition »

August 15, 2012

A Quick Tour Through Drug Development Reality

Email This Entry

Posted by Derek

I wanted to let people know that I'm working on a long, detailed reply to Donald Light's take on drug research, but that I'm also looking at a few other publication venues for it. More on this as it develops.

But in trying to understand his worldview (and Marcia Angell's, et al.), I think I've hit on at least one fundamental misconception that these people have. All of them seem to think that the key step in drug discovery is target ID - once you've got a molecular target, you're pretty much home free, and all that was done by NIH money, etc., etc. It seems that these people have a very odd idea about high-throughput screening: they seem to think that we screen our vast collections of molecules and out pops a drug.

Of course, out is what a drug does not pop, if you follow my meaning. What pops out are hits, some of which are not what they say on the label any more. And some of the remaining ones just don't reproduce when you run the same experiment again. And even some of the ones that do reproduce are showing up as hits not because they're affecting your target, but because they're hosing up your assay by some other means. Once you've cleared all that underbrush out, you can start to talk about leads.

Those lead molecules are not created equal, either. Some of them are more potent than others, but the more potent ones might be much higher molecular weights (and thus not as ligand efficient). Or they might be compounds from another project and already known to hit a target that you don't want to hit. Once you pick out the ones that you actually want to do some chemistry on, you may find, as you start to test new molecules in the series, that some of them have more tractable structure-activity relationships than others. There are singletons out there, or near-singletons: compounds that have some activity as they stand, but for which every change in structure represents a step down. The only way to find that out is to test analogs. You might have some more in your files, or you might be able to buy some from the catalogs. But in many cases, you'll have to make them yourself, and a significant number of those compounds you make will be dead ends. You need to know which ones, though, so that's valuable information.

Now you're all the way up to lead series territory, a set of compounds that look like they can be progressed to be more potent and more selective. As medicinal chemists know, though, there's more to life. You need to see how these compounds act on real cells, and in real animals. Do they attain reasonable blood levels? Why or why not? What kinds of metabolites do they produce - are those going to cause trouble? What sort of toxicity do you see at higher doses, or more long-running ones? Is that related to your mechanism of action (sorry to hear it!), or something off-target to do with that particular structure? Can you work your way out of that problem with more new compound variations without losing all of what you've been building in so far? Prepare to go merrily chasing down some blind alleys while you work all this stuff out; the lights are turned off inside the whole maze, and the only illumination is what you can bring yourself.

Now let's assume that you've made it far enough to narrow down to one single compound, the clinical candidate. The fun begins! How about formulations - can this compound be whipped up into a solid form that resembles a real drug that people can put in their mouths, leave on their medicine cabinet shelves, and stock in their warehouses and pharmacies? Can you make enough of the compound to get to that stage, reliably? Most of the time the chemistry has to change at that point, and you'd better hope that some tiny new impurities from the new route aren't going to pop up and be important. You'd really better hope that some new solid form (polymorph) of your substance doesn't get discovered during that new route, because some of those are bricks and their advent is nearly impossible to predict.

Hey, now it's time to go to the clinic. Break out the checkbook, because the money spent here is going to make the preclinical expenses look like roundoff errors. Real human beings are going to take your compound, and guess what? Of all the compounds (the few, the proud) that actually get this far, all the way up to some volunteer's tongue. . .well, a bit over ninety per cent of those are going to fail in trials. Good luck!

While you're nervously checking the clinical results (blood levels and tolerability in Phase I), you have more questions to ask. Do you have good commercial suppliers for all the starting materials, and the right manufacturing processes in place to make the drug, formulate it, and package it? High time you thought about that stuff; your compound is about to go into the first sick humans it's ever seen, in Phase II. You finally get to find out if that target, that mechanism, actually works in people. And if it does (congratulations!), then comes the prize. You get to spend the real money in Phase III: lots and lots of patients, all sorts of patients, in what's supposed to be a real-world shakedown. Prepare to shell out more than you've spent in the whole process to date, because Phase III trials will empty your pockets for sure.

Is your compound one of the five or ten out of a hundred that makes it through Phase III? Enjoy the sensation, because most medicinal chemists experience that only once in their careers, if that. Now you're only a year or two away from getting your drug through the FDA and seeing if it will succeed or fail on the market. And good luck there, too. Contrary to what you might read, not all drugs earn back their costs, so the ones that do had better earn out big-time.

There. That wasn't so easy, was it? And I know that I've left things out, too. The point of all this is that most people have no idea of all these steps - what they're like, how long they can take, that they even exist. It wouldn't surprise me if many people imagine drug discovery, when they imagine it at all, to be the reach-in-the-pile-and-find-a-drug process that I mentioned in the second paragraph. Everything else is figuring out what color to make the package and how much to overcharge for it.

That's why I started this blog back in 2002 - because I was spending all my time on a fascinating, tricky, important job that no one seemed to know anything about. All these details consume the lives and careers of vast numbers of researchers - it's what I've been doing since 1989 - and I wanted, still want, to let people know that we exist.

In the meantime, for the Donald Lights of the world, the Marcia Angells, and the people who repeat their numbers despite apparently knowing nothing about how drugs actually get developed - well, here are some more details for you. The readers of this site with experience in the field will be able to tell you if I haven't described it pretty much as it is. It's not like I and others haven't tried to tell you before.

