Marcia Angell’s Mistaken View of Pharmaceutical Innovation

At Econ Talk, Marcia Angell discusses big Pharma with Russ Roberts. I think she gets a lot wrong. Here is one exchange on innovation.

Angell: The question of innovation–you said that some people feel, economists feel, [the FDA] slows up innovation: The drug companies do almost no innovation nowadays. Since the Bayh-Dole Act was enacted in 1980 they don’t have to do any innovation….

Roberts: But let’s just get a couple of facts on the table…[The] research and development budget of the pharmaceutical industry is, in 2009, was about $70 billion. That’s a very large sum of money. Are you suggesting that they don’t do anything–that that’s mostly or all marketing? That they are not trying to discover new applications of the basic research? It seems to me basic research is an important part. Putting that research into a form that can make us healthier seems to be a nontrivial thing. You think they are–what are they doing with that money?

Angell: If you look at the budgets of the major drug companies–just go to their annual reports, their Security and Exchange Commission (SEC) filings, you see that Research and Development (R&D) is really the smallest part of their budget. If you look at the big companies you can divide their budget into 4 big categories. One is R&D, one is marketing and administration; the other is profits, and the other is just the cost of making the pills and putting them in the bottles and distributing them. The smallest of those is R&D.

Notice that Angell first claims the pharmaceutical companies do almost no innovation then, when presented with a figure of $70 billion spent on R&D, she switches to an entirely different and irrelevant claim, namely that spending on marketing is even larger. Apple spends more on marketing than on R&D but this doesn’t make Apple any less innovative. Angell’s idea of splitting up company spending into a “budget” is also deeply confused. The budget metaphor suggests firms choose among R&D, marketing, profits and manufacturing costs just like a household chooses between fine dining or cable TV. In fact, if the marketing budget were cut, revenues would fall. Marketing drives sales and (expected) sales drives R&D. Angell is like the financial expert who recommends that a family save money by selling its car forgetting that without a car it makes it much harder to get to work.

Later Angell tries a third claim namely that pharma companies do no innovation because their R&D budget is mostly spent on clinical trials and, “it’s no secret how to do a clinical trial.” I find this line of reasoning bizarre. I define an innovation as the novel creation of value, in this case the novel creation of valuable knowledge. Is Angell claiming that clinical trials do not provide novel and valuable knowledge? (FYI, I have argued that the FDA is overly safety conscious and requires too many trials but Angell breezily and nastily dismisses this argument). In point of fact, most new chemical entities die in clinical trial because what we thought would work in theory doesn’t work in practice. Moreover, the information generated in the clinical trials feeds back into basic research. Angell’s understanding of innovation is cramped and limited, she thinks it begins and ends with basic science in a university lab. Edison was right, however, when he said that genius is one percent inspiration and ninety-nine percent perspiration–both parts are required and there is no one-way line of causation, perspiration can lead to inspiration as well as vice-versa. Read Derek Lowe on the reality of the drug discovery process.

Angell infuses normative claims to the industrial organization of the pharmaceutical industry. Over the past two decades there has been an increase in the number of small biotechnology companies, often funded by venture capital. Most of the small biotechs are failures, they never produce a new molecular entity (NME). But a large number of small, diverse, entrepreneurial firms can explore a big space and individual failure has been good for the small-firm industry which collectively has increased its discovery of NMEs. The small biotechs, however, are not well placed to deal with the FDA and run large clinical trials–the same is also true of university labs. So the industry as a whole is evolving towards a network model in which the smaller firms explore a wide space of targets and those that hit gold partner with one of the larger firms to pursue development. Angell focuses in on one part of the system, the larger firms and denounces them for not being innovative. Innovation, however, should be ascribed not to any single node but to the network, to the system as a whole.

Angell makes some good points about publication bias in clinical trials and the sometimes too-close-for-comfort connections between the FDA, pharmaceutical firms, and researchers. But in making these points she misses the truly important picture. Namely that new pharmaceuticals have driven increases in life expectancy but pharmaceutical productivity is declining as the costs of discovering and bringing a new drug to market are rising rapidly (on average ~1.8 billion per each NME to reach market). In my view, the network model pursued on a global scale and a more flexible and responsive FDA, both of which Angell castigates, are among the best prospects for an increase in pharmaceutical productivity and thus for increases in future life expectancy. Nevertheless, whatever the solutions are, we need to focus on the big problem of productivity if we are to translate scientific breakthroughs into improvements in human welfare.


"new pharmaceuticals have driven increases in life expectancy": that's plausible, but is there good evidence for it?

Follow the link!

And yet, the convincing argument is that increases in hygiene (clean drinking water being the single most important aspect), better nutrition, and improved housing have led to the greatest advances in increased lifespans.

'Non-medical factors, such as better living conditions and hygiene, better food availability, variety and balance and improved nutrition, have also played a significant role but direct and clear evidence for their individual importance is difficult to obtain. Indirectly, however, their potential importance can be judged by the fact that declines in mortality from certain diseases have preceded medical advances in the treatment of the particular diseases. An example of this is tuberculosis (McKeown, 1976). Figure 8 shows that the mortality from tuberculosis in England and Wales, which was very high at the beginning of the 19th century, had declined by about 50% by 1880, the year that the causative tubercle bacillus itself was first identified. Mortality had substantially declined even further by the mid 20th century, when the first effective treatment and vaccines were developed.

