Deadly Medicines and Organised Crime

by Peter Gotzsche

  We had access to both the protocol and the publication for 44 industry-initiated trials and used our sample to study ghostwriting.20 We didn’t find any trial protocol or publication that stated explicitly that the clinical study report or the manuscript was to be written or was written by the clinical investigators, and none of the protocols stated that clinical investigators were to be involved with data analysis. We found evidence of ghost authorship for 75% of the trials, which increased to 91% when we included cases where a person qualifying for authorship was acknowledged rather than appearing as an author. In most trials, the ghost authors we identified were statisticians, but we likely overlooked others, as we had very limited information to identify the possible omission of other people who would have qualified as authors. The trial protocol is an important document, but only five protocols explicitly identified the author. None of these people – all of whom were company employees – were listed as authors of the publications or were thanked in the acknowledgements, even though one protocol noted that the ‘author of this protocol will be included in the list of authors’. The ghosts operate in complete darkness, it seems, and shy from the light.

  A good way of reducing the prevalence of ghosts and guests is to write in the paper who did what, like film credits. This idea was coined by Drummond Rennie in 1996 and the Lancet was the first journal to introduce it, in 1997.21 Here is an example:

  ‘Ms. Yank first conceived of and designed this study; collected, analyzed, and interpreted the data; and wrote the article. Dr. Rennie assisted with refining the concept and design, assisted with data collection, and critically revised the article for important intellectual content.’

  According to internationally accepted criteria for authorship, persons listed as authors are required to have made all of the following substantial contributions: (1) conceived and designed the paper or analysed and interpreted the data; (2) drafted the paper or revised it critically for important intellectual content; and (3) approved the final version of the paper before publication.22 These criteria made it possible to study if authors of original research articles in Lancet papers deserved authorship. Although Yank et al. used a very conservative definition of guest authors, 44% of the authors did not fulfil these lenient criteria for authorship.22

  Studies that rely on what people tell you will underestimate the problems because of social desirability bias. Nonetheless, one such study reported 13% ghost authorship of papers published in six major medical journals and 21% guest authorship.22

  David Healy has described how frank some companies are towards doctors. ‘We have had our ghostwriter produce a first draft based on your published work. I attach it here.’ When Healy was unhappy with the glowing review of a drug and suggested changes, the company replied that he had missed some ‘commercially important’ points and published the paper in another academic’s name.23

  When a little light shines on the ghosts, it is usually in the form of ‘XX provided editorial assistance’, which means ‘XX wrote the paper’, and when only a trace of light comes through the shadows, we are told that the authors thank XX for help. With what? Cooking coffee while the overburdened clinicians analysed the data? Hardly.

  The marketing machine

  With an abundance of flawed literature at hand, it is not difficult to let marketing do the final kill, and even without such literature, marketing works. What is likely the most notorious example of this in medical history is related to treatment of stomach ulcers. Fifty years ago, ulcers were often treated by surgery, but James Black from the US company Smith Kline & French invented cimetidine, a drug that reduces stomach acid. It came on the market in 1977 under the name Tagamet,24 and Black was awarded the Nobel Prize.

  The company’s success was to be beaten by Paul Girolami, a financial controller who had worked his way to the top as CEO of the UK company Glaxo. Glaxo was mostly known for milk formulas for infants and had no operations in the United States. In 1983, Glaxo marketed a very similar drug, ranitidine (Zantac), with a highly unusual strategy. Instead of offering a lower price than Tagamet, the price was about 50% higher to suggest to people that it was a better drug. It wasn’t, but Girolami launched one of the most expensive and aggressive promotional campaigns ever seen. He hired drug salespeople who already worked for Hoffmann-La Roche in the United States and literally exploded the ailment heartburn. Gallup was paid to survey Americans and dutifully came up with the result he wanted, that almost half of Americans suffered heartburn each month, which led to the campaign Heartburn Across America. Glaxo also hired a celebrity, an actress who told the public how Zantac had helped her.