Comments (60) + TrackBacks (0) | Category: Drug Development | Drug Prices


COMMENTS

1. watcher on August 15, 2012 9:38 AM writes...

Acutally, it's not just target ID that often is prsented as the "most important key step". Many people along the continuum, from screening to med chem to biology to ADME to tox to formulation etc. think of themselves as the key to "making the drug". Of course, each is right in the sense that each can play a critical role. But those with experience, open eyes and who WANT to really be involved in the approval for new drugs recognize it is the effective integration of all these disciplines is necessary. The misperception is perpetuated by government groups (eg NIH) as pointed out, and also by hype presented around Biotechs who set objectives of "churning" to make money for investors, not to really take a process all the way for launching new medicines.

Permalink to Comment

2. Jim on August 15, 2012 9:46 AM writes...

I've also often found that people are extremely surprised when they hear about how IP and chemical space issues can delay programs and drain resources. Those challenges that cannot be beaten by working smarter or working harder are ones that many people are unaware of and clearly add another layer of difficulty to this process.

Permalink to Comment

3. Nick K on August 15, 2012 9:56 AM writes...

Donald Light has become remarkably reticent of late. Perhaps he has been kidnapped by the evil minions of Big Pharma to silence him.

Permalink to Comment

4. transposition on August 15, 2012 9:56 AM writes...

Bravo! This is the kind of post that I can show to my former colleagues in academia, just so that they understand the true complexity of drug development.

As a clinician and former biologist, I even take issue with those who believe that a target ID is a good start. How many great targets have been well inhibited, only to find out that the disease we are interested in treating has found a way around our artful trap?

Permalink to Comment

5. transposition on August 15, 2012 9:56 AM writes...

Bravo! This is the kind of post that I can show to my former colleagues in academia, just so that they understand the true complexity of drug development.

As a clinician and former biologist, I even take issue with those who believe that a target ID is a good start. How many great targets have been well inhibited, only to find out that the disease we are interested in treating has found a way around our artful trap?

Permalink to Comment

6. CMCguy on August 15, 2012 10:09 AM writes...

Derek can you do me a favor and put a Warning Line after the header when ever you are going to mention Angell. Her (along with her husband's) misconceptions have been propagated for many years and I now have a automatic response to her name that probably could use a blood-pressure control drug but at least the Warning would give me a chance to do relaxation exercise before reading.

You write a great summary here covering many main areas for R&D however unfortunately such rational arguments rarely seem to have impact on phrama critics who as you indicate appear to have little actual knowledge of what occurs in drug development.

Permalink to Comment

7. Doug Steinman on August 15, 2012 10:31 AM writes...

Derek,
Great post although I think you did leave out some aspects of the research cost but, as you say, those costs are a relatively small piece of the total drug development picture. Now, how do we get this into the mainstream media? Do we have any readers of this blog who have a connection that would give this the exposure that it deserves? Would BMJ be willing to publish it as a response to the Light article? Would anyone from senior pharma management be willing to make a public statement concerning this issue? I know that's a lot of questions without a lot of answers but this issue is not going to go away by itself.

Permalink to Comment

8. Myma on August 15, 2012 10:33 AM writes...

R and then D. Its not like rhythm and blues which happen simultaneously in the genre known as R&B. There is R, and then after R is done, there is D. And that D part of R&D takes as long if not longer as the R part.

Permalink to Comment

9. MadScientist on August 15, 2012 10:34 AM writes...

Nicely done as usual, Derek.

Of course, while those not involved in drug R&D will be blown away by the complexity described in your article, the rest of us know that this cornucopia of hurdles represents just a small tip of the iceberg.

If only it were really this simple!

Permalink to Comment

10. bluefoot on August 15, 2012 10:35 AM writes...

And even this post makes it sound easy. What about those "truly innovative" drugs that Light et al are harping on? New mechanism of action? First-in-class compound? The amount of time and money you'll spend validating the biology of the mechanism/target or the specificity, metabolism and toxicity of the compound - before you even *think* about trials - will make the most seasoned professional wince. What if your novel target/mechanism requires a new assay system or new animal model to demonstrate efficacy?

I could go on....

Permalink to Comment

11. johnnyboy on August 15, 2012 10:53 AM writes...

what bluefoot said - your excellent summary hints at the complexity of the task, but to accurately describe all the hoops a molecule has to go through, you'd need several hundred pages. Just my own area of preclinical tox, which you've briefly alluded to, represents at least a dozen GLP-compliant studies, years of work (a carcinogenicity study takes 2-3 years to complete), and the skills of at least a hundred different people. And even with that, the total cost of preclinical tox is a drop in the bucket of the overall R&D cost of a drug.

Permalink to Comment

12. AndrewD on August 15, 2012 11:04 AM writes...

Derek- Of course once the drug is out in the real world and being used by large numbers of people, a rare side effect is found and the drug is pulled because of public misunderstanding and political pressure and all the money spent is lost

Permalink to Comment

13. Twelve on August 15, 2012 11:06 AM writes...

Well said. Drug discovery has many of the less desirable aspects of betting on a lottery, except that you usually have to keep buying more tickets to stay in the game, the ticket prices keep going up, it takes >10 years to find out if you've won, and nine times out of ten it turns out there was no prize anyway!

