by Ben Goldacre
It’s also interesting to note that the increasing commercial-isation of research has driven many everyday clinicians away from trials, even when the trials come from the more independent end of the spectrum. Three British academics have written recently of their difficulty in getting doctors to help them recruit patients for a study requested by the European medicines regulator, but paid for by Pfizer: ‘Academics wrote the protocols, collaborators are academic, and the study data are owned by the steering committees (on which industry has no say), which also control analyses and publications. A university is the sponsor. Funding is from industry, which has no role in study conduct, data collection, or data interpretation.’10 UK doctors and primary care trusts regarded this study as commercial, and were reluctant to hand over their patients. They are not alone. The Danish Board of Medicine regards these kinds of studies as commercial, which means that any practice participating in them must declare the interest, further reducing recruitment. In the US, meanwhile, the use of private community doctors to conduct trials has expanded enormously, with incentives approaching $1 million a year for the most enterprising medics.11
For a window into the commercial reality of the CRO world, you can look at the way these services are presented when they are being promoted to pharmaceutical companies, and see how far this reality is from the needs of patients, and any spirit of neutral enquiry. Quintiles, the largest company, offers to help its industry customers to ‘better identify, promote, and prove the value of a particular drug to key stakeholders’.12 ‘You’ve spent hundreds of millions of dollars and years bringing your product through the drug development process,’ it says. ‘Now you face multiple opportunities – and possibly more requirements – to demonstrate safety and effectiveness in larger populations.’ There are also cases of CROs and drug companies with contracts that share the risk of a poor outcome between them, increasing the chances of a conflict of interest even further.
These aren’t smoking guns. They simply illustrate the banal commercial reality of what these companies do: they find stuff out, of course, but their main objective is to make a company’s drug look good so that regulators, doctors and patients will swallow it. That’s not ideal in science. It’s not fraud either. It’s just not ideal.
It would be wrong to imagine that this shift in culture has been driven by a hope that CROs would produce more flattering findings than other options. They are attractive because they’re fast, efficient, focused and cheap. And they’re especially cheap because, like many other industries, they outsource their work to poorer countries. As the former chief executive of GSK explained in a recent interview, running a trial in the US costs $30,000 per patient, while a CRO can do it in Romania for $3,000.13 That is why GSK aims to move half of its trials to low-cost countries, and it is part of a global trend.
In the past, only 15 per cent of clinical trials were conducted outside the USA. Now it’s more than half. The average rate of growth in the number of trials in India is 20 per cent a year, in China 47 per cent, in Argentina 27 per cent, and so on, simply because they are better at attracting CRO business, at lower cost. At the same time, trials in the US are falling by 6 per cent a year (and in the UK by 10 per cent a year).14 As a result of these trends, many trials are now being conducted in developing countries, where regulatory oversight is poorer, as is the normal standard of clinical care. This raises a huge number of questions about data integrity, the relevance of findings to developed world populations, and ethics – all issues with which regulators around the world are currently struggling.
There are many anecdotal accounts of bad behaviour with trial data in poorer countries, and it’s clear that the incentives for massaging results are greater in a country where a clinical trial is paying its subjects vastly more than local wages could match. There are also difficulties with regulatory requirements that fall across two countries or two languages, as well as translation issues in patient reports, especially for unexpected side effects. Site visits for monitoring may be of variable quality, and countries differ in how much corruption is routine in public life. There may also be less local familiarity with administrative requirements concerning data integrity (which have been a bone of contention between industry and regulators in the developed world).15
These are, we should be clear, only hints of problems with data. There have been cases where trials from developing countries have produced positive results, while those conducted elsewhere showed no benefit, but to the best of my knowledge, very little quantitative research has so far been conducted comparing the results of trials in poorer countries with those in the US and Western Europe. This means we can draw no firm conclusions on data integrity; it also means – I would suggest – that this is an open field for highly publishable research by anyone reading this book. A barrier to such research, though, would be access to the most basic information. One review of articles in leading medical journals reporting on multi-centre trials found that less than 5 per cent gave any information on recruitment numbers from each individual country.16
There is also the issue of publication bias, where whole trials disappear. We saw in the previous chapter how unflattering data can go missing in action. European and US researchers with stable academic positions have had difficulties retaining the right to publish, sometimes engaging in heated confrontations with drug companies. It’s hard to imagine that such problems wouldn’t be exacerbated in a developing country setting, where commercial research has introduced unprecedented investment for individuals, institutions and communities. This is especially problematic because the trials registers, where protocols should be posted before a trial begins, are often poorly policed around the world; and trials from the developing world – or simply data from new sites – may only come to the attention of the international community after they are completed.
