The invitation from the London Review of Books to review Ben Goldacre’s Bad Pharma™ reads:
“We were unsure, at first, what a review could add that isn’t already in the book – scrappy summaries and bits of praise are not for us. The book is of sufficient importance that the main thing is to get someone who knows what they’re talking about to present the material confidently.. frame the discussion”.
My head said it was inconceivable that the LRB wouldn’t take a review, even if it was at odds with the invitation to praise Bad Pharma. But my gut told me the inconceivable was about to take flesh.
Sure enough LRB turned down the review. Because, they said, their readers would be baffled by it – piquant given that LRB specializes in complex reviews of the esoteric and the obscure.
Now it is unfair to say that if Ben Goldacre didn’t exist, Andrew Witty, the CEO of GlaxoSmithKline, might have had to invent him. But it needs something like this to flag up how perilous our position is and how paradoxical – seemingly beyond the capacity of the editors of LRB to follow. This rejected review will puzzle some, and perhaps annoy those it doesn’t puzzle because it leaves answers for another day.
Ben Goldacre made his name with an earlier book Bad Science. Bad Pharma is equally well written. It will be accessible to anyone interested in health and its politics. It strikes the right note of outrage. It tackles the problems pharmaceutical companies pose to medical care and Evidence Based Medicine. The risk is that the solutions it proposes will aggravate rather than resolve our problems.
Data gone missing
The central problem it targets in the opening section is the clinical trial data missing in company vaults. The story is bang up to date, covering Roche’s withholding of key data on Tamiflu, leading governments worldwide to hand over billions of dollars of your money to stockpile Tamiflu in recent years. The more of this data academics like Peter Doshi and Tom Jefferson have got hold of the more questionable this decision to buy the drug has looked.
The next section deals with what is widely called regulatory capture. The regulatory agencies that approve drugs, such as the US Food and Drugs Agency (FDA) contain many bright and good people. But somehow the system fails patients and doctors, and simply can’t be trusted to ensure that medications on the market are safe and effective.
RCTs & EBM
The third section introduces randomized controlled trials (RCTs) which form the basis for evidence based medicine (EBM). RCTs and EBM sound as good as motherhood and apple pie to most people. After outlining how RCTs and EBM should work, Ben offers a compendium of ways in which companies game RCTs including selected reporting, selected patients, and selected comparators, in trials that are too short, too long or too small, that yield outcomes that are surrogate, composite, incomplete, or stem from subgroup analysis, with clinicians misled by talk of dramatic relative risk reductions when the absolute risk reduction is close to non-existent.
The big picture – RCTs are critical to good medicine and everything would be fine if only companies didn’t “game” them.
The final section entitled “Marketing” contains the juiciest details and the reddest meat. It outlines how advertising to patients works, how companies recruit celebrities to endorse products, sell diseases to sell medication, create patient groups, and employs PR agencies to get their message out and bury their critics.
It covers drug adverts in medical journals, the tune to which journals profit from drug company articles, the corrupting power of trinkets and free lunches, how sales reps track the personalities of doctors, and how companies like IMS Health sell data on the scripts that doctors write so companies know exactly what pitch will work with your doctor.
Ben’s searchlight switches to ghostwriting and how companies now run the continuing education of doctors. This leads to the knotty issue of conflict of interest. He draws complex distinctions between situations and behaviors, and the merits of greater or lesser puritanism, and to his credit is not an extreme puritan.
The book ends with a scene, a question and a plea.
- The scene is Ben being told at a party that Andrew Witty, GSK’s current CEO, is a good guy, who bangs the table in his efforts to ensure things are done better.
- The question is, how will industry respond to the challenge he poses?
- The plea is to any of us who have any other ideas on how to solve all this, to let him know.
Framing the issues
Many readers of Bad Pharma will know much of what is here from books such as Marcia Angell’s The Trouble with Drug Companies or Jerome Kassirer’s On the Take. Where is the problem in re-iterating it eloquently? The central problem is the premium Ben puts on controlled trials not found in other books.
Ben’s Marketing chapter shows how doctors just don’t get it. It confuses sales for marketing. Marketing means understanding doctors better than they understand themselves, and shaping their thinking so that the drug becomes the unquestioned answer to any problem a doctor faces. There is no better way to do this than with clinical trials in eminent journals.
Bad Pharma cites all the research showing that companies get the results they pay for but ignores the research showing that on all quality metrics, whether adequacy of trial design, transparency of procedures, or declaration of conflicting interests, industry trials routinely come out better than academic studies.
Ben cites studies about how doctors don’t believe they can be influenced by sales trinkets. Doctors say they follow the evidence. Ben believes they are influenced by the gifts (sales), and would prefer them to be influenced by the evidence. Andrew believes that swayed by books like Bad Pharma, doctors are increasingly influenced by the evidence (marketing). Andrew uses gifts as decoys.
