*Bad Pharma*, by Ben Goldacre

Here is a simple sentence from Frank Lichtenberg, an economist who studies pharmaceuticals and a highly reputable researcher in the area:

This implies that the incremental cost-effectiveness ratio (cost per life-year gained) of pharmaceutical innovation was about $12,900.

Read the whole paper, and if you wish to go further, you can peruse his entire body of work.

I am thus a little nervous when Ben Goldacre entitles his recent book Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients.  (I have a UK copy, and it is due out in the U.S. this February.)  I do in fact agree with Goldacre’s portrait of a sector wracked with massive corruption and shoddy scientific standards.  And I see many aspects of this book as deserving an “A” or “A+” rating, which I would not hand out lightly.  But I won’t continue down that track, because I suspect the book will receive many very positive reviews, as indeed it did in the UK.

Could he not have called the book Not Nearly as Good as it Could be Pharma: How Corruption is Diminishing One of Our Great Benefactors?  Admittedly that does not roll off the tongue as nicely.

Or how about Slow Pharma: How to Get the New Drug Pipeline Up and Running Again?

Goldacre’s policy recommendations would in general raise the costs of research and development, although they would  likely improve the accuracy of research results and reduce over-prescription and overuse of drugs.  It is quite possible they would lower the rate of return to pharmaceutical innovation, likely I would say.  These trade-offs are neglected, and, much as I admire many features of this book, I cannot help but, alas with trepidation, call some of its central features “Bad Science.”  Bad Economic Science.  The morality of the narrative and the Platonism of his vision distracts him from presenting the policy trade-offs clearly.

Lichtenberg’s name does not appear in Goldacre’s index.  Nor does the word “innovation.”

Recommended, with or without prescription, but use with extreme caution.  And you should “compound” this with other books.

Addendum: I bought this book myself, which included Amazon shipping charges from the UK, and was not sent a free sample or visited by an attractive sales representative.

Second Addendum: There is some back and forth between Goldacre and me in the comments section.


I have the ePub version of Bad Science by Goldacre, and did not like the breezy, layman, strawman, verbose style, but he's got a few titbits like this:

{biblical 'clinical trial'} And Daniel said unto the guard…’Submit us to this test for ten days. Give us only vegetables to eat and water to drink; then compare our looks with those of the young men who have lived on the food assigned by the King and be guided in your treatment of us by what you see.’ The guard listened to what they said and tested them for ten days. At the end of ten days they looked healthier and were better nourished than all the young men who had lived on the food assigned them by the King. So the guard took away the assignment of food and the wine the) were to drink and gave them only the vegetables.
Daniel 1:1-16.

{placebo effect if doctor tells you you'll be OK} Theodor Kocher performed 1,600 thyroidectomies without anaesthesia in Berne in the 1890s, and I take my hat off to a man who can do complicated neck operations on conscious patients. Mitchel in the early twentieth century was performing full amputations and mastectomies, entirely without anaesthesia; and surgeons from before the invention of anaesthesia often described how some patients could tolerate knife cutting through muscle, and saw cutting through bone, perfectly awake, and without even clenching their teeth. You might be tougher than you think.

My theory about acupuncture anaesthesia is that it merely gives you "permission" not to freak out while surgery is performed. You can have some pretty major wounds without much pain. About a month ago, a glass jar fell off my kitchen counter, broke when it hit something on the floor, and bounced into my lower leg. I had a deep wound about 1 1/2 inches long that bled like crazy. I was cleaning up the glass and unaware I was wounded until I noticed I was standing in a rapidly expanding pool of blood. It looked like I lost about a pint of blood, but I felt no pain. Fortunately, the bleeding stopped after I taped some gauze across the wound, and it seems to have healed up just fine.

It actually seems to be burns, dental problems, and some kinds of tiny wounds that cause the most pain. I would not be surprised at all if I could tolerate having a leg sawed off without anaesthesia.

Here is an interview within him on the Science Weekly podcast: http://www.guardian.co.uk/science/audio/2012/oct/22/science-weekly-podcast-ben-goldacre-bad-pharma

Ben Goldacre here.

It would be interesting to know which of my policy suggestions you think would inflate costs, and reduce the return on investment.

The key argument of the book is that the results of clinical trials should not be withheld from doctors and patients. Industry spokespeople continue to deny that this happens at all, in the face of overwhelming evidence, in a testament to how easy they have found it to evade serious public discussion.