I sometimes explain the relationship between academic basic research and industrial drug research with an analogy to the IT business, which is the background for many well-meaning critics of pharma: It could be rationally argued that John Von Neumann and Alan Turing created the modern computing industry. It can't be rationally argued that they created the iPad.

Permalink to Comment

14. Chrispy on August 15, 2012 11:21 AM writes...

As you point out, Derek, it seems that one of the biggest costs which is being missed is that of supporting all the things which ultimately fail. If you just look at the costs involved with the development of one successful drug you'll probably be an order of magnitude low.

Permalink to Comment

15. Andy on August 15, 2012 11:29 AM writes...

Left out some significant post-lead preclinical work. Activity in animal models of your disease, non-GLP tox studies, GLP tox studies. GLP/GMP process. All the mandatory regulatory paperwork.

Permalink to Comment

16. Anonymous on August 15, 2012 11:34 AM writes...

Unfortunately, people that want to remain ignorant will do so. And often they have an audience that will also choose to do so. That's how people like Light and Angell can continue to propagate their ill-conceived opinions as fact.

Permalink to Comment

17. JC on August 15, 2012 11:40 AM writes...

You also have to have the back-up, the back-up in another class, the metabolites & the deuterium analogs

Permalink to Comment

18. John on August 15, 2012 11:50 AM writes...

Has anyone pointed out the annual operating costs of a Schreiber or Bertozzi lab to professor Light? I think he might find it interesting to compare their annual costs with his figure for bringing a drug to market.

Permalink to Comment

19. anchor on August 15, 2012 11:57 AM writes...

Derek: as always great post! While all the above are true that I have experienced myself in big pharmaceutical company. In addition, I believe in one of those program we also addressed the issues related to "Drug-Drug-Interaction" even as the drug was going through clinical trials. Donald Lights of the world may be informed of this issue that literally consumed us! When I think of drug program, am reminded of snake-chute game, really.

Permalink to Comment

20. Helical_Investor on August 15, 2012 12:13 PM writes...

This topic got some play on a Motley Fool discussion board where I had linked some of your and other posts/articles. To cut to the chase, I made the following comment on Dr. Lights post.

"But ... to invest here for the long term or to believe that the innovation crises is 'FAKE', you either have to believe one of two things. 1) That efficiency in the industry will reverse and climb from here or 2) that the industry will continue to be able to make up for this loss of productivity with price / revenue increases.

I don't and I don't!!!"

Zz (or TMFHelical, or whatever)

Permalink to Comment

21. cookingwithsolvents on August 15, 2012 12:21 PM writes...

Excellent post, Derek. I started coming to your blog for just such insight

I direct all sorts of "R" folks in academics to your site because the overall R&D process and pitfalls are pretty similar across many types of technology, even if the precise details are different.

Permalink to Comment

22. Chemjobber on August 15, 2012 12:39 PM writes...

In case people didn't believe Derek about what Marcia Angell says about pharma, here's a little tidbit from her 2004 book "The Truth About the Drug Companies: How They Deceive Us and What To Do About It" (the Google Books version is linked in my handle):

"The development part of R&D is itself divided into two stages - preclinical and clinical. The preclinical stage has to do with finding promising drug candidates and then studying their properties in animals and cell cultures. Companies keep vast libraries of drug candidates -- molecules that can now be screened very rapidly by computerized methods to see if they will target the Achilles' heel found by the basic research. In addition, new molecules can be synthesized or extracted from animal, plant, or mineral sources. Only the small fraction of drug candidates that make it through preclinical development go on to be tested in humans -- the all-important clinical stage (more on that later.)"

It's that easy, folks.

Permalink to Comment

23. Anonymous on August 15, 2012 12:53 PM writes...

Derek,

I tried to comment on Monday's Light post, but it seems to still be in moderation (too many links, perhaps). I wanted to call it to your attention because it debunked one of Light & Lexchin's specific claims - that pharma's own numbers supposedly prove they only spend ~1.3% of revenues discovering new molecules.

Turns out they base that claim on a 1999 NSF survey that asked pharmas how much $ they used for basic research, applied research, and development. Not surprisingly, the NSF survey defined basic research as projects "for the advancement of scientific knowledge and which do not have specific immediate commercial objectives."


Clearly, screening compounds to find hits against a known target for a specific disease doesn't fit that definition of basic research. Yet Light & Lexchin use those basic research numbers to come up with their claim that "government reports indicate that companies have been spending only 1.3% of revenues on basic research to discover new molecules, net of taxpayer subsidies." (Br. Med. J. 2012)

Full details are in my moderated comment, in case you want to address this point in your planned reply. If the comment is lost, feel free to email me for a resend.

It's not that big an issue, I know, but I fear it may be just one example where their analysis is faulty and leads to erroneous claims.

Permalink to Comment

24. Hap on August 15, 2012 12:58 PM writes...

No one is as blind as the one who will not see.

Dr. Lowe, if you publish your response somewhere else, can you post a link or tell readers where so we can get to it? Thank you.