Bad Pharma spots the sneaky little sales gnats and swallows the marketing camel, and while the advocates of evidence based medicine do just this, the only outcome can be growing perplexity at worsening indigestion and ever shriller calls for more regulation and policing of conflicting interests – which pleases Andrew.
Bad Pharma calls for more regulations but the more regulations there are the stronger industry becomes. The Chinese sage Chuang Tzu caught the dilemma beautifully in 323 BC:
“For security against robbers who snatch purses, rifle luggage, and crack safes, one must fasten all property with ropes, lock it with locks, bolt it with bolts. This is elementary good sense. But when a strong thief comes along he picks up the whole lot, puts it on his back, and goes on his way with only one fear; that ropes, locks and bolts may give way.”
The problem for healthcare worldwide goes back to a set of regulations – the 1962 amendments to the US Food and Drugs Act put in place following Thalidomide. These were later adopted everywhere else. They had three critical components to them:
1. Product patents supersize profits and risks
The first component involved maintaining the US system of product patents on drugs rather than the German system of process patents (Dancing in the Dark). Congressional staffers in 1962 had demonstrated that process patents led to more innovation in pharmaceutical development and produced cheaper drugs than product patents but instead the Germans were persuaded to change to product patents.
Universal product patents laid the basis for the development of blockbuster drugs pioneered by Glaxo two decades later. Blockbusters are drugs that earn a billion dollars or more per year for a company. They are not life-saving compounds. They are highly marketable products that are literally worth more than their weight in gold. This distorts company priorities. It puts a premium on hyping the benefits and concealing the hazards of a few unimportant drugs like: Prozac for depression, Lipitor for cholesterol lowering, and Fosamax for osteoporosis, all of which we could have easily managed without.
But companies can now go out of business if a significant hazard of their blockbuster comes to light. So company ingenuity is funneled into hiding hazards. Bad Pharma is silent on the role of patents in creating this perverse incentive. By putting the premium it does on clinical trials, it plays straight into pharma’s hands.
2. Prescription-only status addicts doctors
A second set of bolts put in place in 1962 was prescription-only status for medicines (Dance of the Sugar Plum Fairies). The idea that drugs should be prescription-only fits Ben’s worldview – no matter how health literate you are doctors have spent more time being educated in these things than you and they’re here to manage you.
But prescription-only was a mechanism designed to control addicts and it makes doctors rather than their patients the consumers of medicines. And pharmaceutical companies bring more money to bear on influencing this small band of consumers – $50,000 per annum per doctor – than other companies bring to bear on influencing an entire population. When they slip on the Ring of Prescription Privileges, doctors become visible to Eye of Sauron and shrivel.
The hazards of thalidomide came to light in Germany because it was over-the-counter there and doctors didn’t make a living out it. Ditto with tobacco and alcohol.
In contrast, the significant hazards of prescription drugs now take 15 – 20 years to be accepted. The combination of trials and prescription-only status has fashioned doctors into a risk-laundering service for companies. Bad Pharma mentions nothing about this.
Controlled trials were the third set of ropes and bolts put in place in 1962, and Andrew is praying they don’t give way (Dance to the Music of Time).
Louis Lasagna was the Ben Goldacre of his day. He was the first Professor of Clinical Pharmacology, the author of the version of the Hippocratic Oath that many doctors take on qualifying, and one of the earliest exponents of controlled trials. In 1962, almost by accident, he slipped the use of controlled trials as a test for efficacy into the FDA amendments and set us on the road to tragedy.
Regulations up till 1962 had been about safety. Forcing companies to show a drug worked for something was intended as an additional safety step. Eliminating drugs that don’t work seemed the perfect way to “first do no harm”. Except that’s not what happened.
All of a sudden this new scientific tool, whose uses and quirks are still not fully understood, became a tickbox for bureaucrats – if there are some positive studies regulators can let a drug on the market and no-one can criticize them afterwards no matter what goes wrong.
This is the point at which Bad Pharma makes a category mistake. It treats a box ticking exercise as though this could inform the practice of medicine.
Regulators deal with both food and drugs. All they want is a way, in the case of food, of determining if this yellow slab that could be lard colored to look like butter meets criteria for butter. If it does they can let companies label their product as butter. It’s not the regulators job to decide if this is good butter, what butter might be used for, whether it’s healthy for you – that’s up to consumer groups and physicians.