Sharing all trial results would not cost any more money, since the cost in a trial is in conducting it, rather than dissemination. The risk to industry is that doctors and providers of healthcare services would have a clearer picture of the absolute benefits of drugs, and of which is best, or most cost effective. So the downside would perhaps be that drugs might become more like commodities, as purchasers got better information.

This would be easier to make good money from good drugs, but harder to make money from successfully obfuscating the inadequacies of mediocre drugs. While that might be tough on companies with only mediocre drugs, it might be beneficial for innovation that success is rewarded much better than mediocrity and spin.

Would making researchers log all trials centrally before they are conducted help? Or is this already done? Shouldn't that ensure no publication bias?

Even under a full disclosure regime why should a drug become a commodity? Aren't today's patents long enough to protect against competitors?

For a start, if you don't know whether a path will produce good or mediocre drugs, and if suddenly you profit less from mediocre drugs, you will invest less up front in going down these paths overall. It is a marginal decision for many companies, which gets obscured with a phrase such as "While that might be tough on companies with only mediocre drugs". A simple test is whether the best pharma companies would favor your proposals, or rather think it raised their costs. I predict the latter. I don't doubt that your proposals also bring benefits, including greater accuracy and less over-use, and I do very clearly cite that in my review above. But your analysis is not considering this trade-off (among others).

I think Tyler's reasoning is bizarre. A farmer may want to market the Lemons in its harvest as Strawberries and sure this might make an average Strawberry a little cheaper but what of all those Lemons that got sold at essentially Strawberry prices?

It doesn't help society to encourage this marketing of Lemons as Strawberries.

In the long run, transparency makes farmers better at getting more Strawberries in their harvest and that's exactly what we need. And no, asking the company what it prefers is no test at all.

'...if you don’t know whether a path will produce good or mediocre drugs, and if suddenly you profit less from mediocre drugs, you will invest less up front in going down these paths overall.'

You are honestly defending a method that profits from creating mediocre drugs? I would have have thought that companies finding paths where they would create mediocre drugs less often would be better, both for patient health, innovation, and for pharmaceutical company profits, but then, I'm not an economist.

'A simple test is whether the best pharma companies would favor your proposals, or rather think it raised their costs.'

Seriously, Dean Baker has this area covered too. Maybe you should add him to your reading list? Though possibly, the term 'best' doesn't imply 'most profitable' in this context. Because I can't imagine any 'best' pharma company wanting to develop mediocre products except by the not exactly reassuring perspective of increased profits, regardless of patient benefits.

'But your analysis is not considering this trade-off'

A trade off which still seems to rest on the idea that developing mediocre drugs is what the 'best' pharma companies are interested in - which really leads one to wonder what that definition of 'best' is.

Having worked on both blockbuster drugs (Lipitor) and duds, I can assure you that going in, no one says, "Let's make a mediocre drug." What happens is that you try many, many targets and try to isolate the few that look to be the best combination of efficacy, side effects, and cost. Over time the company stops less promising lines of research and evaluates the surviving ones. Each step down the pathway from initial molecule synthesis to commercialized, mass produced drug costs a modest fortune. Eventually the best few make it to market. Up front there is very little information about which drugs will be mediocre and which will be good. When a drug starts to look iffy, it normally gets cut and it tends to be decades before we know for sure if it could have saved thousands or millions of lives.

Consider it like this: a drug company begins with 100 drug lines. Of those 50 will die at the first cut, costing perhaps $250K each. The next 25 will die at the next cut, costing say $1M each (total). 12 will fall out at the next cut costing around $4M each. 7 don't make the next cut (costing around $16M each). Everything from here on out actually gets commercialized to one degree or another. Figuring that actually mass producing the drugs costs about $10M each and you end up with total up front costs around $353.5M. You have 6 drugs that you actually sell. One of them makes you $250M before going off-patent, 2 of them will do so-so ($40M a piece), and 3 of the others make less than their development costs ($20M each). All told, the company makes $36.5M profit. (Note these are mostly conservative WAG estimates, real drug development costs have been estimated to be about $1.7 billion by venture capital firms).

Now suppose you know going in that you will make only 250/30/10 each because mediocre drugs make less profit under the new regime, well now you lose $13.5M each run through the full pipeline.

But couldn't you just be harsher and reject more the drug candidates? Well, yes that is what the industry is doing, the downside is that this limits your risk (both from selling mediocre drugs and getting sued for something like Vioxx) but also limits your potential for blockbusters. You have to spend a small fortune for each drug before you can possibly have a clue if it is a good or a strong drug. Anything that tips the equilibrium toward killing drug lines sooner means we end up with fewer drugs.

