‘If I only pulled half the people out of the way of an oncoming vehicle, when I could have pulled all of them out of the way, then you would rightly say that I was a bit of a bastard,” says Ben Goldacre. But that, he says, is the problem with modern medicine. That articulated truck bearing down on all of us is the avoidable harm of pharmaceutical drugs. And it is simply not good enough, he argues, that the drug industry is so relaxed about the massive gaps that exist in our knowledge about their effectiveness, safety and potential side effects.
Goldacre, a British doctor, epidemiologist and geek provocateur, attracted plenty of attention with his previous best-seller, Bad Science, a scathing critique of the lack of evidence-based science in the media. But his new book, Bad Pharma, seems to have struck an even bigger nerve by pointing out that, in many cases, humans are being used as guinea pigs by the drug industry. To the public, it may seem that drugs are exhaustively tested before they go to market. The truth, says Goldacre, is the main judge of their effectiveness is often the manufacturers themselves. And often they are relying on poorly designed trials; on small numbers of unrepresentative patients; and on flawed analysis techniques that tend to produce biased results.
He is not against drugs or even the companies that produce them. “Drug companies around the world have produced some of the most amazing innovations of the past 50 years, saving lives on an epic scale,” he writes. “But that does not allow them to hide data, mislead doctors and harm patients.” Think he’s exaggerating? Consider Tamiflu, the brand name of oseltamivir. This drug from Roche, kept by governments around the world in billion-dollar stockpiles, was the prime weapon during the 2009 H1N1 swine flu epidemic. New Zealanders who wanted it had to pay up to $80 a dose. Yet according to Cochrane Collaboration, the world’s gold-standard healthcare number-crunchers, about 60% of Roche’s data from phase 3 trials of the drug have never been published, so its effectiveness for treating pneumonia and reducing deaths is still unknown. “When a company can withhold information from doctors and patients like that, that’s clearly a very broken system,” says Goldacre.
And that’s just the tip of the drug information deficit, he claims. In 2004, a popular anti-inflammatory drug known as Vioxx (rofecoxib) was withdrawn worldwide by Merck Sharp & Dohme after years in the market and having made billions of dollars for the company. A study had found using it for longer than 18 months could measurably increase the risk of heart attack and stroke. Medical journal the Lancet said unacceptable risks were apparent years earlier. At the time, about 15,000 New Zealanders were using Vioxx for pain and inflammation. Subsequent studies, including a Swiss study that combined 31 randomised, controlled trials in various countries involving more than 100,000 patients, have since confirmed the high level of risk of many popular painkillers, known as non-steroidal anti-inflammatory drugs (NSAIDs), when taken long-term. Brands still sold in New Zealand include Naprosyn (naproxen), Nurofen (ibuprofen) and Voltaren (diclofenac). And in 2007, another anti-inflammatory, Prexige, was withdrawn from sale following reports of severe liver damage in people taking more than 200mg a day. At the time, more than 1000 New Zealanders were reported to be taking twice that dose.
In July, GlaxoSmithKline was hit with US$3 billion in fines for misbranding the anti-depressants Paxil and Wellbutrin and marketing them for unapproved uses. The company was also fined for holding back data on Avandia, a diabetes drug. And this month, Myriad Genetics, which makes a gene test for breast and ovarian cancers, was accused of deliberately withholding data that could help other scientists understand cancer genetics.
ECOSYSTEM OF FAILURE
Goldacre’s view, backed by examples and footnotes throughout his 400-page book, is not that drug companies are “evil” – he acknowledges they employ many smart, well-intentioned people – but that patients and GPs are let down because of “a huge ecosystem of failures and problems that reinforce each other”. The crux of the problem, he says, is the pharmaceutical industry appears all too willing to gloss over unfavourable outcomes. Trials with positive results are about twice as likely to be published as those with negative findings, and companies can choose not to reveal data that cast doubt on a drug’s effectiveness. Something like a third of all new drugs are tested against a placebo – a substance containing no medicine – instead of a currently available treatment, so their overall value is unclear. And most astonishingly, the results of about half of all drug trials are never published.
The result, says Goldacre, is that for many important diseases we have treatments that are better than nothing, but we don’t have enough information to know which of them works best. “We flatter ourselves that it’s okay to not give people the best treatment, because we do at least give people some kind of treatment. [But] if you don’t give the best treatment, you’re exposing people to avoidable harm.” And sometimes you may be paying for an expensive treatment when a cheaper one may be just as good.