But there is a more interesting problem with running trials in such diverse populations: people, and medicine, are not the same the world over. It is known – we shall see more later – that trials are generally conducted in very unrepresentative ‘ideal’ patients, who are often less sick than real-world patients, and on fewer other drugs. In trials conducted in developing countries, these problems can be exacerbated. A typical patient in Berlin or Seattle with high blood pressure may have been on several drugs for several years. Now, you might collect data on the benefits of a new blood-pressure drug in Romania, or India, where the patients may not have been on any other medication, because access to what would be regarded as normal medical treatment in the West is much less common. Are those findings really transferable, and relevant, to American patients, on all their tablets?
Beyond differences in routine treatment, there will also be a different social context. Are patients diagnosed with depression in China really the same as patients diagnosed with depression in California? And then there are genetic differences. You might know, from drinking with friends, that many Oriental people metabolise drugs, especially alcohol, differently from Westerners: if a drug has few side effects at a particular dose in Botswana, can you really rely on that data for your patients in Tokyo?17
There are other cultural considerations. Trials are not simply a one-way street: they are also a way to create new markets, in countries like Brazil, say, by reshaping norms of clinical practice, and modifying patients’ expectations. Sometimes this may be a good thing, but trials can also create expectations for drugs that cannot be afforded. And they can even, by distorting local employment markets, draw good physicians away from clinical work in their own communities, and off into research jobs (just as Europe has taken expensively-trained doctors and nurses away from developing countries by emigration).
But more than anything, these trials raise enormous issues about the ethics of research, and meaningful informed consent.18 The incentives offered to participants in developing countries can exceed the average annual wage. Some countries have a culture of ‘doctor knows best’, where patients will be more likely to accept unusual
or experimental treatments simply because their doctor has offered them (a doctor with a significant personal financial interest, we should note, since they are paid for each recruited patient). The background and risks – that a drug is new, that they may actually be taking a dummy placebo pill – may not be clearly communicated to patients. Informed consent may not be adequately policed. Standards of ethics oversight can also vary: in a survey of researchers in developing countries, half said their research was not reviewed at all by an institutional review board.19 A review of Chinese trial publications found that only 11 per cent mentioned ethics approval, and only 18 per cent discussed informed consent.20 This is a very different ethical context for research to that in Europe or the US; and while international regulators have tried to keep up, it’s not yet clear whether the steps they’ve taken will be successful.21 What’s more, oversight is especially problematic, since these trials are often used to bolster the marketing case for a drug after it has come to market, and are not included in the bundle of documents presented for regulatory approval, meaning they are less subject to Western regulatory control.
Outsourced CRO trials in the developing world also present the issue of fairness that we met earlier in our discussion of phase 1 trials: the people participating in trials are supposed to come from a population that could reasonably expect to benefit from their results. In several damning cases, especially from Africa, it is very clear that this was not the case. In some cases, more horrifically, it seems that effective treatment which was available may have been withheld as part of the drug company’s effort to conduct a trial.
The most notorious story is the Trovan antibiotic study conducted by Pfizer in Kano, Nigeria, during a meningitis epidemic. An experimental new antibiotic was compared, in a randomised trial, with a low dose of a competing antibiotic that was known to be effective. Eleven children died, roughly the same number from each group. Crucially, the participants were apparently not informed about the experimental nature of the treatments, and moreover, they were not informed that a treatment known to be effective was available, immediately, from Médecins sans Frontières next door at the very same facility.
Pfizer argued in court – successfully – that there was no international norm requiring it to get informed consent for a trial involving experimental drugs in Africa, so the cases relating to the trial should be heard in Nigeria only. That’s a chilling thing to hear a company claim about experimental drug trials, and it was knocked back in 2006 when the Nigerian Ministry of Health released its report on the trial. This stated that Pfizer had violated Nigerian law, the UN Convention on the Rights of the Child and the Declaration of Helsinki.