How about charging more for the winners? Perhaps doable, but let's say we charge more for our one good, strong drug, is society really any better off than what we have now? No matter how me-tooish the mediocre drugs, there will be people for whom they provide a benefit. Of course, we are already dealing with very limited competition thanks to patents so drug prices are probably already pretty near having monopoly pricing and too much higher means that people will be priced out of the market.

How about take a lower rate of return? Well then you run into the problem that investors will just sell or liquidate your drug company and poor the money into something with a better profit margin (like say a new social media venture). Society really benefits when somebody (like a pension plan or a new software magnate who just sold out) with a billion bucks says to themselves - I want to park this capital and make a profit, let me do that in life-saving drugs. There are literally thousands of drug/treatment ideas kicking around that we can't pursue because the payouts are huge, but the chances of success are slim (for instance I have worked a bit on a fun idea to cure cancer; at best it has a 1% chance to effectively cure every rapid growth cancer known to man). The harder you make it turn a profit developing life saving drugs, the more money will move from drug research into other areas (like say video games).

Seems we need a definition of "mediocre drug."

A drug that isn't great, but is a significantly better treatment than th ebest alternative shouldn't be considered mediocre. I'd say that mediocre means it offers little improvement in either cost or results. Isn't that the set of drugs that will go out of use if their benefits or lack thereof are widely understood?

Suppose you continue to sell such a drug, using misinformation to promote its use, on Tyler's argument, and yours, that the revenue is needed to encourage R&D. It seems to me that you are arguing that the amount of R&D that will be done otherwise is suboptimal, and that the solution is to make some sick people give the pharmaceutical companies money for useless products, which will stimulate R&D spending.

Is that really what we want to do?

Perhaps rather than patent protection, drugs should be subject to government "grand challenge awards". So the government could say, "I will award $10B to the first company to develop a Lipitor successor that is X% more effective (as established by something like the FDA) who's formulation will be put in the public domain".

I'm not necessarily opposed to government rewards, but for pharma you have to choose between impossible-to-meet thresholds or thresholds so vacuous that anyone can win.

We probably aren't going to find a drug that lowers cholesterol as good as (let alone better than) Lipitor without serious side effects. (As is, some people really hate Lipitor's side effects.) Then you have to try to legislate what side effects are good or not.

Imagine a drug that doubled your risk of lung cancer while halving your risk of a heart attack. That's a pretty serious side effect, but for people with low risks of lung cancer and high risks of heart attack it is probably a godsend. How can you write a challenge that allows this?

I'm not sure I understand your argument here. You seem to be saying that allowing companies to push mediocre drugs has the beneficial effect of encouraging R&D, thereby producing more good drugs, because mediocre products will still be somewhat profitable. Is that right?

If so, I have a few observations.

First, you are paying for these new, better drugs by providing poor treatment via mediocre drugs to lots of patients. Can that be right?

Second, the assumption of fewer good drugs under the changed regime is not obviously correct.

If mediocre drugs are suddenly unprofitable, because they are widely known to be mediocre, then won't good drugs become more profitable, by virtue of facing less competition?

Further, while you may not know at the start what a path will produce, you probably do know once you're some distance down that path. If the mediocre products aren't going to generate profits then you can stop them early, and devote the resources to more promising efforts. That could mean less R&D expenditure, with not much reduction in the quantity of good product produced. How can it be a good idea to encourage expenditure on likely mediocre drugs by allowing them to be misrepresented?

When you develop a "good" drug (e.g. Lipitor) that has a major health impact and becomes vastly prescribed (and hence earning the drug company lots of money), you virtually always end up with several drugs that can improve conditions for some people - but won't have enough impact to be vastly prescribed. These latter drugs are the mediocre ones, they may not lower your cholesterol any more than Lipitor, but they have different side effect or drug interaction profiles.

If you make it harder for those types of drugs to earn money, like say not granting FDA approval for "me-too" drugs or having extremely harsh conditions for getting the mediocre drug over the a more cost effective alternative, you make it harder for the drug company to recoup its costs.

"First, you are paying for these new, better drugs by providing poor treatment via mediocre drugs to lots of patients. Can that be right?"

We aren't paying for R&D by foisting ineffective pills on people, we are defraying a portion of R&D costs by allowing people & their doctors the option of prescribing from a wider selection of drugs. For every truly "poor" treatment you stop, you will stop dozens of incrementally better ones (though perhaps not cost effectively better); this in turn does mean that capital will leave the pharmaceutical sector in search of more profitable ventures.