Permalink to Comment

25. SP on August 15, 2012 2:25 PM writes...

I think something that people tend to not believe is the exponentially increasing costs as you move down the pipeline. As someone on the early side, I can tell you exactly what those parts cost- $W per well to do HTS, $X per FTE chemist doing Y compounds a week, $Z for a CRO to run these ADME assays. Where do the hundreds of millions for phase 3 go, specifically? It's certainly not production of the drug or data handling. Is it salaries for clinical statisticians? Compensation to subjects? Fees to clinical sites? I honestly have no idea, it would be good for someone to lay out specifics within an order of magnitude about where exactly the money goes.

Permalink to Comment

26. clinicaltrialist on August 15, 2012 2:40 PM writes...

As a clinical trialist, let me pitch in and enumerate just some of the things that are necessary after you start clinical trials.

Protocols
Case report forms
Clinical database programming
Clinical database validation
Adverse report forms
Adverse report database programming
Adverse report database validation
Informed consent forms
Translation of ICF
Backtranslation of ICF
Validation of backtranslated ICF
Validation of ICF for appropriate grade level
Sponsor review of each site's modification of ICF before submission to IRB (not necessary in some countries)
Contracts with investigational sites
Regulatory approvals from foreign countries
Scientific review board
Ethics committees
Statistical analysis plan
Data quality plan
Site monitoring plan
Site audit plan
CRO oversight plan
CRO contract
CRO audit plan
CRO audit
Audit follow up and closeout
Feasibility study
Site selection
Site qualification visit
Site initiation visit
Site audit
Manufacturing site audit
Manufacturing agreement
Data management vendor selection and audit
Data management vendor agreement
Pharmacovigilence vendor selection and audit
PV vendor agreement
Randomization system programming
Randomization system validation
Data safety monitoring board formation
Data safety monitoring board charter
DSMB liaison, statistician selection

OK, you get the idea. I don't think this even covers 5% of what's involved.

Permalink to Comment

27. Ed on August 15, 2012 2:46 PM writes...

Well SP, the folk in clinical development at m last employer earned approximately double what a PhD scientist did. So I guess that contract research FTE rates in clinical must be horrendous.

Permalink to Comment

28. clinicaltrialist on August 15, 2012 2:47 PM writes...

In response to #25, the rule of thumb used to be a third to sites, a third to CRO, and a third for internal sponsor costs. The CRO and site costs have gone up a bit.

The costs also be split up by function. The rule of thumb used to be a third for data management, a third for site monitoring (CRAs flying out to sites to hand verify each data point in the medical charts against the data on the case report forms) and a third for everything else. Data management costs have decreased because we use web-based forms rather than paper so it's about 25% data management and 40% monitoring now.

In short, most of the clinical trial costs are FTEs. Boots on the ground.

Permalink to Comment

29. SwedenCalling on August 15, 2012 3:18 PM writes...

Nice post! Very impressive how elegantly you summarized it Derek. One thing you (deliberately?) did not mention was having to defend the your drug project from being closed for 'strategic reasons'.

Permalink to Comment

30. qetzal on August 15, 2012 3:22 PM writes...

Comment #23 was me. Sorry.
(Rassifrassin' iPhone!)

Permalink to Comment

31. MoMo on August 15, 2012 3:36 PM writes...

Excellent! You forgot about the years wasted in research where fellow employees derail solid performing drugs because their Egos get in the way. That'll cost 43M$ alone, Dr. Lightner.

Now go away Lightner et al. and leave drug discovery and development to the professionals.

Permalink to Comment

32. Respisci on August 15, 2012 3:37 PM writes...

To #25 SP
You asked: Where do the hundreds of millions for phase 3 go, specifically? It's certainly not production of the drug or data handling. Is it salaries for clinical statisticians? Compensation to subjects? Fees to clinical sites? I honestly have no idea, it would be good for someone to lay out specifics within an order of magnitude about where exactly the money goes.

Response: I disagree. Drug production can be a chunk of money if you consider that you are making material in quantities far greater than before for Phase 1 and 2. Assume a 5 mg dose per day, in Phase 2 you may treat 100 patients for 3 months, and then Phase 3 you treat 1000 patients for 1 year— now requiring you to manufacture kg of drug. Ask anyone in CMC how easy it is to scale up GMP!

Data handling-Again, think of the volume of data from a 1000 person Phase III study over 1 year (or longer) compared to a 3 month 100 person Phase II study. The increase in the amount of data monitoring, audit trails, verifications even before you begin the analyzing data.

As for the study itself, each test performed at each visit has its associated cost. These can quickly add up to per patient costs of 10K and much, much, much (much, much) higher.

I think Derek’s line of how clinical development costs make “preclinical expenses look like roundoff errors” is completely apt.

And believe me when a clinical study fails or is inconclusive, your heart absolutely breaks when you consider all the hard work, efforts and costs that were invested.

Permalink to Comment

33. okemist on August 15, 2012 3:46 PM writes...

One stupid cost I could not believe was before going into phase I 20 yrs ago, we had to pay 50K for a company to propose 3 names for our NCE. There were all kinds of rules for the name, may not imply disease, or cure etc.

Another thing is you have all the remaining big Pharma's doing the same thing with that identified TARGET. There may be 6 companies chasing that one little pocket of the protein. These costs are the same for all of them, but Light would have you believe that 5 of them are just me toos and that is much cheaper, even though you are second or third to market for a smaller slice of the pie.

Permalink to Comment

34. lynn on August 15, 2012 4:10 PM writes...