In the case of tuberculosis, therefore, the advancement of medicine was obviously secondary to that of other changes occurring within this particular population. Among the likely explanations for the decline in tuberculosis before the advent of medical treatment, are improved housing conditions and better hygiene.'

Just don't expect any pharmaceutical company to point this reality out.

We live longer because by historical standards we live better - clean water, more than adequate safe food, and housing which is well heated but not filled with smoke or parasites, are the greatest contributors to lifespan advances in the modern era.

And a note - tuberculosis was the leading cause death in the U.S. in the early 20th century ('Tuberculosis (TB), once the leading cause of death in the United States, appeared to be receding into history by the latter part of the 20th century' - )

The same, non-medical related decline in TB death rates is undoubtedly shared by the UK and the U.S. - as loyal comment readers might remember, the campaign to stop spitting in public in the U.S. was motivated by the desire to help stop the spread of the leading cause of death in America at the time.

Two words: Protease inhibitors.

And yes, I know protease inhibitors are treatments for HIV. But if you are going to bang on big Pharma, it is worth noting that Roche developed the first FDA approved protease inhibitor. And if you want to talk about the history of TB, you really can't leave out streptomycin or isoniazid.

Seriously. Like I was just telling my rheumatologist. I suffer from a horrible disease called ankylosing spondylitis which essentially renders me an invalid. He prescribed me Humira, after which I was basically cured and my life was livable again and I put aside thoughts of suicide, etc. He was all like, "so it worked!" And I'm all like, "What? You think some big pharmaceutical company cured me? Fat chance, pharma stooge! It was my government-provided tap water!" Then I slapped the chart out of his hand and walked out.


And yet, the reality that if you were drinking the water typical of a third world country, you would also have a third world life span seems to be too obvious to be actually worth discussing.

Or did that TB information just go over your head?

No one is arguing that pharmaceuticals have no place in medicine - just that when looking at historical information, the clearest contributor to longer life spans was the provision of clean drinking water, adequate food, and decent shelter. Longer life spans are not about any particular medicine solving any individual's disease - they are about entire populations.

Yet another citation based on historical reality and population records -

'To explain recent improvements in the quality of life in the West, we must return to the 19th Century, when the hygiene revolution began. More specifically, attention should be given to progress in the field of sanitation since that time.

Life expectancy remained stable during the first part of the 19th Century. However, mid century, a new event occurred, that is that pure water was first provided to the cities in the Lyons region (second largest city in France situated in the south-east part of the country). Thus, around 1850, clean water became abundant. Although there was some progress, the situation in Paris was not so favourable.

However, progress in Paris only became significant in the last twenty years of the century. It was then that new pumps along the Seine made it possible to double the provision of water. Furthermore, the number of sewers increased significantly; comparing the situation in 1870 to the one in 1900, the number of sewers had more than doubled during that period of time. Finally, in Marseilles (south of France by the Mediterranean sea), the situation only improved in 1890, when a double sewerage system was installed. Around 1900, life expectancy increased considerably over the whole of France. In fact, at that time, Paris was considered to be the cleanest city in the world.

In developed countries, it is possible to say that progress in life expectancy only became significant once the provision of water and sanitation systems had been completed in the principal towns.'


If you aren't arguing that that pharmaceuticals have no place in medicine, what are you arguing in regards to pharmaceuticals? Sanitation and living standards helped, likely a very large amount. No one is arguing that they didn't. Great, we're in agreement, now what?

It probably went over my head! I've never been able to make heads or tails of Wikipedia entries. I'll have to consider further whether the introduction of clean water and sewage is currently driving increases in life expectancy. I hope it is!

Yes but the high priests of statist dogma have ruled that you don't fit into their narrative.

Note the two ways of parsing "new pharmaceuticals have driven increases in life expectancy":

1. "All that increase in life expectancy? It was driven by pharmaceuticals."

2. "Pharmaceuticals have increased life expectancy."

Good point. However, average life span being considered an aggregate measure, one needs to look at measures effecting it in aggregate. Vaccinations come to mind - a concept that predates the pharma industry, it must be noted. As does doctors cleaning their hands and instruments before performing medical procedures, most noticeable in terms of maternal and infant survival rates, as history demonstrates. Antibiotics have undoubtedly saved many lives, and it is safe to assume would have an effect noticeable in aggregate.

And of the three examples noted, only one has anything to do with actual pharmaceutical innovation, though vaccinations are clearly something which are now considered a part of the pharmaceutical industry's business.


the clearest contributor to longer life spans was the provision of clean drinking water, adequate food, and decent shelter.

One weak test of your hypothesis is to look at the historical data on mortality of nobles and royals. Water, food and shelter were never a problem for royalty so I'd expect a cleaner picture of how much drugs actually helped.

I don't know the answer.

Rahul, Prince Albert, Queen Victoria's husband, died in the last great Typhus outbreak in the UK. Precisely because the Royals weren't having any of these modern new-fangled things like sewage.

Don't assume that water isn't a problem for the upper class. It is hard to retrofit drainage to a 12th C castle.

It was the CEO of a pharmacutical company that first told me the "truth" that improved sanitation has done more to improve health than any other development.

So, prior_approval, would you suggest we quit looking for any new antibiotics since the old ones never did us any good anyway?