Just how New Zealand fits into this big picture is a controversial subject. The Government’s drug-buying agency, Pharmac, is renowned for its no-nonsense attitude to funding new pharmaceuticals with big claims but less than compelling evidence. In fact, its parsimony has caused so much grumbling from the American drug industry that the future of Pharmac has become a key issue in trade talks between New Zealand and the US, as part of a proposal for a potentially powerful free-trade alliance known as the Trans-Pacific Partnership (TPP). The next round of negotiations on the TPP is due to take place in Auckland next month. In September, a US lobbyist for the drug industry visited New Zealand and pointed out that our pharmaceutical spending as a share of total health spending was among the lowest in the OECD. Yet with the exception of 100 or so local trials, all our drug-testing information comes from elsewhere. We are also one of only three developed countries that allow direct-to-consumer drug advertising and there are few restrictions on drug company representatives schmoozing GPs.
In 2007, Les Toop, a professor in the Department of Public Health and General Practice at the University of Otago, and his colleague, associate professor Dee Mangin, wrote a critical report on direct-to-consumer drug ads for the British Medical Journal. They found that most health professional, consumer and patient groups are opposed to them. A poll of 1600 GPs showed they were overwhelmingly against consumer drug ads, believing “it harmed the doctor-patient relationship and public health”. According to Goldacre, drug advertising encourages doctors to make bad prescribing decisions. “When there’s a big advertising campaign, when patients ask their doctor for antidepressants for depression, they’re much more likely to get them.” He points out that consumer ads have been banned in most countries since the 1940s for the simple reason that they work. They are prohibited in Australia and across Europe, but have been permitted in the US and New Zealand since the 1980s, along with Pakistan and South Korea. Mangin says they work so well the return on ad expenditure is something like three to one. That means for every $1000 spent on drug ads, the drug company earns $3000. In 2001, Toop notes, New Zealand came “within a whisker” of having consumer drug ads banned, but the then Labour Government decided to tighten regulation instead. The Health Minister at the time, Annette King, acknowledged debate over the issue had been “extremely vigorous”.
Pharmaceuticals lobby group Medicines New Zealand argues advertising is simply one of the methods of helping consumers inform themselves about what is available. Consumer ads are used “as a means of enhancing patient engagement”, it says. It also claims the advertising of prescription medicines has provoked “very few complaints” to the Advertising Standards Authority and is “well regulated” by the Medicines New Zealand code of practice. The Ministry of Health unit that regulates medicines and devices sold in New Zealand, Medsafe, won’t say whether consumer drug advertising should be banned. Toop and Mangin’s BMJ report, however, says reviews show direct-to-consumer ads have poor compliance with guidelines and are not independently assessed for balance or the scientific validity of the claims unless someone complains. It also notes that Vioxx and Celebrex were “heavily and effectively” promoted in New Zealand “despite the awareness of their cardiac risks”. It is worth noting that these two drugs are among those not funded by Pharmac.
Medsafe, which is funded by fees from drug applications and licensing, says regulators require that all data – both positive and negative – be provided to them. “Steps have been taken in New Zealand and globally to improve transparency, including clinical trial registries, and the requirement that data from all studies is published,” says group manager Dr Stewart Jessamine. “Work is ongoing internationally in these areas.”
According to Medsafe, all phase 2 and 3 studies (phase 3 drug trials are final testing; phase 4 are those that may be done after approval is granted) are registered in publicly accessible clinical trials websites, such as international site clinicaltrials.gov, and the New Zealand and Australia clinical trials register. It points out that regulators and other journals also publish material on medicines’ safety, such as minutes of expert advisory meetings and information on noncompany-funded studies. Kevin Sheehy, Medicines New Zealand’s general manager, says sponsors of clinical research now routinely register each trial. He also notes drug companies are responding to calls for greater transparency by creating clinical trials databases, such as roche-trials. com, that can be accessed by anyone. This is not a trivial endeavour: Medsafe says a typical new drug’s clinical data will run to 300,000-500,000 pages. Companies have agreed to present the results within a year of a trial being completed, and in Roche’s case, it offers more detailed reports to investigators working on its trials.
This reassuring response matches how the industry in the UK responded to Goldacre’s book: it insisted all the problems had been fixed or were in the process of being fixed. But according to Goldacre and an army of science-literate, internet-savvy fact-checkers, this is simply not true. Despite some efforts by governments and regulators, says Goldacre, nothing much has changed. The European Medicines Agency doesn’t exercise its powers to make companies reveal data. The clinicaltrials.gov website cited by Medsafe hasn’t worked, he says. “The FDA passed a regulation saying everybody has to post the results of their trials on clinicaltrials.gov within a year of completion, and we know that only one in five trials has met that obligation. We didn’t find that out for four years because nobody at the FDA bothered to do a public audit.”
Pharmac’s medical director, Dr Peter Moodie, says the agency considers both published and unpublished evidence or analysis, some of which is provided by companies when they make funding applications. Companies must explain why studies weren’t published, he says, but he confirms the agency has no statutory right to see company data on the safety or effectiveness of drugs. Drug companies argue such information can be commercially sensitive, and would reduce their incentive to spend billions of dollars on risky research. If they fail to provide complete drug trial data, Medsafe can legally sanction them by demanding that drugs be removed from the market. But according to Goldacre, regulators are often reluctant to withdraw drugs, “in case that is seen as an admission of their failure to spot problems in the first place”.