"If mediocre drugs are suddenly unprofitable, because they are widely known to be mediocre, then won’t good drugs become more profitable, by virtue of facing less competition?"
No, the opposite actually (at least until there are fewer drug companies). The more restrictions you put on what is profitable, the more likely you are to end up with convergent research. Past a certain point, it becomes vastly more profitable to use corporate espionage and race the competition resulting in far, far fewer good drugs.

"Further, while you may not know at the start what a path will produce, you probably do know once you’re some distance down that path. If the mediocre products aren’t going to generate profits then you can stop them early, and devote the resources to more promising efforts."
This is EXACTLY what we do now, the problem is getting the data needed requires things like Phase II clinical trials. That takes huge amounts of time and huge amounts of money. Some of the biggest killers of promising drugs are unforeseen side effects. Something that ends requiring a box warning just about automatically loses half (or more) of its profit potential; there is zero way to see that coming before clinical trials.

Now I'm not saying that the current mix of profit incentives are the best option out there, but seriously, That Ain't No Such Thing As A Free Lunch. You want fewer mediocre drugs? Fine. You will pay more for the good drugs or you will have fewer of them. There is a trade off whenever you make it harder for a company to turn a profit. Sometimes that trade off is still worth it, but don't be obtuse and say that drastically changing profit incentives won't have impacts.

So do we want to encourage crappy cars, poor bridges, and bad food as well, because each of those encourages larger markets, and hope larger markets mean more investment, and hope that more investment means better drugs?

Note that with the exception of the bridges, generally people at least have opportunities to choose. With drug choices, basically, there is none, unless you are extremely wealthy with a condition that allows time to think, a doctor or self-medicating.

That reasoning also assumes there is a market function in US healthcare. Which, of course, there is, but the end consumer is less a participant than a product.


you seem to be suggesting that it is good for companies to be able to make money from mediocre drugs, by hiding evidence of their mediocrity, because they might invest more in R&D if they know they can make money even from a substandard product. The idea seems to be that softening the risks will boost innovation, or that subsidising R&D by giving money to failures will encourage innovation.

This seems tenuous to me. Rewarding success, rather than failure, is surely more likely to stimulate innovation in the long run.

It is also important to be clear – though this does not necessarily affect your position - that there is more than money at stake. If clinical trial results are distorted, for example, so that doctors are misled into using an expensive new drug, where an older cheaper one is more effective, then money is wasted; but patients are also actively harmed, by being given a less effective treatment, and deprived of the best one.

The point (which was already made above, by the way) is that nobody starts out saying "let's make a mediocre me-too drug!"

Pharma development is not like playing with Lego blocks. It is not the kind of "add baking soda to vinegar, boil water to extract salt from solution" chemistry you learned in grade school. If anything, it's like going into a casino and putting one quarter in every slot machine. Once in a great while, one of the machines pays off big. Occasionally you'll win enough to break even. But what usually happens is that *none* of them win and all your quarters are wasted.

" If clinical trial results are distorted, for example, so that doctors are misled into using an expensive new drug, where an older cheaper one is more effective, then money is wasted"


this is what's known as major medical fraud

people go to jail forever over that kind of thing

there do not need to be new laws or "market disincentives" about it

Publishing all clinical trial data would cost little and increase the signal-to-noise ratio, it is true. However, it is important that this not make it harder for drugs to get approved.

Data fudging and publication bias didn't occur in a vacuum - they were part of an arms race as FDA standards for approval got increasingly onerous and expensive. If we start publishing all the clinical trial data without adjusting the FDA approval process, we'll simply choke off the flow of innovation at its source.

Cleaning up publishing AND fixing the regulatory regime would be a great win. Doing either alone would be a grave mistake.

"Data fudging and publication bias didn’t occur in a vacuum – they were part of an arms race as FDA standards for approval got increasingly onerous and expensive"

- probably the best comment in the whole thread

"This would be easier to make good money from good drugs"

should be

"This would make it easier to make good money from good drugs"

"it might be beneficial for innovation that success is rewarded much better than mediocrity and spin."

- now that's a _Good Sentence_

Tyler mentions this alternative title:

"Slow Pharma: How to Get the New Drug Pipeline Up and Running Again?"

I suspect that the answer to that question is that the low-hanging fruits of making-people-better-by-making-them-ingest-bulk-chemicals have been picked. Pharma is broken because there is a limit to how much you can fix the body subject to the constraint of finding a druggable target within the disease mechanism which can be drugged with a compound that has no side-effects even though you intend to soak the entire body(!) in it.