We're all aware that there were huge returns on picking up the low-hanging fruit. You've written a thousand words in this thread, and informed no one.

CV and diabetes mortality in the US have dropped the last 10-20-30-40 years. That's not from hygiene. It's from ACEI, statins, clopidigrel, etc.

How can profits ever be part of a budget? Isn't a budget on expenses? Is her accounting kosher?

She presumably is describing what happens to revenue, and is saying that profits are higher than R&D spending. But the language indeed betrays a certain amount of ignorance.

Are you sure about marketing's multiplier effect? What about in a liquidity trap?

In any case, the companies themselves are quite sure about marketing being something that makes them more money.

They don't spend on it thinking it's losing them money; they spend on it precisely because they expect to get more in return than they spent.

Now, in some cases they might be wrong, I suppose ... but they've got much better leads on the knowledge than Angell does, necessarily. They're much more likely to know how well their marketing works than she is.

Alex, I just stopped by to say that I am enjoying seeing you more active on MR.

People trot out Heinlein's bad luck quote a little too often, but right here is a place where it is totally appropriate.

I liked this part "In fact, if the marketing budget were cut, revenues would fall. Marketing drives sales and (expected) sales drives R&D".
It remembered me about something along this line "government shouldn't cut spending, because government spending drives income and (expected) income drives investment".

that graph plots a real variable against a nominal variable.

Looks like an innocuous time series to me. What's wrong?

Also, those two graphs shouldn't be superimposed at all. This creates the impression that they should be correlated. But the variables are number of NMEs and cost per NME, so there's really no particular reason to expect any kind of correlation between the two.

There is a possible corellation, that as the cost per NME rises,you get less of them, since they are more expensive to produce. That would be a sensible relationship, and a sensible corelation. However I see more random up and down fluxuation in the number of NME's plot than I see steady decline, so the corelation is not that clear to me.

I listened to the podcast and one exchange struck me. Angell was complaining about "me-too" drugs, which were described by her as being very similar to ones that the drug companies had developed but for which the patent had expired and the profits vanished. The idea was that they could get monopoly profits for a drug that was little better than the (now) generic one. Russ suggested that patients be allowed to select the cheaper drug. Angell seemed astonished that patients should be allowed such a choice. I realized that Angell, who is at Harvard and a big advocate of single-payer, is probably not exposed to other than standard-issue "progressive" views. Such is our "elite".

I'm not one to pile on, but what struck me was that everything had an obvious, intrinsic government intervention that I would humbly propose would be the key to why things are the way they are.

The me-too drugs, at least the over-emphasis on them, are an emergent unintended consequence of the patents they are allowed and the status provided to them by the regulatory regime.

Angell just accepts those as given and aims to fix the methodology of the trials. Why would we take issue with the existence of me-too drugs? Filling out the molecules and their biological effects is an unmitigated good. The problem is how they are profitable and this is almost entirely due to the ins and outs of the regulations.

Especially with brain-drugs, the "me-too" drugs can be very valuable to patients. Sometimes one drug works where another one does not, even while the biochemists say they ought to be just about the same.

(Not to pick on biochemists. They readily admit they are tackling a huge problem they only have a tiny handle on when they try to decode the brain.)

Oh yeah. That was another thing that crossed my mind listening. Talk to a health professional and they'll decry the me-toos. Talk to a relative and they'll tell you about the several drugs they tried before one worked and didn't cause unbearable side-effects.

Patients are already allowed to select a generic alternative. Angell was pointing out that it didn't work--patients are still selecting brand name drugs over generics despite there being no difference between the two. Russ agreed, stating that perhaps this is because people don't directly pay for healthcare, or they pay for healthcare with other people's money.

I don't remember single-payer even coming up in the podcast.

I looked her up and that's where I found that she was a single-payer advocate. I wondered where she was coming from and it made her comments more understandable. I don't think having participants pay more of their own money for healthcare would even occur to her.

When you have a high deductible plan like I do, you think about these things before you go ahead.

"Angell was pointing out that it didn’t work–"

She asserted that it doesn't work because she assumes that once a generic or me-too is available then it is identical. What we were talking about is that people try different drugs that have similar benefits because they may have dissimilar side effects.

If you allow the insurance companies to charge higher copays for brand name than generics, patients will notice, and ask for generics, unless a non-generic me too drug produces better results. Angell sounds like a big gov leftist hack, with no understanding of the market.

Oftentimes, me-too drugs have no utility whatsoever. I am pasting a summary below for those who didn't hear the podcast. It is an example of company conniving and marketing beating commonsense and costing this country--including us, the taxpayers--hundreds of millions of dollars.

Nexium was developed by the same company that was already making Prilosec, which used to be the most prescribed drug in America with around 60 million scripts. It was going off patent, however. So, the company created Nexium, which is just the mirror image of the Prilosec molecule, and turned Prilosec into an OTC product. Really genius if you think about it. Prilosec costs about $30 per month, while Nexium is $200.

But isn't the real idiot / scoundrel in this scam the doctor? Pharma companies can market all they want; but the doctors need not prescribe the me-too. Unless the company lies about the effectiveness. But that'd be a far more serious allegation.