Last week I attended [and spoke at] an NIAID workshop on Bottlenecks in Development of Anti-infectives. I wasn't sure who would be attending. Academics? Biotechs? It turned out that 50% of the attendees were from NIH, NIAID, other government funding agencies - because they finally decided to reach out to Drug discoverers and Developers (mostly, but not all, from industry). They initially had included only one session on drug discovery (me) and they wanted me to talk about screening strategies. I told them that this was not the main bottleneck in antibacterial discovery. So, they invited some other discovery people while I did a session on bottlenecks in discovery. Which, in antibacterials includes chemical libraries with the wrong physicial properties, the difficulties of getting compounds into bacteria (especially gram negatives), and the difficulty in choosing the right targets, because inhibitors of single enzyme targets are highly prone to resistance [just like with HIV or HCV or TB]. There have been no novel classes of marketed antibacterials discovered since 1987, and almost all classes were discovered by serendipity and empirical screening (mostly but not all of natural products). Genomics and target picking for HTS screening for antibacterials has been amazingly low yielding. It is just so hard to discover, let alone develop, candidates for this area. Whereas creating modifications of the "old" useful classes, specifically to overcome existing resistance has been successful. People put this down as Pharma making me-toos. But it seems that the old multi-targeted drugs that are not subject to spontaneous single-step high-level resistance are better than what has been tried for the last twenty years. Like in HIV and HCV and TB, antibacterials with single enzyme targets will have to be used in combinations - but, currently, there is no clear path for developing such combinations. Reading about Light's and Angell's thoughts on discovery is very depressing. But I am a bit more hopeful that the NIH folks seemed to get a lot out of the workshop, hearing from all of us how hard the process is - and how all parts of it have to be integrated - and how difficult it is to outsource parts. First of all, biologist and med chemists have to talk to each other. All the time. Anyway, thanks for the post, Derek, and I look forward to reading your "final" response to these people.

Permalink to Comment

35. clinicaltrialist on August 15, 2012 4:56 PM writes...

I should add, for biologics trials, cost of goods can be material part of the cost. API for antibodies cost about $300/g or more, and typically, you would dose about 1 mg/kg once or week (roughly). That's 100 mg per week, or 5 g per year per patient. Once you add in the fill/finish, labeling and QA costs, it's about $500/g X 5 g = couple of thousand dollars per patient.

Also, with regard to data management and monitoring, they cost so much because

1) There are many steps involved. Data collection by study site. Entry into data system. Double entry into data system. Edit checks. Queries generation. Query transmission Query resolution form generation. Query resolution approval by site. Query resolution re-check. Reconciliation with safety database. Mapping of adverse events to standard nomenclature dictionary. Travel of monitor to the site. Finding the medical records (which is not trivial in some hospitals). Locating medical source data. Verification of source data with record in database. Tracking down site coordinator for missing or wrong data. Query resolution form generation. Etc. Etc. - OK, you get the idea.

2) Pharmas typically have very high quality standards. Many insist on double data entry (data entered twice by two different people). 100% source data verification (every data point verified by CRA). Double programming (every data table and listing programmed independently by two sets of programmers so insure accuracy), etc.

The 100% data verification/quality is a killer (though I think 100% SDV era is finally ending due to cost constraints and new guidances from the FDA). As in many things, the last 10% takes 90% of the work. In one 17,000 patient study I was involved with a while back, cleaning the data in the first 16,900 patients was a lot of work but the missing data in the last 100 patients took an enormous amount of effort as well, especially since one of the patients was a prisoner who has escape from prison in the middle of the study, and even the police couldn't track him down. We finally convinced the data managers to close the database without him, but he did so reluctantly.

Permalink to Comment

36. wprocessing on August 15, 2012 5:50 PM writes...

API cost can be an issue. The last campaign I ran cost ~$2M. For two batches. That doesn't cover all of Phase II - only part. Plus all the upfront cost of working out the synthesis strategy, impurity profile, polymorph stuff, GMP stuff...etc. This is not as expensive as clinical, but its not insignificant.

Remember, we don't just hop from ID of a candidate molecule to formulation of the API. Someone has to make the first 1-10 kgs. And then the next batch of 100 kg, and then 350 kg, and then 2000 kg...have YOU ever tried to scale up?

I don't think Light has taken into account much of any of the full demands of Research...and he seems to know even less about the Development process.

Permalink to Comment

37. clinicaltrialist on August 15, 2012 7:21 PM writes...

I should add that the reason that pharma is so insistent on doing so much work on quality assurance and validation of clinical data is for the following reason: most clinical sites do clinical trials as a side project.

The doctors who enroll and treat the patients and record the data (or have the data recorded) usually do not do clinical trials as their day job. Running a practice (or for academics, teaching/research/treating patients) is typically their day job. Except in places like Eastern Europe (where clinical sites often do see clinical trials as their main job), physicians usually put patient care first and trials second.

The doctors and the site coordinators typically don't have professional training specific to running clinical trials either. They kind of learn on the job, and some are willing to spend more time than others to learn the ropes.

As a result, if you didn't do a lot of data cleaning and verification (though I would argue many companies overdo it), the data would be so dirty that you couldn't really use it. There are thousands of moving parts in a clinical trial and it's nontrivial to do everything properly.