I suspect that there will be a revolution in healthcare outcomes and therefore in cost-effectiveness ratios when science learns how to move beyond the bulk-chemical-drug paradigm. Already we are learning how to attach a drug or even a radioactive atom to an antibody which selectively targets something specific in the body before it releases its payload. We need to move forward towards making artificial macrophages, to understand signalling pathways between cells, to treat diseases more like a robotics/computer science problem than a chemistry problem.

Of course, Ben's point about rewarding real innovation still stands. Maybe if we decrease the return on fake pharmaceuticals, some drug companies will invest more than 13% of their revenue in R&D.

What we really need is a sort of 21st century "Bell Labs" of pharma where they go and invent the future. Maybe the economists can work out a way of making that kind of thing pay off better than short-term bullsh*t like "me too" drugs and manipulating clinical trials.

I don't think we have picked all the low hanging fruit. I think we, and by we I mean American trial lawyers, the FDA and European socialists, have made the development of new drugs nearly impossible.

The problem is that "no side effects". Has any drug ever been produced with no side effects? Are we better off with the insane approach we have to risk? Well, put it this way, neither aspirin or penicillin would be approved these days. Aspirin kills guinea pigs and so wouldn't even make it to a human trial. A small number of people are allergic to penicillin and so it wouldn't pass a human trial. Have the benefits of these drugs out weighed the loss of human life? I would think so. Which suggests that the rules are wrong.

Take fen-phen. A now forgotten drug that helped people lose weight. That caused damage to heart valves and may have contributed to some small number of deaths - on the order of dozens rather than thousands. Wyeth, the manufacturer had to pay out some $20 billion - bearing in mind that when the company was taken over later, it was only worth some $60 billion. It was a serious payout.

The point is the FDA approved it. Eventually. It could have helped hundreds of thousands of people who are going to get other serious medical conditions because of their weight. Millions of people these days. But it was pulled from the market and the company pushed into take over.

The result is that if you're a pharmaceutical company, you are insane to work on drugs that actually help sick people. Sick people die. Work on something that helps healthy but unhappy people. Prozac. Viagra. Vioxx. Well maybe not. Even that approach is not working out so well these days.

So, great, we probably could develop more drugs using the antibody approach with radioactive attachments. But we won't. Because anyone who does will be sued out of existence. We could go the GM route and tailor individual drugs to patients. But we won't. John Edwards made very good money suing doctors because patients killed themselves or were born with cystic fibrosis. No one in their right mind thinks doctors were responsible in most cases. Maybe all cases. He nearly became President. The media loved him. Juries loved him. The science was never on his side, but it didn't matter. So you're going to put something radioactive into someone's body and three decades from now they are going to get cancer, and once some nice telegenic lawyer with great hair mentions the word "radiation" or "genetic mutation", you're looking at a nine figure payout.

It ain't going to happen. I usually admire Ben Goldacre, but this sort of piling into the pharmaceutical industry simply makes our future clear - we will have fewer and fewer new drugs, eventually resulting in none. Diseases will continue to become more and more resistant to the drugs we have. But no new ones will be made. More and more people will die of things that were curable in the 1960s before Thalidomide. And people like Ben Goldacre will write books blaming the now no-longer-existent pharmaceutical industry for being greedy selfish capitalists.

Perhaps someone like Obama will fight a War on Disease, that will be just as successful as the War on (fun) Drugs and the War on Homelessness. Perhaps they will even call it something like Bell Labs. But it won't work as long as the structural problems remain.

To channel Glenn Reynolds:


Throughout history, poverty is the normal condition of man. Advances which permit this norm to be exceeded — here and there, now and then — are the work of an extremely small minority, frequently despised, often condemned, and almost always opposed by all right-thinking people. Whenever this tiny minority is kept from creating, or (as sometimes happens) is driven out of a society, the people then slip back into abject poverty.

This is known as “bad luck.”

I think there is some truth to this perspective, too.

It would probably be best to drastically reduce government regulation of drugs, and to rely more on government ratings panels which would rate a drug on a scale of 1-10. People could then make their own minds up. Of course you would still have to ban drugs which were obviously fraudulent and were designed to take advantage of people's stupidity; note that before the FDA etc existed in the USA, people were fraudulently sold all kinds of "Elixirs" and other nonsense.