It also helps if insurance companies have higher copays for non-generics. That will lead to the patiient complaining to the doctor about the high copay, which, if the generic and non-generic really are the same, will lead the doctor to change to the generic. But if the insurance has the same copay for both, then there is no incentive for either doctor or patient to go generic. I wonder how rigid the obamacare rules are, and whether they give the insurance companies the flexibility and incentive to make useful distinctions like this. And note, allowing insurance to charge different copays like this is not mainly to soak patients and fatten profits, but lowers insurance purchase cost for all, although it is more costly for those whose only choice is a non-generic drug. An even more sensible rule for insurance might be 3 copays, lowest for generic, next for non-generic that has no generic equivalent, and the highest for non-generic with generic equivalent. But the way to get this sensible flexibility is for the insurance companies to make these rules, not big gov.

"Just the mirror image"? Sorry, but this sentence shows that you are probably not qualified to post.
Does thalidomide ring a bell?
And, given the long development times for drugs, nobody starts developing a me-too-drug after the first drug is on the market. The players more or less start at the same time working on the same target and one company has to be first.
And. Oftentimes, me-too drugs have a significant benefit, like once-daily instead of every-two-hours.
And. Reading the article under your link, I sometimes wonder if the litigation system in the US is responsible for this "problem". If a doctor prescribes prilosec and the patient gets problems after these quoted 8 weeks and there is a paper showing that the enantiopure drug would have worked better, I wonder what would happen next...

I'm not a drug company biochemist phd, though I did study molecular biology as an undergrad and have some grasp of this science. I said that _oftentimes_ me-toos have no additional utility. Not always the case, but yes that is very often the case--they are basically no different in the case of true me-toos. Can you really make the case that Nexium is worth the additional cost when most patients taking it don't even have any idea why are they are taking that instead of Prilosec? It was the same company who developed Prilosec and Nexium, so your scenario of parallel development isn't applicable here, though I admit sometimes that is the case.

It is an interesting theory about the litigation system, though I've never heard of a patient suing his doctor for using an inadequate drug for excess stomach acid.

What I wonder is how come no competitor pounced on the enantiomeric patent trick immediately after Omeprazole (Prilosec) was patented by AstraZ?

Note that AstraZ only patented the EsOmeprazole just before the Prilosec patent was about to expire; so competition would have had tons of time on their hands.

Do Drug patents need a clinical trial?

It is largely a problem with the patent and regulatory system that the businesses respond to.

The system outlandishly rewards marketing a specific novel chemical to a specific diagnosis after running a (probably too large) safety and efficacy trial.

If you want to fix that then fix that.

Your own link contradicts the claim that Nexium is "just the mirror image" of Prilosec. Nexium is a pure enantiomer. Prilosec is a racemate. Not what you said.

Also, the article complains that Nexium 20 mg dosage is compared to Prilosec 20 mg dosage. Nexium 20 mg would have twice the desired enantiomer, but *none* of the non-desired enantiomer. What are the effects of the non-desired enantiomer? Your blog post doesn't say.

Third, I am no expert but I think the blog author's claim that patents on nexium are invalid since prilosec was already patented are false.

Fourth, the argument seems to boil down to "Nexium has some proven benefit over Prilosec, but in my opinion the benefits aren't worth the increased costs; therefore, it's a scam." That's tendentious in the extreme.

Before I read your link I had no opinion or knowledge on Nexium's comparative benefits relative to Prilosec. Now from reading your article I am *more* convinced that it is likely useful and beneficial in some situations.

Wikipedia reveals that the other enantiomer is not a hazard, not even inactive, only marginally slower.

"[ Is is ] a mixture of two mirror-imaged molecules, R and S. Both are converted to the same active molecule in the body. Some people metabolise R-omeprazole slowly than S-omeprazole"

Just a reminder that the chemistry is irrelevant and also not having to know any of this stuff is supposedly what the government is supposed to do for us.

Rich, I also heard that exchange. Her point was that there's no way to know if a new drug is better or worse than the existing standard treatment because new drugs are only tested against placebos not existing drugs. You may be fixated on single payer, her fixation is safety and effectiveness.

One comment that I have to disagree with is "the information generated in the clinical trials feeds back into basic research". Knowing a specific compound does not work or is killing patients is not terribly useful information and may even send the wrong signals unless the mechanism of failure is really clear. I tell students that informing me a reaction does not work is an almost information free statement; you need to know exactly what happened. Detailed pathway information is hard to get from a clinical trial from the basic research point of view. You can tell if an endpoint is reached but not always what is going on in the body.

She is wrong about Pharma not trying to innovate. They are, but they also engage in a lot of practices that bump profits that have nothing to do with innovation. They have paid generic producers to not produce. They have simply changed the dosing in pills and called it a new med to extend patents. Many of the me-too drugs are little different than older, cheaper drugs, but Marketing (which Alex lauds) allows them to sell what is essentially the same drug at a large mark up. Let us have cost effectiveness studies done on these drugs to see if are getting a good ROI.


The point Ross brings up is that in every other facet of life that doesn't have an FDA 'protecting us' marketing does not cause this.

Her tone really was off-putting when she compared that to shirts because shirts don't matter and we cannot possibly decide which heart valve to select. For one, we aren't talking about shirts, we are talking about EVERYTHING and we weren't talking about heart valves but drugs that she claims everyone knows are equivalent.