For drugs, the criterion is two positive trials. These regulations were not designed to have anything to do with the practice of medicine, they regulate adverts. The FDA in the US, along with regulators in the UK and elsewhere, are ticking boxes perfectly well – and as the law stands cannot do otherwise. It is hardly surprising that pharma have found many ways around the regulations. Consumer groups and physicians step up to the plate when it comes to butter but not when it comes to medicines. Who’s captured?
Introducing trials into the regulatory apparatus has created a mess. What’s been captured is the production of evidence and the more books like Bad Pharma fetish clinical trials the more captured medicine becomes.
The limitations of RCTs
RCTs can give the wrong answer
RCTs can be very helpful when they disprove claims that a drug works. This can rein in the unscrupulous huckster, the credulous physician and protect the vulnerable patient.
But RCTs can also be unhelpful. For many conditions they simply do not produce a meaningful outcome. For instance, where superficially similar problems can be caused by both drug and illness, such as suicidality on antidepressants, RCTs may perversely show that drugs that clearly cause a problem seemingly don’t cause it. Within the mental health domain it’s debatable if RCTs can show anything “works”. Using the clinical trial procedures used to bring Prozac on the market, alcohol could have been shown to be as effective an antidepressant as and safer than Prozac (Shadow Dance).
Rather than being used as initially intended – to first do no harm – trials have given rise to an efficacy fetish. As a result we are now all put on an increasing number of drugs that have been approved by regulators and there seems to be no way to limit the number of drugs we end up on. End up on because these drugs that work have not been shown to save lives. The result is that drug induced death is now the third leading cause of death. This is a tragedy of mythic proportions.
RCTs tell us almost nothing about cause and effect. They discover nothing. They likely block the discovery of many treatments. But for the fact that Barry Marshall in 1980 took a laboratory rather than a clinical trials approach to showing ulcers were caused by h pylori and could be cured by antibiotics, Glaxo would have used RCTs to bury the evidence that antibiotics cure ulcers to defend the very first blockbuster drug, their H2 antagonist for ulcers, Zantac.
More generally in a 400 page book about treating illnesses and drugs, Bad Pharma has only 3 pages (101-104) that come close to dealing with biology, even though biology is the primary driver of the superficial associations that controlled trials throw up.
Efficacy for something is another component of the tragedy. This element restricted companies to selling medicines to doctors for diseases only. In 1962, no-one foresaw that if constrained in this way companies would convert what had been a series of vicissitudes of everyday life and normal variation in terms of beauty and functionality into a set of diseases. And so we have all become depressed, osteoporotic, with hypercholesterolemia where otherwise we might have had burn-out, aging bones that could be managed by exercise, and a diet-related issue that is only significant against a background of more important cardiac risk factors.
Companies now speak the language of medicine to an extent that doctors fail to appreciate. Not only have vicissitudes been transformed into illnesses, but acute illnesses have become chronic and the moral imperative to treat that is brought to tuberculosis has been co-opted to sell products for restless legs and fibromyalgia. In 1962, the drive was to temper patient enthusiasm for treatment with medical skepticism. Fifty years later skeptical patients try in vain to temper the zeal of doctors following the supposed evidence embodied in the latest guideline based on published trials.
Fixing a bad system
Bad Pharma describes the superficial features of the problem but it fails to analyze how we got here. It just assumes Pharma is Bad and that everything else follows from this.
This analysis in contrast appeals to an historical moment – things that might have been otherwise. Some might argue the insertion of controlled trials into drug regulation was an accident waiting to happen. Others might say this is an unparalleled demonstration of the perils of interfering with the market. Whatever frame you opt for, the result is the strangest market on earth in which mega-corporations spouting a rhetoric of market freedom operate in a totalitarian way to control the least free market ever seen.
Ben locates missing data as the key problem and access to the data as the solution. This is a hot issue in Europe with the British Medical Journal leading a campaign for access to clinical trial data.
Is Andrew feeling nervous? Probably not. Companies have been managing this scenario for years. When recent calls for access to trial data emerged (access to trial), GSK immediately offered access to their data once the right academic board and governance is in place. With control of the analytic approach to the data, GSK can achieve both the result they want from “independent” analysis and the kudos of transparency all in one go.
Andrew has patient groups lined up – not paid for by companies as Bad Pharma would have you believe – to recast the argument in terms of academic dilettantes versus patient rights to privacy and confidentiality. For most casual readers there can only be one winner to this argument. Andrew has politicians on board with the need to manage the communication of risk in the public domain – and in a supreme irony on this issue the primers for politicians cite Ben’s Bad Science as their number one text.
It’s like pitting a major league outfit against a set of minnows. If the minnows win they are likely to make a bad situation worse. Good intentions aren’t enough. Unless the good guys understand the system they are up against they can make things a lot worse – just ask Lou Lasagna. Access to trial data on GSK’s terms will make the days of Bad Pharma look halcyon. The good guys don’t think there could be any adverse effects to good clinical trials. If the outcomes aren’t right, like a religious belief it will always be because the trials weren’t good enough, not because there are so many things for which trials aren’t the answer.