I agree it is absurd that people expect drugs to have no side effects. But it is entirely reasonable for people to expect that information about the safety and effectiveness of medicines should not be actively withheld from doctors and patients.

You then go on to present a rather melodramatic and apocalyptic vision. I have merely suggested that drug companies should stop withholding information about the safety and effectiveness of their drugs from doctors and patients. I don't believe that this will present a barrier to innovation, driving us back to the stone age: drug companies should be able to make lots of money from the best drug in their class, and they shouldn't expect to make money by withholding evidence showing that a drug is less effective. This is a perfectly simple and reasonable position, and it is no sense a wild attack on industry, the profit motive, or innovation.

Speaking as a former biotech executive (and current biotech VC), I don't think any good case (I almost said any reasonable case) can be made against clinical trial transparency. The incremental cost should be minimal, and the benefits tremendous. I'd suggest FDA require full disclosure of clinical trials as a condition for any product seeking US registration. So i think Mr. Goldacre's has that one nailed precisely. I have yet to read his book, but I will say the tone of the introduction, gives me reasons for concern along the lines Tyler Cowen indicates. If my day frees up I can provide some examples. And Mr. Goldacre, thanks for joining the conversation here.

Thanks for your unsourced opinion. It could well be you are a biotech exec who profits off others innovation--being second to market and/or a patent infringer is quite lucrative, as even Big Pharma will tell you. I say there should be no need for full disclosure.

"I am thus a little nervous when Ben Goldacre entitles his recent book Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients."

Even Tyler Cowen can be guilty of mood affiliation :)

I am also guilty as charged. As a long time fan of Ben Goldacre, I was disappointed in the title (thought he had sold out) and passed on the book when it came out. Guess I'll buy it now.

Why don't companies already do it?

Why is this such a big problem in the industry arguably where we do the most to combat it?

They do! See my post below.

FWIW, I spent virtually my entire working career in the pharmaceutical/biotechnology industry (with a short stint at the National Institutes of Health). Transparency of clinical trials is already a reality. Companies are already posting all trials on the National Library of Medicine website: http://clinicaltrials.gov/ As per the Principles for Clinical Trial Research developed by the Pharmaceutical Research and Manufacturers of America (I had a hand in developing these), companies post the results of these trials once they are completed (http://www.phrma.org/about/principles-guidelines/clinical-trials). This is also a requirement imposed by all of the major medical journals that publish papers on advances in medicine. This is an issue that was put to bed some time ago and most of what anyone is writing about is just rehashing old stories.

A major difficulty for pharmaceutical companies is that new drug targets are proving to be elusive and for many diseases we have therapies that are highly effective (and increasingly becoming generic drugs thus lowering the cost of health care; which is why the Medicare drug benefit is not costing as much as was predicted). Companies discontinuing research in a number of disease categories for these reasons (very little is being done in cardiovascular, arthritis and rheumatism, allergy and immunology, and most disturbingly anti-infectives).

Nobody starts out to develop a mediocre drug. It just won't be profitable. Insurers, whether private payers or the government only want to pay for what works. There also are not any "me too" drugs despite what one reads. In virtually all cases where this is mentioned, it is a consequence of simultaneous development by a number of companies with one company getting regulatory approval first. Lipitor, the best selling statin in the class, was about the fifth or sixth approved. Should it not have been developed after the first one was approved.

Finally, in about 95% of the cases it is not the FDA (or any other sophisticated regulatory agency) that is at fault. It is either an inept company (there are more than you think, particularly in the biotech sector) or the drug just doesn't work (or has too many safety problems).

Don't know that I will get the Goldacre book; I've read too many of them. One should probably wait for John LaMattina's book, Devalued and Distrusted: Can the Pharmaceutical Industry Restore its Broken Image (http://www.amazon.com/Devalued-Distrusted-Pharmaceutical-Industry-Restore/dp/1118487478/ref=sr_1_2?s=books&ie=UTF8&qid=1354295470&sr=1-2&keywords=john+lamattina). His earlier book, "Drug Truths: Dispelling the Myths about Pharma R&D" is good reading.

Goldacre claims that only one in five clinical trials that were supposed to be posted to clinicaltrials.gov were actually posted. What is your response to that?

Source: http://www.badscience.net/2012/10/gsk-have-promised-should-we-trust-them/

He's wrong. His agenda and reality don't coincide. It's that simple. You can go to the database and see for yourself. Inconvenient truth for Goldacre.

I did go to the database, but found only links for finding specific topics. Where do I find the statistics for the database as a whole, including posting percentages?