Mike Sherer, former chief economist at the FTC, has done excellent work on innovation and risk adjusted returns, including adjustements for technology lag, and found slightly higher returns consistent with the goal of risk taking innovation.

You can find his papers, including the one that appeared in a 2007 book, by googling his name and the term pharmaceutical innovation.

I am puzzled, though, that Alex take this position given his hostility to patents.

Not all patents are created equal. Pharma patents are pretty good.

Actually, that's too categorical of a statement.

For those checking for the book recommended by Bill, the author's last name is Scherer. Also, although everyone calls him Mike, he lists himself formally as F.M. Scherer.

Yes, I call him Mike. And, you can too.

I am also worried that there are many ways in which even trying to solve seemingly simple problems like publication bias can backfire.
Performing 5 clinical trials and reporting only the two which favor the drugs efficacy looks very unethical and violates what is taught in even the most basic statistics course.

However, the statistical approach ignores that performing additional trials is not cost neutral and longer time before approval reduces the revenue from marketing a patented drug. A policy that allows some selective publication of trial results implicitly lowers the significance threshold for drugs with high expected revenue. It is therefore possible that some measures to reduce publication bias negatively impact patients, provided that revenue is positively correlated with the welfare of the patient population. (The same is not true for chery-picking data within an individual trial, since this does not involve the cost of additional experiments)

Similarly, forcing publication of all trial results has the effect of raising the marginal cost of a clinical trial, which might be positive or negative, depending on external factors. Additionally, if clinical trials were performed by less revenue and cost aware organisations, like government agencies, these trade-offs need to be made explicitly by a political decision process, which has its own difficulties.

Big Pharma doesn't spend their R&D money on innovating. They spend their R&D money outsourcing innovation. This may seem like a trivial difference, but it's not. The innovators are small, mom-and-pop biotech firms that behave a lot like IT start-ups. They try to come up with one blockbuster new molecule to sell to a major pharmaceutical company so that they can all retire. Every now and then, these small companies grow into larger pharma companies, which then get bought-out, but still...

The point is, her claim that Big Pharma doesn't spend money innovating is basically true. Big Pharma is a different subset of people than biotech innovators. They do business with each other, but it's not "all the same," even if industry outsiders have a hard time perceiving that fact.

Are you talking about spending money on R&D research or actually conducting R&D research? By your own account, Big Pharma does a lot of the former even if they don't do much of the latter, so it seems to me that Angell is quite wrong.

It starts to become a matter of perspective after a while. It's like suggesting that auto manufacturers spend a lot of money on producing steel. They don't, but they need steel to manufacture cars. Same deal here: R&D expenditures exist for Big Pharma, but that money is solely being used to outsource new drug development to other companies. They're not R-ing or D-ing, they're buying R&D.

So what do we say about that? Does Big Pharma engage in R&D? Only in a manner of speaking. I'm not defending Angell, but I am suggesting that the underlying perspective is somewhat understandable.

When I was still working in the pharma industry I had the dubious pleasure to debate Angell on at least three occasions. I can't say whether I won or lost since she refuses to look at actual evidence but rather is subject to preconceptions that she will never let go of. Space and time are way too short to rebut everything that she claims. Sure there is some 'non-innovation' that goes on such as line extension (changing a formulation to require less dosing for example) but by and large the industry has been one of the most productive of all US industries (and it's important to note that even foreign-based pharma companies do the majority of the research & development here in the US). Biotechs have not been a significant player in the development of new drugs and that trend will continue. They are however a good sink hole where money can be thrown into and never seen again.

Her points seem well practiced. But upon reflection there is always another side to the coin. The point about the Bayh-Dole act catalyzing the dropping of R&D by Pharma in favor of outsourcing it to NIH can be restated as "WTF was NIH funding doing beforehand?" If Big Pharma's job is basically to run clinical trials, then what is the problem, really?

The problem is that she is totally wrong about Bayh-Dole. Pharma research has always been strong and independent of NIH. The huge increase in NIH funding is only a relatively recent phenomenon. Pharma companies have always run their own clinical trials despite what Dr. Angell says and they do a much better job of it than do academicians. In fact, many academic trials need to be repeated by the pharma company since they usually don't meet FDA regulatory standards. Angell tries to reduce everything to a sound bite and she is successful for certain audiences.

Why would Academics attempt clinical trials at all?

As for Biotech, they are supposed to be catologing the 9999 ways not to invent the light bulb.

Biotechs have not been a significant player in the development of new drugs and that trend will continue. They are however a good sink hole where money can be thrown into and never seen again.

Can you elaborate on that? It was a strong statement but I suspect you know what you are talking about.

Strong, and wrong.

Off the top of my head...Amgen; Genentech; Biogen; Vertex; Regeneron...
All rather profitable for investors.

Most fail. But there again, so do most pharma trials.

'Off the top of my head…Amgen; Genentech; Biogen; Vertex; Regeneron… All rather profitable for investors.' and there are a couple of others such as Gilead but look at all the $$$$s that were thrown into biotechs in general (and there were probably over 1000 companies in total that attracted venture money. Some are still around such as Xoma and they really never developed the Sepsis drug that the company was set up for. Others were acquired by big pharma along the way and written down (remember Genetic Systems and Hybritech that were bought by BMS and Lilly respectively? Neither company recouped their investment). The number of pure biotech drugs is still on the small side if you compare that number to the total number of small molecule drugs approved over the same time period (start with the alpha-Interferon approval in 1981). Some of the biotech drugs never realized the market that was forcast (Interferons, human growth hormone) and others ended up having significant safety problems that led to their withdrawal from the market or restricted use, some were too expensive compared to existing treatments and proved no better (TPA vs streptokinase for clot busting follow heart attack or stroke).