Toop agrees that GPs’ and other health professionals’ knowledge of the flaws in drug trial data is “patchy”. And according to Mangin, patients are even less likely to be aware of the flaws in clinical drug trials, or that the data from trials they participate in may never be made available. She describes this practice as “unethical”. “Patients bear all the risks and companies make all the money.” Patient advocates are also concerned that consumers don’t always get the information they need to make informed decisions. Women’s Health Action director Julie Radford- Poupard says her organisation’s main concern is that new and existing medicines are tested and safe, and she believes drug companies should be forced to register and report trial data. Women’s Health Action has lobbied for better regulation and an even larger role for Pharmac.
HITTING BACK AT PHARMAC
Meanwhile, the drug industry has its own concerns. Sheehy says Medicines NZ supported a recent health select committee recommendation that Pharmac should provide guidance to industry about trial design. This recommendation was not followed, he says. Sheehy also finds fault with Pharmac’s funding policies. As Goldacre notes, the agency has a worldwide reputation for “holding the line” on drug claims, and an aggressive attitude to costs. For example, the prices it pays for cholesterol-lowering drugs such as statins are a fraction of those in Australia. Pharmac achieves this through a range of buying strategies, including tenders and “reference pricing”. The latter is where Pharmac pays the same subsidy for medicines that have comparable therapeutic effects. If suppliers set higher prices, patients have to pay the difference above the subsidy – and tend to choose the fully subsidised medicine.
“The New Zealand system of medicine funding doesn’t incentivise innovation, as the uptake of innovative medicines here is too low and slow,” says Sheehy. “The best method of incentivising innovation in medicine development is for the funding system to appropriately value those medicines that provide additional benefits to patients.” He says Pharmac’s funding role also affects patient choice. “This effectively takes a lot of the decision-making about therapeutic alternatives away from the GP or patient.” The industry would also like more access to patient and prescription data. Sheehy says Medicines New Zealand has repeatedly recommended Pharmac and other health agencies consider improving the use of the National Health Identifier number for research, including to monitor the outcomes related to medicine use. “We believe New Zealand is in a very strong position to embark on some world-leading research using this system, which is almost unique in the world.” But regulators and GPs are wary. They are fearful of continued publication bias – positive results being published ahead of negative ones – and say drug safety data should be managed by health authorities, not industry.
As for suggestions that consumers sometimes find drug claims difficult to decipher, Sheehy says Medicines New Zealand requires of its members that research data be presented clearly and with appropriate comparisons. But in practice, claims that a drug is “50% better”, for example, do occur. In relative terms, the “50% better” claim is correct, even though in absolute terms it means three in 100 patients might benefit from a drug rather than two in 100. Statisticians say both numbers should be included. But Sheehy says his organisation cannot specify that all data should be presented as a “relative” versus “absolute” benefit, “as there are instances where each method is appropriate”.
There are signs of progress. In Canterbury, the district health board has helped ensure 80% of the region’s GPs get regular, independent information on new and existing drugs. Information is sifted by a team of 10. GPs are key to the initiative, and are paid to attend regular meetings with their peers and to run evidence-based examinations of their own prescribing data. The programme has been running since 1992 and was initially paid for by the “wastage” from funding newer drugs, says Mangin. “Most drug reps have disappeared from town now.” The scheme is unique in New Zealand, but Toop says there is interest from the Royal NZ College of GPs. All it needs is for GPs to get involved and for health boards to fund it, he says. Mangin believes the Canterbury programme works better than Goldacre’s suggestion of banning drug reps from hospitals and doctors’ offices, as that would simply shift promotion into other areas, such as paid talks and dinners.
Toop also notes New Zealand has the highest per-capita level in the world of “spontaneous reports” by GPs of negative drug reactions, although it is still only about 5% of the total. And there are more than 100 trials ranging from phase 1 to phase 4 under way in New Zealand. According to Sheehy, trials for gout, diabetes, osteoporosis and cancers are world-leading. The number of international organisations devoted to evidence-based science related to drug trials is also growing. Mangin is involved in Rxisk.org, a global scheme for patients to report adverse effects. It can take about 10 years for safety risks to manifest themselves, she says, but she hopes such schemes will shorten the time it takes to discover “the next Vioxx or Avandia”. There is also a local pilot scheme to make more use of GP clinical records, but there are privacy concerns and they require a sophisticated search engine.