  Already 3 years after the launch, Zantac surpassed Tagamet in sales and became the best-selling drug on earth, and Girolami was knighted by Queen Elizabeth.

  This looked more like an evil plot concocted by an imaginative novel writer than reality, but unfortunately it was real, and it showed to the world that even research at Nobel Prize level couldn’t beat marketing. It changed the drug industry forever after, as they say in fairy tales, and marked the beginning of an era with a terrible waste of taxpayers’ money on industry marketing and very little innovation.

  Drug companies institutionalised deception,24 and Pfizer won the race to the ethical bottom. Right from its foundation in 1849, the company has shown a knack for getting people to take more medicine, so it’s no surprise that Pfizer became the world’s biggest drug company. When its CEO retired in 2000, he said that he recently bought a boat but as he had nowhere to put it, he bought a marina too.24

  The tight information control goes under the radar of most doctors, but their patients may think otherwise:25

  ‘My patient scanned the prescription I had handed her, then idly glanced at the elegant ballpoint pen I had used to sign it. The same brand name appeared on both. She said nothing, but I knew what she was thinking.’

  General practitioners rely on the drug industry as their main information source.11,26,27 In one study, 86% of them reported seeing drug salespeople,27 and in Australia, 86% of the specialists had seen a drug salesperson within the last year.28 Free samples of drugs are usually left behind after such visits,29 and such samples are highly effective in getting doctors to use expensive drugs. This explains why the value of the samples amounted to about one-quarter of the industry’s total marketing costs in 2004.30 It is a nice gesture to give away a pill box for free, but some doctors actually sell them to their patients or bill their government for them.31,32

  Doctors are surprisingly naïve and don’t realise how much they are being manipulated. Most doctors believe the information they get from the industry is helpful for them.27,33,34,35 When interviewed, they question the objectivity of the industry, but nevertheless consider the information to be factually accurate and also feel able to separate credible from misleading information.27 The truth – which has been demonstrated in many research studies – is that doctors are not able to separate correct from misleading information.26,33,35 How could they when they are only presented with misleading information?35

  Physicians believe that their actions are motivated by how good the drugs are, but studies have shown that their beliefs more closely match marketing claims. A survey of 85 physicians, of which one-third were specialists in internal medicine, showed that 71% believed that impaired cerebral blood flow was a major cause of senile dementia, and one-third had found cerebral vasodilators useful in managing confused geriatric patients.26 However, dementia isn’t caused by impaired blood flow and the drugs didn’t work! Half of these doctors also believed that a morphine derivative, propoxyphene, is more effective than aspirin, although it’s worse and hardly better than a placebo.

  I doubt that these same doctors would privately buy a washing machine that costs 10 times more than other machines, just because the maker has compared it with the cheaper machines and claims that his machine is best. But healthcare is different. Doctors are not held financially accountable for their choices and often prescribe drugs that a
re 10 times or more expensive than older drugs, although the only information they have comes from the manufacturer.

  Because marketing is so effective, industry spends vast amounts of money on it. Already 20 years ago, the industry spent $8000–$15 000 per physician every year in the United States.36 The current expenditure exceeds $1 billion a year in the United States; there is one salesperson per five office-based physicians, and 12% of a random sample of doctors had received financial incentives to participate in studies. You and I pay for all this through our taxes. We not only pay for the extravagant marketing but also for reimbursement of the drugs because they are so expensive that people cannot afford them.

  Meeting with drug salespeople leads to formulary addition requests for the company’s drugs, although most of the requested drugs present little or no therapeutic advantage over existing drugs; it leads to higher drug costs and decreased prescribing of generic drugs; and it leads to irrational prescribing in other ways.33 A study showed that physicians were more likely than other physicians to request these drugs to be added if they had met with salespeople from the companies (odds ratio 13) or had accepted money from them (odds ratio 19).37

  Sponsored meals lead to formulary addition requests, even when the information spread at lunch rounds about the sponsor’s and competitor’s drugs is inaccurate.33 In a study where the salespeople knew their pitches were tape-recorded, 11% of the statements about the sponsor’s drug were inaccurate and favoured the drug, whereas none of the statements about competitors’ drugs were favourable.34 There is reason to suspect it’s much worse when it’s a one-to-one interaction without witnesses.