It's like you're asking your local neighborhood engineers and mechanics who have some time on their hands to come and help put together a 747. They may be very talented, but if it's not their day job, you're going to have to do a lot of QA and QC.

Permalink to Comment

38. CMCguy on August 15, 2012 7:47 PM writes...

#25, #32 & #36 while it is indeed true the supply cost of goods and process development can be a significant contributor to absolute expense of drug development project it usually on a relative basis very much smaller than the clinical demands and other aspects (QA, Reg) as the requirements aptly eluded in clinicaltiralists multiple elaborations. For small molecules if COGs projections exceed 5% the execution of the project will likely be a tough sell and even though biologics can be higher tolerance (maybe 15%) there is often much scrutiny during times when often don't really have a scalable process so can even define how much will cost to make the drug.

Permalink to Comment

39. Anonymous BMS Researcher on August 15, 2012 8:51 PM writes...

Maybe some enterprising reporter can ask pharmas, "may I spend a few weeks sitting in on meetings of midstage discovery groups to learn what it's really like if I sign an agreement promising not to reveal anything proprietary?"

That reporter will mostly see many powerpoint slides of structure-activity relationships, ADME, PK, preclinical Tox, etc. "These compounds bind the target best, these compounds have the best solubility, these compounds have the best PK, these have the least off-target effects..." Somebody proposes a structural model of what might explain some trend in the SAR, after which six people start waving their hands around discussing how the latest compounds might fit into the binding pocket. There is the usual argument between folks who want to try more compounds with fewer asays per compound versus those who want to run fewer compounds through the assays in order to do more thorough assays of each molecule. Etc., etc., etc. Makes Rubik's cube seem simple. Seems like all too often whatever improves one property takes you away from the desired profile in some other aspect.

What academics call "leads," we usually call "tool compounds."

Permalink to Comment

40. scalar on August 15, 2012 9:25 PM writes...

@ #32 Respisci

I'd like to add that the amount of clinical drug substance/API is also sometimes quite small compared to the amounts needed for long term tox!

Permalink to Comment

41. rjw on August 15, 2012 10:41 PM writes...

Don't forget dose finding in phase II - complex, fraught and expensive.

Permalink to Comment

42. insilicoconsulting on August 15, 2012 11:15 PM writes...

No one disputes that drug discovery is extremely complex, fraught with challenges, failures and other UNK UNK's and known unknowns etc.

But rather than pick a fight over 43 million or 300 million or just try to justify why it takes 800 million or more, it would be great if you comment about what solutions could be explored at every stage of the process. Both the science and economics and IP issues [change patent law?] could be addressed.

After all even if standards were a bit lax and low hanging fruit were more plentify, people did discover drug prior to the 1990's.

Should we consider polypharmacology for most diseases? Should we use phenotypic screening in most situations? Should we use phase 0 more often? Should patent law be fundamentally revisited? Should DD efforts start only when much more of the underlying biology and etiology is understood , rather than unproven hunches [e.g. alzheimer's]. How can clinical trials be simplified in cost and complexity while maintaining ethics and statistical certainty?

And many more...

Permalink to Comment

43. Anonymous on August 16, 2012 12:21 AM writes...

I involved in one FIH Ph1 clinical study in cancer patients (~50) that costed 20 million USD. Imagine, how much would it be to complete Ph2 and Ph3 studies for getting approval.

Permalink to Comment

44. matt on August 16, 2012 2:39 AM writes...

@42, insilicoconsulting:

The reason for these posts is that you and Light and Angell DO dispute the complexity of the drug discovery process, and have publicly disputed it. The fight over 43 million is that it completely disparages more than 90 percent of the work that is being done, more than 90 percent of the workers who rack their brains to find solutions. With tens of thousands of people being laid off, and thus being already disparaged by their employer, being further minimized in the academic and medical literature and now in popular media, it becomes a bit much to take.

It would be great if you could read some of the previous posts on this blog. They have explored some thoughts on solutions, at least in the lead discovery part of the process where Dr. Lowe's experience qualifies him to make observations.

Of course there are many posts, so here's what you can do: at the bottom of each blog post are some links beside the word "Categories:". For this post, those 2 categories were Drug Prices, and Drug Development. Click on those and you will get a list of all the posts on these topics in reverse chronological order. You will find some suggestions of what seems promising, some of what to avoid, and some comments on news in these areas. In particular, I'd direct your attention to the post on Eroom's Law for a gloomy thought.

With regard to phenotypic screening, I believe Dr. Lowe has said he's a fan. With regard to clinical trials, you should probably ask "clinicaltrialist" whose username links to the website clinicaltrialist.com. I'll bet that's a good resource to check out for ideas on that question.

Permalink to Comment

45. newnickname on August 16, 2012 5:18 AM writes...

I am surely in the group of disbelievers when it comes to Light's $43 MM figure and Marcia Angell's low-balling but I do want to point out that, IMO, Angell did a lot of good stuff at NEJM such as requiring full disclosure of conflicts of interest, pushing for publication of negative results and, basically, supporting, then and now, evidence based medicine. Angell's public persona is, at least, that of a quackbuster who wants to elevate public awareness of many medical and health care issues. I haven't read as much Angell as Pipeline :-) (well, maybe I have) but I assume that these are fair labels.