This month, the BMJ was cautiously optimistic about GlaxoSmithKline allowing access to anonymised patient-level data from its clinical trials. At the same time, the journal announced that from next year it would publish industry-funded trials “only where there is a commitment to make the relevant anonymised patient-level data available on reasonable request”. The announcement specifically mentioned Goldacre’s book, but other medical journal editors have yet to take this step. Goldacre says the most valid criticism anyone can make of his book is that he’s demanding perfection. “And that’s absolutely bloody right. It’s a life and death matter. If you were choosing a surgeon, you would want the best surgeon. If you were choosing a drug, you should want the best drug. And you would expect that the systems society had put in place for assessing your drug were the best they could be.”
BAD PHARMA: HOW DRUG COMPANIES MISLEAD DOCTORS AND HARM PATIENTS (4th Estate, $34.99).
Hits & myths
Statins halve your risk of having a heart attack, says Ben Goldacre. “But if one of them is only 1% better than one of the others at preventing death, because heart attack is such a common cause of death and because so many people are taking statins around the world, that would be an enormous piece of information that we were missing out on.” The side effects of statins, of which simvastatin, pravastatin and atorvastatin are funded in New Zealand, can include depression, sleep disturbances, memory loss and sexual dysfunction. Pharmac says: “A number of studies compare individual statins with other statins, but nearly always they use surrogate intermediate outcomes – for example, for lowering of cholesterol levels or atheroma volume [fatty artery deposits] – that are not guaranteed to reflect real-world outcomes.”
SSRI anti-depressants – selective serotonin reuptake inhibitors such as Prozac (fluoxetine hydrochloride) and citalopram – work because they increase serotonin levels in your brain, right? Wrong. The hypothesis was always shaky, says Goldacre, and the evidence for SSRIs treating at least mild and medium-level depression is now “hugely contradictory”. What’s more, there’s a drug called tianeptine that actually reduces serotonin levels, and yet it is “a pretty effective treatment for depression”, he says.
FEMALE SEXUAL DYSFUNCTION
As recently as this month, the media and academic press have been reporting imminent breakthrough treatments for female sexual dysfunction. The latest reports have referred to a “female Viagra” that “could soon be available as a nasal spray”. Goldacre is concerned the issue is being seen “as a molecular problem rather than a relationship problem”. And he is sceptical that such drugs will ever materialise.
Herceptin: the right decision?
The jury is still out on the best use of the costly breast cancer drug.
In 2007 Pharmac announced, after a couple of years’ consideration, that it would fund the breast cancer drug Herceptin (trastuzumab) for treatments of nine weeks’ duration. The country’s drug-buying agency, Pharmac, said there was no evidence longer treatment was more effective. Two years before, Medsafe had granted provisional consent for the drug to treat early breast cancer. Roche, the maker, had been pushing for 12-month treatments. Herceptin, often in conjunction with other treatments, had been shown to increase survival rates for those with HER2-positive breast cancers, an often aggressive form that accounts for perhaps a quarter of all breast cancers. But it is expensive – in 2006 it cost $70,000 a patient a year – and can have serious cardiac side effects.
In June 2007, eight women filed proceedings in the High Court seeking a judicial review of Pharmac’s decision, and the court eventually directed it to hold public consultations. In July 2008, after consultation with breast cancer patients and oncologists, Pharmac confirmed its decision to fund nine-week treatments. The National Party promised to overrule Pharmac and fund Herceptin for 12 months if it was elected at the end of 2008. (It had costed the election promise at an extra $9 million a year for three years, and several million dollars more each year turned out to be needed for administration.) Around this time, an article in UK medical journal the Lancet said Herceptin used over 12 months was much less effective than first thought because important clinical trial data from 985 women given the drug for 12 months was missing. “It’s a chemotherapy drug for treating cancer,” says Ben Goldacre. “It’s not something that you would enter into trivially, giving somebody 12 months of it, if you could get away with nine weeks and still have the same benefit.”
For every funding decision, there’s an opportunity cost – another treatment doesn’t get funded. And it’s a complex calculation to do the most good and the least harm. Sharron Cole, a member of Pharmac’s consumer advisory committee, wrote in 2008: “Of every 100 women treated with HER2-positive breast cancer, 77 without Herceptin will be alive and disease-free two years later and with Herceptin, 86. Statistically, this means that of the 23 women who would have had a return of their cancer, nine would be disease-free if they had Herceptin – in other words, a 39% risk reduction. But as 77% of the women would not have had a return of their cancer anyway, the absolute risk reduction is 9%. “Because it is not known which nine women will benefit, all 100 women must be treated, at the average cost of $70,000 a woman, if the Roche 12-month treatment is accepted.”
Groups such as Women’s Health Action continue to say that the evidence does not support 12 months’ treatment, and have confidence in Pharmac’s processes. Pharmac’s Peter Moodie says “robust” head-to-head scientific studies are still examining the drug. “The optimal duration and sequencing” – it can be given with or after other treatment – “of trastuzumab therapy in HER2-positive early breast cancer patients remains an open question.”