  Whenever research studies have examined dose–response relationships, they existed.33 Thus, the more exposure to industry people, the worse for the patients and our national economies.

  So-called educational events are not any better. The sponsor’s drug is always preferentially highlighted and prescribing practice changes in favour of the sponsor’s drug.33

  One of the best things a company can do is to invite doctors to a lavish resort. On an all-expenses-paid trip to the Caribbean, the doctors learned about a new intravenous antibiotic and a new intravenous cardiovascular drug.38 Only one of 20 doctors admitted that such a trip could possibly influence prescribing decisions; the other 19 denied it. However, usage at the hospitals more than trebled for the first drug and more than doubled for the second, whereas little happened in national usage patterns for the two drugs. Interestingly, the new drugs did not replace the old ones; they simply increased overall drug usage just as we have seen for other areas, e.g. NSAIDs (see Chapter 14) and SSRIs (Chapter 17). For some reason, the names of the drugs were not revealed, but hospital owners and taxpayers pay for trips like this many times over what they cost for the companies.

  It is not surprising that a major source of income for the drug industry is their me-too drugs. They are rarely any better than old drugs,2,39 but we are usually left in the dark about this, as the industry generally avoids to perform head-to-head comparisons of similar drugs, and as those that are carried out are often rigged.2-15,40,41 Publicly funded trials that compare a new drug with an old one commonly reveal that we have wasted vast amounts of money on drugs that were not any better than cheaper alternatives.40,41,42,43,44,45

  A study from British Columbia showed that even with a generous definition of what constitutes a therapeutic advance, 80% of the increase in drug expenditure between 1996 and 2003 was explained by new, patented drugs that didn’t offer a substantial improvement.39 If only half of the me-too drugs had been priced to compete with the older alternatives, the state could have saved a quarter of its total expenditure on prescription drugs.

  Doctors say they don’t take drug ads in medical journals seriously, but they are influenced by them, otherwise they wouldn’t be there. A 2003 paper reported on 287 advertisements for anti-hypertensive or lipid-lowering drugs and found 125 promotional claims with references.46 However, 23 references were unretrievable, as they referred to data on file or inaccessible monographs, and 45 of 102 referenced claims were not supported by the reference provided, which was therefore pure window-dressing to make the ad look ‘scientific’.

  An analysis of 109 full-page advertisements in 10 leading medical journals showed that in half of cases, they would lead to improper prescribing if the physician had no other information about the drug.47

  An industry insider who responded to a paper in the BMJ where we had compared Cochrane reviews with industry-supported meta-analyses of the same drugs in the same disease48 gave an amusing account of the issue of the unretrievable references under the heading, ‘Pharmaceutical lies’:49

  We have doctors from all over the world who drop into Australia on a first class all expenses paid junket/trip telling us how great a particular medicine is. If you read the small print on a pharmaceutical company flier you will find most references are ‘on file’ or have been presented at a midnight session of the Darfur Cardiologists Conference. As a medical director of a pharmaceutical company I learnt how to get articles published in journals with one journal promising publication if we purchased 2000 reprints at $10 each.

  Summing all this up, a systematic review of 58 studies showed that the information from the drug industry leads to higher prescribing frequency, higher costs and lower prescribing quality.50 We should ask our politicians to forbid marketing of drugs, as it is harmful,33,34,35,36,37,38,51,52 just like marketing of tobacco is, which is why we have prohibited tobacco advertisements.