Investors did well in a handful of companies and got taken to the cleaners in the vast majority. It's not a pretty picture. I never invested in a biotech (even the ones I thought would succeed) because the odds were just too long. One made more money on MSFT and APPL than in the biotechs.

Where do you think those drug ideas come from? Let me clue you in most of them come out of university laboratories studying biological mechanisms underlying various diseases. And where does the money come from to support those studies!!! Tada-the NIH. Your federal tax dollars at work innovating for the future.

I think Apple is an apt comparison to big pharma. Apple does have a smaller R&D budget than a marketing budget because while I'm sure they are 'innovators' in some form or another they certainly don't drive research innovation. Nor have they ever. They take existing technologies and advancements in interface design that other have produced, and they figure out how to make them work in the mass market. Nothing about that involves Apple doing the necessary R&D themselves. Similarly, large pharmaceutical companies are in the business of acquiring other groups' research, and then convincing regulatory agencies and the public at large that they want to buy the finished product. I would argue that both Apple and big pharma drive basic research in their field only to the extent that they are a potential market for some of the byproducts of basic research. This is true moreso than ever, now that medical researchers are having to find other sources of funding after coming to grips with the extent to which pharmaceutical industry funded research has contributed to research bias.

I think Apple is an apt comparison to big pharma. ... They take existing technologies and advancements in interface design that other(s) have produced, and they figure out how to make them work in the mass market.

I just want to add that what Nathan ascribes to Apple is not easy. If it were, a whole bunch of ex-Xerox PARC employees would be billionaires and the late Steve Jobs would just be a guy who shared fame with Steve Wozniak for the Apple II, if anything.

What really needs to be parsed is R&D, research and development, which are two different functions. Many pharma companies spend a majority of dollars in the Development Phase, running clinical trials to prove efficacy and safety. If we look at many Big Pharma pipelines out there, most of the projects are very low risk and not entirely innovative (there are always examples of truly innovative treatments, but it is nothing like the mid nineties). The true Research and Innovation is being driven by smaller players (such as biotech firms) who are far more willing to accept the greater risk associated with that phase of the process. Specialization at work! One can suggest that the FDA has driven the industry to focus on development of low risk projects due to the requirements needed for approval. The last piece of the puzzle (for me) is even if I am able to drive an innovation (or new molecular entity) to approval, but it has no better impact on actual outcomes such as survival how innovative is that product? How much of a premium should one be willing to pay for it?

Why isn't the defense of pharma innovation a list of important new drugs discovered since 1990, to optimistically purge the public innovation discoveries that suddenly became viable because drug testing was now protected with private patents?

The only new drugs developed privately since 1990 I can think of are the statins, but not initially in the US.

The rate of antibiotic discovery is so low that I believe that innovation is driven by public funding because the private sector considers them unprofitable. The flu vaccine innovation has been publicly funded even though the egg based productions has been long considered in desperate need of replacement.

Where is the evidence of private sector drug innovation??? Name the innovations!!!

Surely you don't consider VIOXX to be innovative?!?

"The flu vaccine innovation has been publicly funded even though the egg based productions has been long considered in desperate need of replacement."

There are plenty of systems in development to replace egg based production. Here's an example (a biotech which was recently acquired by a pharma):

Flu vaccine innovation is both publicly and commercially funded. Most of the hard problems are, and for good reason.

'Surely you don’t consider VIOXX to be innovative?!?' But the premise behind the Cox-2 inhibitors was innovative but unfortunately the safety data later showed that there was a problem with these drugs (and that problem is the same for other NSAIDs such as ibuprofen and naproxen which all have the same side effects as the Cox-2 inhibitors). You are also wrong about the statins - first one was developed by Merck under the direction of Roy Vagelos who later became the CEO. There are a host of other good drugs that have been developed since 1990 so don't go making a blanket statement.

That's what were waiting for the list of drugs or actually drug families that the pharma copies developed from concept to product. I've got 20 small appendages to count them on.

Why isn’t the defense of pharma innovation a list of important new drugs discovered since 1990

Fair enough. By my count, 6 of the 11 top drugs by sales WW (or 5 of 10) are *unambiguously* highly innovative.