It might also be helpful for me to post Goldacre's source, cited in the article I linked, from the British Medical Journal.


Abstract ("Results" section):

"The ClinicalTrials.gov registry contained 83 579 entries for interventional trials, of which 5642 were completed within the timescale of interest. We identified trials as falling within the mandatory reporting rules if they were covered by the FDAAA (trials of a drug, device, or biological agent, which have at least one US site, and are of phase II or later) and if they investigated a drug that already had approval from the Food and Drug Administration. Of these, 163/738 (22%) had reported results within one year of completion of the trial compared with 76/727 (10%) trials that were not subject to mandatory reporting (95% confidence interval for the difference in proportions 7.8% to 15.5%; χ2 test, P=2.6×10−9). Later phase trials were more likely to report results (P=4.4×10−11), as were industry funded trials (P=2.2×10−16)."

Thanks for posting the link to the BMJ study Jeff.

Sadly it is very common to see people from industry denying simple facts, such as this proven failure to comply with the clinical trials.gov legislation.

John LaMattina himself has also falsely claimed that this law, which has been almost universally ignored, has fixed the problem of missing trials.

This is the clearest evidence that the industry is, for done reason, not yet able to engage in serious discussion about how they can join a modern era of open data and transparency.

I hope this will change, but the change will only come from involving the public in the discussion: voluntary codes, promises, and negotiations behind closed doors have sadly failed.

Goldacre reported that "In 2007 the FDA Amendment Act promised that all trial results would be posted online at clinicaltrials.gov within a year; but in 2012 it was shown that only one in five trials had met this commitment."

His source was http://www.bmj.com/content/344/bmj.d7373, which used the database at clinicaltrials.gov and reported that of studied trials falling within the mandatory reporting rules, "163/738 (22%) had reported results within one year of completion of the trial".

"Orange14" claims that these issues have been fixed. This is a false claim.

He says that trials must be registered on a public registry before publication. This is not true. The International Committee of Medical Journal Editors claimed they would enforce this principle, but in reality they did not. A half of all trials published after this rule was passed weren't properly registered, and a quarter weren't registered at all.


He says that all trials must be published on clinicaltrials.gov. In fact this law has also been ignored. An audit was published in the British Medical Journal this year: only one trial in five has complied.


These are fake fixes that have not been implemented. The problem of missing trial data persists for the simple reason that industry representatives, such as "Orange14" - I trust he will tell us his real name in his response - continue to deny it, in the face of all the evidence.

This is in many respects extraordinary behaviour, and although I'm being naive, it is hard to picture any other serious industry in which people would so casually state that black is white and night is day on vitally important matters of patient safety. The only thing that will move this discussion on is the public, and journalists, become sufficiently aware of the basics of the issue that such obviously false claims are no longer possible in a serious public forum.

"Transparency of clinical trials is already a reality. Companies are already posting all trials on the National Library of Medicine website: http://clinicaltrials.gov/ As per the Principles for Clinical Trial Research developed by the Pharmaceutical Research and Manufacturers of America (I had a hand in developing these), companies post the results of these trials once they are completed (http://www.phrma.org/about/principles-guidelines/clinical-trials). This is also a requirement imposed by all of the major medical journals that publish papers on advances in medicine. This is an issue that was put to bed some time ago and most of what anyone is writing about is just rehashing old stories."

Apologies for missing words, am on a poor connection!

You are a terrible writer as well. Verbose, breezy, and prone to gossip, based on your book Bad Science. But at least I did not pay for it (got it free at Piratebay--that's 'full transparency' for you--I hope you like the New Free ideology as applied to authors and not just Big Pharma).

Laying into Ben over his writing style is irrelevant to the point we are discussing. Even if he is the worst writer in the world, he may still be correct about missing clinical trial data.

Nice cherry picking of the data Mr. Goldacre! The JAMA paper was from 2009 and of course the data don't look good because it includes a lot of trials that were undertaken PRIOR to the implementation of the FDAMA law in 2008. It's largely irrelevant to the current practice by all researchers in terms of posting to the NLM database. If you want to continue to argue bout non-compliance for studies completed prior to 2008 that's your right but it's also dishonest.

The key quote from the BMJ article is: "We have no data on the number of studies that applied for exemptions from the requirement for reporting. Industry funded studies are perhaps most likely to form part of an application for licensing/marketing, and therefore most likely to apply for an exemption, but we have found that industry funded studies are more likely to report results than are studies funded by other means. We also note that trials investigating new indications for drugs previously approved by the FDA are not required to upload results until two years after completion, but our methods did not allow us to identify these trials."