Maybe Angell will reconsider her estimates of the costs of bringing new drugs to market in the face of better evidence.

DISCLAIMER: I do not know Angell or Relman. I have no financial interest in anything they do.

Permalink to Comment

46. Pharmevangelist on August 16, 2012 6:36 AM writes...

W need to do some edjamacatin' on a herd of academics, too. If anyone knows of good online videos or podcasts that briefly describe the drug development process at a level of detail that is suitable for post-graduate biomedical scientists, it would be great if you could post the link(s). Thanks.

Permalink to Comment

47. newnickname on August 16, 2012 7:04 AM writes...

@46 Pharmevangelist: IMO, there is a lot of stuff out there ... including the non-video, non-podcast Pipeline, mostly prepared by and for the choir.

The problem is reaching the target audience. In order to be "good" (to them), you have to speak in their language, play to that audience and make it available in the places they frequent so THEY will share it and blog about it. For that, I have no answer.

(I apologize for many typos and mistakes in some of my posts. There isn't always time to check / edit things and there is no way to edit once posted.)

Permalink to Comment

48. simpl re #33 name costs on August 16, 2012 7:38 AM writes...

Okemist, there are ways to optimise the costs of names like INNs, and one obvious one that we use is to wait till later (phase III). The WHO is okay with this, they like the attrition rate to be low. The small R&D partner usually goes earlier, so they can 1) give their product a handle, and possibly 2) give the impression they have something close to market.

Permalink to Comment

49. Respisci on August 16, 2012 9:04 AM writes...

#40
I heartily concur.

Permalink to Comment

50. yonemoto on August 16, 2012 11:41 AM writes...

Light's analysis on the costs might be wrong, but the margin analysis is not. Margins, are total revenue minus total cost.

Major pharmaceuticals pull 16% margins, minor pharmaceuticals pull 14% margins. There are few sectors which pull that level of margin (publishing, hedge funds, and precious metal mining - though that's assuredly temporary). Generics, which do have to do some process R&D, but no lead and development R&D, pull a whopping... 6%. But they, of course, have to compete with each other and enjoy no patent protection.

http://biz.yahoo.com/p/sum_qpmd.html

Permalink to Comment

51. clinicaltrialist on August 16, 2012 1:01 PM writes...

insilicoconsulting (#42), your question is an important one, and as matt (#44) says, one that has been discussed a lot. But there are a lot of ways to ask the question.

The following questions are all different:
1) "how much does it cost to put a drug through development"
2) "why does it cost that much" which can be divided into
___a) "where (what activities) does the money go"
___b) "why does it go to those things"
___c) "why is the price so high"
3) "should it cost that much"
4) "what can be done to reduce it"

And all those questions are efficiency questions which are different from productivity questions. There's efficiency, which is cost per drug candidate. And there is productivity, which is cost per successful drug.

The problems in drug industry are twofold. The cost per shot on goal is going up, and the success rate is not budging (or perhaps declining).

I've discussed where the money goes and a bit about why with respect to clinical trials, but if you want to know how to reduce it, it's a very long and involved story, and I can't go into all the details. Suffice it to say that most experienced clinical trialists would agree that a substantial reduction in cost is possible, but the barriers are nontrivial. The barriers are cultural (risk tolerance), economic, etc. I'll just a handful of examples, but this is a small subset.

Cultural/Risk Tolerance: We think we're unique as an industry and we're slow to adopt innovations from other fields. And we don't really want to go around fixing things that work just to save money if it hasn't been thoroughly proven. For example, it took a long time for us to get rid of carbon paper (actually, I lie - some of us still use carbon paper) and move to web based data forms.

Economic/structural: The way that most CROs get paid is by the hour, not by milestones. This means that longer a trial runs, and more errors there are, more money the CROs make. Not that they are causing errors on purpose, but their top priority is not to make the trials as short and efficient as possible.

(And don't even talk to me about change orders. That's where a lot of CROs make their profit margin - it's very much like hiring a contractor to build a house.)

Economic/cultural: there are too many trials. There are too many companies going after oncology patients, for example. That is driving up demand and the supply is limited. Better weeding out of less promising drugs earlier in the development process would lower demand for sites, patients, CROs, etc. But in an industry populated by optimists (you'd have to be, to work in an industry with 5% or less success rate) this seems to be the situation.

Permalink to Comment

52. yonemoto on August 16, 2012 3:58 PM writes...

just out of personal interest, Derek, could you go through the differences between an anticancer drug and a general pharma compound? Obviously phase I trials are totally different, success rates probably, too, right?

Permalink to Comment

53. FormerPharma on August 16, 2012 6:39 PM writes...

@47 newnickname:

I think this is always the problem - how can you reach the target audience, which in this case should be the general population who have no choice but to lap up whatever is fed to them by the popular media? Unfortunately, stories with titles like "Drug R&D is extremely difficult and only a tiny fraction of the costs can really be attributed to public funding" or "Drug R&D costs claimed by major pharmaceutical companies are not bogus" just do not make for stories that would appeal to the general population, and as such would not be published by major media outlets or bloggers for them.