They are: (drug/date of first approval/comments)

Lipitor 1996 (statin, 5th or 6th approved, but the best, as it turned out. How innovative one can decide for oneself)
Plavix 1998 Innovative
Remicade 1998 (hugely innovative TNFa blockade product; biotech product)
Humira 2003 (same class as Remicade; significantly improved dosing regimen)
Advair 1998 (combo steroid/laba for asthma -- my sense is that this was really about a better device/drug combo, but this is not an therapeutic area where I know the history well)
Enbrel (1998) same *target* as Remicade, but radically different approach for therapeutic blockade. Enbrel and Humira were developed more or less in parallel by competing groups. This is a hugely innovative, transformative class of drug)
Crestor 2002 (another statin. minimally differentiated if at all from Lipitor)
Rituxan 1997 (hugely innovative)
Seroquel 1998 (one of many atypical antipsychotics. Large debate exists on how valuable they are as a class)
Avastin 1998 (innovative, likely [in my humble view] overprescribed given benefit in many cancers. A *tremendous* drug in age-related macular degneration, and there's a not hugely positive-for-pharma story on how Roche/Genentech tried to capture that value (search for Luncentis vs. Avastin in AMD)
Herceptin 1998 (I extended this list to eleven just because Herceptin is so great. Innovative drug which has had trans formative impact on the treatment of a subtype of breast cancer)

These are MBA run companies. They're at best profitable for MBAs. Apple is wildly profitable only when compared to truly horribly run hardware companies like Dell. Even now the fact that Apple spent more money on marketing and less on R&D is starting to show. Just like Pharma they have fantastic brands, but their product pipeline is drying up.
But I'm sure the MBAs in marketing will start pointing fingers at the MBAs in sales and vice-versa. The fact there is no product to sell is the last thing that gets blamed.

What makes Dell "horribly run"? And if it is, which is a well-run hardware company?

Dell made Apple look bankrupt circa 1997 and now Apple makes Dell look bankrupt.

Bankruptcy is measured by stock price - going up, not bankrupt; going down, bankrupt; staying flat, zombie.

The myth of the $1 billion drug has been debunked numerous times. Innovation is getting harder, yes, but it costs nowhere near that much (or Alex's even more ludicrous $1.8 billion) to develop a drug.

Of of course it costs nowhere near $1 Billion to bring a given drug to market successfully. However, when you compare the amount spent vs. the drugs that actually get approved, 800MM to 1B is a fair figure. Too many candidates fall by the wayside at some point in the process.

As for spending too much on marketing, drug approval takes so long that manufacturers need to maximize the value before the patent runs out. Sometimes they only have seven or eight years left.

How elastic are sales of within-patent drugs to marketing? Anyone know?

Do we really know absolutely nothing? I'd say it is even worse. For example, what I'm doing might be described as trying to develop a new drug. I'm sure I'll fail on that part because that is not the main thrust of my work and my costs will never be included in the total cost to develop on successful drug.

It can't really be $55M can it? If so, let's just toss $10B at it and a wrap party in Vegas later we've got 180 new drugs.

We must be talking about different things:

One of my ideas to lower the cost and failure rate of drugs is to reduce the dosage because in toxicology the dose makes the poison, more or less. We should use drug cocktails instead of swinging for the fences of the miracle drug home run. This will require a MUCH more sophisticated clinical trial and post-approval epidemiological tracking system.

Perhaps you can make a case that the companies ought to count their sunk costs on other drugs that didn't make it. But did you know that the $1-1.8 billion estimate that they tout also includes the estimated forgone _profits_ of other failed drugs? Can you give me an example of any other product development sector that uses this kind of accounting? It is totally disingenuous.

Pharma companies used to have some of the highest profit margins out there. Every other country except New Zealand bans direct to consumer advertising of prescription drugs. Maximizing value indeed.

I found no use of accounting for the foregone profit of failed drugs. The paper you linked mentions that the estimates given here make use of "cost of capital" which is quite common. As the authors note, comparing your expected profits from some long term venture against potential returns from a safe and conservative investment alternative is a regular test of business. Normally, this is not included in the cost estimate for many firms, but normally product lifecycles are neither as long in development (e.g. Intel's current state-of-the-art Ivy Bridge was started around 2005 when Sandy Bridge architecture was being investigated, it is in full mass production now in 2012) nor have as short a lifespan for sale (e.g. Pentiums are still being sold for some uses in spite of being introduced).

At the end of the day, cost bookkeeping is irrelevant. Capital owners (be a drug company or an investor with cash) always have the option of doing something else with their money. If the risk-adjusted-return for investing in drug development doesn't beat the risk-adjusted-return for investing in something else, people will just invest in that something else. This is fine for a lot of endeavors, if your R&D doesn't generate enough return the business liquidates, the remaining businesses increase their prices and returns, the liquidated capital increases demand for other investments lowering their returns and a new equilibrium is reached. The public doesn't have any vested interest in the IP generated except in the final consumer products each firm offers.

With pharmaceuticals, things are different. The drug companies make their profits off the sales whilst the drug is on patent. This lasts, at most, for two decades. The public continues to use drugs developed 60 years ago and likely will continue to do so for the foreseeable future. In short, the public's main interest is not in new patented drugs, but in fat tail of QALYs derived after the drug goes off patent. If you tell the drug companies, "reach a stable equilibrium with other investments" and don't account for cost of capital, then fine - they can do that and they will invest less and the public will get fewer new drugs that eventually go off patent.

I, for one, think that we WANT life-saving research to be vastly more profitable than financing yet another social media venture. If we change the incentive structure to make drug R&D less profitable than it is currently, somewhere on the margin that means we will invest less in new drugs and more in something else. If we do this through regulation and restriction, it will almost certainly mean new investment in less regulated sectors of the economy.

I think we are getting bogged down in unimportant stuff.
Whatever the pharma industry+academia+others spent, not a whole lot of really new, effective drugs have emerged - although, afaik, there is no standard: no one knows what an productive industry looks like interms of lives saved per year per new drug.
so all this stuff about how to apportion the 70 large, fugheddaboutit

The real question is, why do we know so little about the human body, and what can be done ?
why can't we model drugs in silico against hi res xray or NMR structures - why do we need wet chemistry at all ?

those are the important questions.