What are these exemptions? Also, two years seems an excessively generous period to upload data; is that really necessary?

Orange14, do you have links to better data than what Goldacre cited? If not, I don't see how you can claim that the problem has been "put to bed".

I agree. This is a remarkable response from an unnamed anonymous industry executive who has earlier claimed that the clinicaltrials.gov regulations have fixed the problems of researchers and industry keeping trial results secret. One in five posted results within a year. What delay are we expected to believe is acceptable, for patients to be misled about the overall benefits of a treatment they are taking? He denies that four out of five are withheld - with no evidence - but what proportion of trial results does "Orange14" think it is acceptable for people to withhold from patients?

Your ad hominems aren't much better. Many people participate here by pseudonym.

(Not that I'm convinced by Orange14, either. Each of you is waving around studies and calling the other side a liar, making it very hard for a dispassionate third-party to decide which of you to believe.)

"There also are not any “me too” drugs despite what one reads. In virtually all cases where this is mentioned, it is a consequence of simultaneous development by a number of companies with one company getting regulatory approval first."

No. 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. True price discrimination. So let's not claim that the issue of "me too" drugs is a ghost, rather it is a total fraud.


Thanks Orange. I came to write the exact same.

"Nobody starts out to develop a mediocre drug. It just won’t be profitable." Derek Lowe also says this. Everybody *tries* to hit the one-in-a-million jackpot, but the sheer difficulty and unforeseen obstacles prevent this. Things don't scale linearly, Phase II superstars turn into Phase III flops, etc.

Even so, a mildly different drug (in terms of efficacy) with a side-effect profile that is unrelated to the first drug can be quite beneficial.

Tyler: "A simple test is whether the best pharma companies would favor your proposals, or rather think it raised their costs. I predict the latter."

No need for predictions, this is obviously true right now. Big pharma doesn't want transparency etc. because it would raise their costs (or lower their revenues, whichever you prefer). Ben clearly believes this, as he has written in the comments section and in his book. The industry is apparently opposed to his proposals. Tyler is saying this is prima facie evidence that Ben's proposals would be in some way costly to the pharma industry.

Some commenters seem to think Tyler is defending the industry. He's not. He's only saying a cost-benefit analysis needs to be done by Ben, not just a cost analysis (or just a benefit analysis). Ben's book is a great analysis of the cost of bad science as practiced by Big Pharm. But Tyler is suggesting there is a benefit from rent-seeking, e.g. maybe some of those rents are plowed back into R&D, etc. etc. That is not a defense of the industry, it is a call for good science.

*MAY* be a benefit from rent-seeking

Is there any data available on whether or not it would actually increase their costs? If you assume that there are two companies doing research on 100 possible drugs of which 10 are viable products then if they work in total secrecy until a drug is ready to market then each would do 100 trials and end up with 5 products (on average). If they announce failed trials as well as successful ones then each company would do 50 trials and end with 5 products (again, on average). Clearly this would save them money. This is obviously extremely simplified and I'd expect real-world benefits to be significantly less, largely because the pool of potential drugs is so large. However it doesn't seem quite so cut and dried that it should be considered automatically true.

I would have thought that the increased uncertainty would be more of a problem for the drug companies. At the moment there's room for several competing products on the market, but that's at least partly an artifact of limited knowledge. If we know definitely which is best then only one is going to make serious money. That would mean that if a company had an unlucky couple of years then it could be in trouble, and you'd expect to see some management changes.

What a prick!

Sorry. Wrong thread.

Here's another idea: government targets broad, explicit health goals: a combination of longevity, Quality Adjusted Life Years and other objective data - national or even global. It issues and backs a large number of tradable non-interest bearing bonds, redeemable for (say) $1m each once the target health goals have all been achieved and sustained. The bonds would be floated on the free market and would fetch a low price if the goals were ambitious and long term. As steps are taken to improve health, the value of the bonds rises in the market. The upshot is that government rewards successful initiatives for improving health (including, but by no means limited to, new drugs) regardless of how these initiatives work, or who implements them. Government still articulates society's broad desired health outcomes, and still raises the revenue for their achievement. But it contracts out the achievement to a motivated private sector in a way that rewards success, and only success in improving health. Such a bond is one application of the Social Policy Bond idea (socialgoals.com), would stimulate diverse, adaptive ways of achieving goals and refocus efforts on the long term.

Comments for this post are closed