Permalink to Comment

54. Rock on August 16, 2012 7:42 PM writes...

Derek, you suggested that your readers would correct you if you were off base. I would like to challenge one of your statements. I do not believe that most medicinal chemists will get one drug through Phase III in their career. It is probably more like 10% of medicinal chemists will get a drug through Phase III. Perhaps less.
Also, don't forget John LaMattina's book "Drug Truths" which also attempts to counter the claims of Angell and others like her. He has gotten some exposure in the media because of it, but as you can read on his blog, with limited success in changing misconceptions.

Permalink to Comment

55. Nick K on August 16, 2012 8:48 PM writes...

Rock #54: The very first thing they told me when I started my career at Pfizer back in 1986 was to forget any fantasies about making a marketed drug. The probability was just too small. And never did make a marketed drug, though one of my compounds was at least nominated as a back-up.

Permalink to Comment

56. S Silverstein MD on August 16, 2012 10:35 PM writes...

Derek,

Thank you very much for this exposition of the realities drug discovery.

It should be required reading of all non-scientists who have decision making authority in pharma, especially over money, people and strategy.

In my experience, people who did not understand these issues were making decisions they ought not to have been empowered to make.

Permalink to Comment

57. petros on August 17, 2012 8:07 AM writes...

#54,#55

In 20 years as a med chemist my best shot was one into phase II, which primarily went down on cost of goods and one candidate which failed tox

Permalink to Comment

58. LJStewartTweet on August 19, 2012 1:10 AM writes...

This is a very important topic and it’s tremendous to see the debate unfold here. Arguably, this forum is where the aggregate truth can be found. Thank you all.

Derek, I applaud Vertex for considering working on totally novel targets with unprecedented therapeutic areas. This is indeed where investments need to be made…! No one can argue with that point, not even Dr. Light, because Vertex isn’t pursuing “me too” incremental improvement drugs which seems to be the thinking about Pharma and of a lot of discussion at http://www.pharmamyths.net/

I worked for deCODE genetics for 6+ years all the way until 3 days before it’s bankruptcy. deCODE burned through ~$1.2 Billion and no approved drugs resulted. But we probably only spent ~$100M on drug discovery. The other $1.1 Billion was spent on genetic based target discovery (is that drug discovery?). This is just one of many biotechnology companies that burned investor money with no approved drug products to show for it.

Regardless of how you digest all the economic numbers… or how you treat net R&D spending etc. The truth comes down to people’s experience.

I’ve met 100s of people in pharma world who have worked all their careers trying to develop drugs. Unfortunately, I can’t name more than a handful of people I know who actually were successful in being involved in developing an approved drug product.

So if you look at it from this perspective, I know for sure that the cost of drug development is many-fold more than Dr. Light’s numbers.

Another way to look at it is if the cost of drug discovery and development were in aggregate $43M then it would be completely irresponsible for governments not to spend this kind of money to address the ~$1 Trillion annual dollar economic costs estimated for dealing with Alzhiemer’s Disease by the year 2025. When it comes to AD, you already have easily more than $1 Trillion burned in Pharma trying to address AD, and so far nothing working.

If $43M were the true costs of drug development then it would be further irresponsible for any large Pharma CEO not to have a major internal effort in CNS to develop drugs to profitably treat the many unmet medical needs in CNS. Yet, many Pharmas have pulled back significantly in CNS.

Is it really ~$1.3 Billon to develop a new approved drug? I don’t know the answer to this, but one thing is for sure the number is not $43M.

There is no Pharma conspiracy. It’s just really very hard to develop a new drug, and frankly we just don’t understand biological systems well enough to do it very cost effectively.

I sure wish then number were $43 M, but its not !

Permalink to Comment

59. Jay Dee on December 3, 2012 6:31 PM writes...

You forgot the last step. Your drug is being sold and Hear Come the Lawyers! These come in two forms:

1) Your competition doing everything it can to bust your patents.

2) Class action scumbags putting together pools of anyone who ever had so much as a headache while taking your product and trolling for a huge settlement. 1-800-BAD-DRUG indeed!

Permalink to Comment

60. eman on December 4, 2012 12:41 AM writes...

An API has to be treated throughout the entire development process as if it is certain to succeed. You can't hold back. Process chemistry must take the synthetic scheme used by med chemists and revise it so the company can make multi-kilogram batches for multiple purposes (tox batches, GMP, etc.) and can not wait for trial data that says, "Sorry, epic tox fail."

You need, just on the chemistry side of things, high yield, minimized cost, low impurities, the right polymorph, and sometimes even the right particle size distribution. All of it needs to be well-controlled, robust, validated, and ready to go at any scale anywhere in the World.

The sad truth is a drug can die anywhere in the process for a whole host of reasons and there is nothing we can do about it. We must spend and test and innovate as best we can, fingers crossed.

Permalink to Comment

POST A COMMENT




Remember Me?



EMAIL THIS ENTRY TO A FRIEND

Email this entry to:

Your email address:

Message (optional):




RELATED ENTRIES
Messed-Up Clinical Studies: A First-Hand Report
Pharma and Ebola
Lilly Steps In for AstraZeneca's Secretase Inhibitor
Update on Alnylam (And the Direction of Things to Come)
There Must Have Been Multiple Chances to Catch This
Weirdly, Tramadol Is Not a Natural Product After All
Thiola, Retrophin, Martin Shkrell, Reddit, and More
The Most Unconscionable Drug Price Hike I Have Yet Seen