“If the outcomes aren’t right, it will always be because
the trials weren’t good enough, not because there
are so many things for which trials aren’t the answer.”
Missing data are critical but Bad Pharma misses the critical data. Andrew probably purrs at Ben’s dismissal of adverse event reports, at his missing company perversities around hiding adverse event data that are more astonishing than ghostwriting. For instance companies in the US now encourage patients and doctors to report adverse events directly to FDA rather than to the company. Enlightened? Very – companies have a duty to follow up adverse event reports and to establish causality whereas FDA doesn’t. Diverting reports to FDA saves money, reduces legal liability and relegates adverse event reports to the status of anecdotes (See AW and AW2).
The ultimate perversity is when in the face of hundreds of compelling reports of a treatment induced problem, when employees using standard causality algorithms have said their drug caused this problem, companies will under oath deny there is any link between the problem and their drug (Burn in Hell). Thanks to EBM, they can get away with this. How? Well built into RCTs is the notion that the only findings that count are those that are statistically significant. And because RCTs are designed to see whether one event is statistically significant or not – whether the drug works – they cannot properly look at what else the drug might be doing and so the safety of drugs which was in 1962 the primary goal of regulation has been eclipsed.
The fetishing of clinical trials means it’s now difficult for doctors or patients to believe the evidence of their own eyes. Restoring this ability – which clinical trials almost by definition attempt to undermine – is more important to the practice of medicine than access to data about “butter”.
“The plural of clinical trials is not medicine.”
The individual patient once sat at the top of the EBM hierarchy – but no longer – following a relentless repetition of mantras like “the plural of anecdotes is not data”. Well the plural of clinical trials is not medicine.
The candidate of hope
At times you have to stand back and admire how good Andrew is at his job. Is he really banging the table just to get GSK to be a bit more like what Ben would like? Ben calls for more industry whistleblowers but most of what he cites comes from industry whistleblowers in the US and very little from doctors anywhere. If I was one of the many ex-pharma employees whose whistleblowing cases have not worked out, I would feel bitter and ask just where have doctors been in all this?
The problems could be settled without trying to make Pharma good. Doctors could but don’t take the simple step of banning industry “trials” from their journals. Because the New York Times takes the checking of the primary sources of an article more seriously than the New England Journal of Medicine does, we might all be safer if clinical trials were published in the NYT rather than the NEJM – which is a damming indictment not of pharma but of doctors.
Bad Pharma is relentlessly British in its focus. Ben wants the Royal Colleges to stand up and make a difference. Within the terms of the debate he outlines, there is almost nothing Britain can do. The one thing that might rattle Andrew would be an exhortation to doctors and patients to start believing the evidence of their own eyes again – but this Ben cannot say.
Instead Andrew offered controlled access to clinical trial data. In response as the BMJ noted “[Witty] was singled out for praise by one of the industry’s harshest critics. Ben Goldacre called it a “cartwheel moment”.
The BMJ issue in which this comment features had a three-page spread on Andrew Witty featuring his image on its front cover in political poster format, resembling no-one more than Barack Obama, with a single word Hope underneath. Making the case for GSK, as Not so Bad Pharma, he says “the 100,000 people who work for GSK are just like you, right? I’m sure everybody who reads the BMJ has friends who work for drug companies… They’re normal people. Many of them are doctors”.
EBM like Freud has been wonderful in bringing the biases of both doctors and patients to light. But just as the Catholic Church has learnt there comes a point where routine dismissal of what people say to you because some of it is mistaken leads to a crisis, medicine needs to realize that EBM has led us into a comparable crisis.
Doctors need to realize their profession is at stake (Professional Suicide, Model Doctors). If clinical judgment and discretion no longer matter, nurses and others will be far cheaper prescribers and can more readily be ordered to stick to guidelines than doctors. Anyone can get you on a drug; it takes an expert to get you off. How does an expert know when or how to do this? Not from clinical trials.
Bad Pharma™ hypes the RCT and EBM brand, and denies the existence of any side effects. The only other place in medicine where this is done to a comparable extent is with blockbuster™ drugs. A Freud might think the™ revealing.
Lou Lasagna stands as an extraordinary symbol of how misguided this approach might be. He pushed for an inclusion of placebo-controlled trials in the 1962 Food and Drugs Act in the belief that establishing the efficacy of a drug would make a contribution to the safety of all us. But as of 1962, only one drug prior to marketing had been demonstrated in a placebo-controlled trial to be safe and effective – thalidomide.