Jan December 3, 2012 at 4:31 pm - great catch, thanks, I've always been suspicous of the tufts study for hte reasons cited.

Why can't we model drugs in silico? Because Schroedinger's equation is hard to solve in a really fundamental mathematical sense, and the short-cuts are unreliable.

And even once you've convinced yourself that a thing with a pyridine there and two trifluoromethyls on that bit would fit well in the site, you'd better make the compound to check, and the task of seeing whether it inhibits some uncharacterised but essential protein in mouse liver is much more easily carried out using a mouse.

Someone recently said, maybe Nassim Taleb on a Russ Roberts podcast, that essentially no medical treatments have originated from design. They are almost all trial-and-error. Then we build the narrative after the fact. It's shocking but rings true to me.

Marcia Angell ... Proof that women should be kept away from basic numbers, for all our own good.

"we need to focus on the big problem of productivity if we are to translate scientific breakthroughs into improvements in human welfare."

That's just on the supply side. There is also a demand-side challenge, namely that asymmetrical information -> adverse selection. Here Russ was a little too complacent and Angell at least had the problem right. Russ seemed to think consumers could just figure out which drugs were right for them, or which doctors and drug companies were honest, or whatever. Fat chance, I think. Medicine is an informational minefield.

I don't mean to suggest the government or anyone else should be making choices for consumers, just that "translating scientific breakthroughs into improvements in human welfare" does require better information on the demand-side in this case.

Medicine is an information minefield partly because of the mandates and the third-payer. You don't need many consumers to step up in order for the rest of the market to piggy back. I'm one of those consumers. The problem is we get treated like assholes because the people with the full inertia of the system behind them think they are just doing their job when they don't raise a finger.

Angell at least had the problem right.

Fat chance. Angell seems to think bureaucrats could just figure out which drugs were right for consumers, or which government agencies, officials, doctors and drug companies were honest, or whatever. That compounds the number of variables, adding complexity. Government is not science and has no privileged access to truth or unbiased behavior. A CSICOP sort like Angell should know better.

Why does the FDA approve things like shoddy design?

Much like Walter Williams point about education. No matter what is going on, they don't have to hand out passing grades.

One could say "well they are captured" and I would, but you don't blame capture on the capturer, that is their job.

This reminds me of an exchange that happened on an episode of West Wing.

They issue people were discussing was anti-aids medicine in Africa and how much should the US provide - and how much of that burden should Pharma shoulder. Pharma said they'd dispense X amount of pills for free. The folks on other side of the table wanted them to dispense 3X amount of pills for free. Their argument was that "It only costs 1.90 a pill. That's nothing to a company that big." Then the reply came, "No, it's the second pill that costs $1.90. The first one cost 2 billion."

I always liked that.

Back to water and hygene and such. Have there been such sweeping changes in water, food, and shelter over the past 50 years or so in America that would explain the spike in life expectancy over the same time.
As far as the cases of TB declining before medicine existed to treat it, wouldn't you expect the virus to wipe out the most susceptible at first, leaving a smaller number to infect, and those who'd shown to be more resistant? And wasn't there a sort of quarantining of those infected? (Sorry, but most of my TB knowledge comes from stories about Doc Holliday and films like Tombstone - the ER Version) Seclusion, or institutionalization is pretty miserable to those infected, but wouldn't that mean exposure to the infected would have decreased and that may have kept numbers down as well. Like I said, I'm far from an expert, but improved water, food and shelter over a span of decades doesn't seem like a cure for TB.

Marcia Angell has had a vendetta against the Pharma industry for many years. Some of her hatred was earned, by actions of a few companies in sales and marketing that were pretty sleazy. Most of that's cleaned up now, thanks to case after case of very big fines for those who got caught.

Angell has never remotely understood just how difficult it is to design and execute large Phase III programs. She completely dismisses the value added by the information gained in these trials.

I agree that the problem of publication bias is real, and the attempts so far to deal with it (clinical trial registries, e.g.) are insufficient. But something like 3% of the NMEs that enter clinical programs in cancer ever make it to the market. The results from the other 97% clearly aren't being suppressed.

People don't understand that there are three possible outcomes to clinical trials:

1. It works
2. It doesn't work
3. The trial results don't allow us to draw a conclusion

The third outcome occurs very frequently, for a variety of reasons. In the media, outcome 3 is routinely reported as "It didn't work." This is extremely misleading. In my experience, most of the unpublished trials fall into category 3. I still would prefer to seem them published, and in a timely way, but I think the problem is less pernicious than many people assume.

Well done Tyler.

I found the publication bias component interesting, since in academia there is an inherent publication bias. Those academic doctors need pubs to further their career, and they know negative results do not publish. I know many statisticians in the industry: I have heard of few cases where pharma was pushing to publish something dodgy, I know plenty where a doctor was. So is the over-publication of positive results because of pharma or the doctors pursuit of academic careers?

The Indepent Institute's research director, Alex Tabarrok, was the person that posted this comment.

The graph said it all. NME is being tremendously punched out by Very High Production Cost. What's gonna be for us ordinary meds taker?

Comments for this post are closed