  Drug companies use the rhetoric of liberalism to defend their right to advertise but liberalism is about the right of people to do what they like as long as it doesn’t harm others, not about companies having a right to do immense harm to people and society with impunity.11

  Actually, most physicians agree that drug salespeople as speakers should be banned,23 but they are highly inconsistent as most of them meet with salespeople every week.33 And it gets worse all the time. In 2004, there were 237 000 meetings and talks in the United States sponsored by drug companies featuring doctor salespeople as speakers and 134 000 led by company salespeople; just 6 years earlier, doctors and industry salespeople delivered together only 60 000 talks.53

  There are also the planted messages. The industry has armies of paid bloggers that distribute pharma material disguised as opinion on the internet, and most major media outlets have pharma ties. For example, James Murdoch, son of Rupert Murdoch, was on the board of GlaxoSmithKline and Time Inc’s CEO Laura Lang formerly worked at Pfizer and Bristol-Myers Squibb. This helps explain why we so often see completely uncritical articles in the media that are copy-and-paste versions of company press releases about their wonder drugs. Like the drug industry, the media are immensely powerful, and when the two join forces, falsehoods are at their worst. The industry also tries to get access to making changes in Wikipedia to ensure pharma friendly messages appear there too.

  Hard sell ad nauseam

  Drugs against nausea and vomiting tell a story about how the voluntary efforts of 100 000 patients were wasted because of poor research conduct. Ondansetron is a showcase for this. When 108 trial reports were examined more closely, it turned out that 14 of them were not new trials but reports that included some of the same patients reported on before.54 None of these additional reports had a clear cross-reference to the original reports, although this is required, and some had a completely new set of authors. Some had combined data from two trials, added a new treatment arm, added more data, used a different anaesthetic, used other numbers of patients or reported other patient characteristics than in the original report. One would have thought it impossible to have a new treatment arm and to use a different anaesthetic in the same trial as reported elsewhere.

  The trials published more than once were the most positive ones. The NNT to prevent vomiting compared with placebo was 16 for the trials that were not duplicated and only 3 for the duplicated ones. The manipulations, which give the readers a false impression
of the drug, were generally not detected, as papers and a textbook cited the same very favourable trial more than once, as if it were separate trials.

  Ondansetron was originally marketed by GlaxoWellcome for nausea and vomiting after chemotherapy, but the company wanted to sell it also for postoperative problems. In 1993, an advertisement in the BMJ talked about ‘Making history of postoperative nausea and vomiting’, but all five references were to studies in cancer.55 In 1994, 18 placebo-controlled trials of ondansetron for postoperative problems had been published, compared to only four trials with an active comparator. Considering that several effective medicines were already available, this wealth of placebo-controlled trials was neither ethical nor helpful for the patients and their doctors, but it was certainly helpful for Glaxo’s marketing machine: Although ondansetron was very expensive, it was highly used instead of the much cheaper alternatives.

  When ondansetron ran out of patent, its effectiveness evaporated overnight it seemed, as there were now other patented ‘setrons’ that were much more expensive. One was granisetron. Its effect on the prevention of postoperative nausea was assessed in the largest Cochrane review ever performed.56 It runs over 785 pages and includes 737 trials (103 237 patients) comparing a drug with placebo or another drug, or doses or timing of administration. This is a colossal waste of resources and abuse for a commercial purpose of the patients’ trust in medical research. Much fewer trials and patients would have sufficed to tell us what we need to know. However, these trials inadvertently show us something about fraud and other manipulations with the data. The nausea trials do not show a symmetric pattern as in Figure 9.1, and the bias in trials comparing granisetron with placebo is huge (see Figure 9.2). The most dramatic effects were seen in small trials, and it is clear that many small trials with poor effect, or showing that placebo was better than the drug, are missing. The bias was similarly large in trials that had compared granisetron with an old, cheap drug, droperidol. Trials that had been performed by a particularly prolific author, Yoshitaka Fujii, were also heavily biased; he was later found to have fabricated his data in 172 studies of which 126 were randomised trials.57,58 This is a world record.