This all took place in 1996, and was the inspiration for John le Carré’s novel The Constant Gardener. You may think 1996 was a long time ago, but the facts in these matters are always on a delay, and in contentious or litigated issues the truth can move very slowly. In fact, Pfizer only settled the case out of court in 2009, and several disturbing new elements of what is clearly an ongoing saga emerged in the WikiLeaks diplomatic cables made public in 2010.22 One US diplomatic cable describes a meeting in April 2009 between Pfizer’s country manager and US officials at the American embassy in Abuja, where smears of a Nigerian official involved in the litigation are casually discussed.
According to [Pfizer’s country manager], Pfizer had hired investigators to uncover corruption links to Federal Attorney General Michael Aondoakaa to expose him and put pressure on him to drop the federal cases. He said Pfizer’s investigators were passing this information to local media. A series of damaging articles detailing Aondoakaa’s ‘alleged’ corruption ties were published in February and March. Liggeri contended that Pfizer had much more damaging information on Aondoakaa and that Aondoakaa’s cronies were pressuring him to drop the suit for fear of further negative articles.23
Pfizer deny any wrongdoing in the Trovan trials, and say the statements contained in the cable are false.24 Its $75 million settlement was subject to a confidentiality clause.
These issues are disturbing in themselves, but they must also be seen against the wider context of trials in the developing world on drugs that are not available for normal clinical use in those countries. It’s a classic ethics essay quandary, but very much set in real life: imagine you’re in a country where modern AIDS medicines cannot be afforded. Is it reasonable to run a trial of an expensive new AIDS drug in that setting? Even if it has been broadly demonstrated to be safe? What if the control group in your trial are simply receiving dummy placebo pills, which is to say, effectively, nothing? In the USA, no patient would receive dummy sugar pills for AIDS. In this African country, perhaps ‘nothing’ is a commonplace treatment.
This is an area of considerable obfuscation and embarrassment, tangled up in complex regulatory frameworks which are starting to change in a worrying direction. In 2009, three researchers wrote in the Lancet drawing attention to one very notable shift.25 For years, they explained, the FDA had insisted that when a company applied for a marketing authorisation for a drug in the US, all foreign trials given as evidence had to show that they were compliant with the Declaration of Helsinki.26 In 2008 this requirement changed, but only for foreign trials, and the FDA shifted to the International Conference on Harmonisation (ICH) Good Clinical Practice (GCP) guidelines. These are not terrible, but they are only voted on by members from the EU, the USA and Japan. They are also more focused on procedures, while Helsinki clearly articulates moral principles. But most concerning are the differences in detail, when you consider that GCP is now the main ethical regulation for trials in the developing world.
Helsinki says that research must benefit the health needs of the populations where it is conducted. GCP does not. Helsinki discusses the moral need for access to treatment after a trial has finished. GCP does not. Helsinki restricts the use of dummy placebo sugar pills in trials, where there are effective treatments available. GCP does not. Helsinki also, incidentally, encourages investigators to disclose funding and sponsors, post the study design publicly, publish negative findings, and report results accurately. GCP does not. So this was not a reassuring regulatory shift, specifically for trials conducted outside the US, and specifically in 2008, at a time when studies were moving outside the US and the EU at a very rapid pace.
It’s also worth mentioning that the pharmaceutical industry plays hardball with developing countries over the price of medicines. Like much else we are discussing, this is worthy of a book in its own right, but here is just one illustrative story. In 2007 Thailand tried to take a stand against the drug company Abbott over its drug Kaletra. There are more than half a million people living with HIV in Thailand (many of them can thank Western sex tourists for that), and 120,000 have AIDS. The country can afford first-line AIDS drugs, but many become ineffective with time, through acquired resistance. Abbott had been charging $2,200 a year for Kaletra in Thailand, which was – by morbid coincidence – roughly the same as the gross income per capita.
We give drug companies exclusive rights to manufacture the treatments they have discovered for a limited period of time – usually about eighteen years – in order to incentivise innovation. It’s unlikely that the revenue available from selling drugs in poorer countries will ever incentivise innovation of new treatments to any great extent (we can see this very clearly from the fact that so many medical conditions that occur mainly in developing countries are neglected by the pharmaceutical industry). Because of this there are various international treaties, such as the Doha Declaration of 2001, under which a government can declare a public-health emergency, and start manufacturing or buying copies of a patented drug. One memorable use of these ‘compulsory licences’ was when the US government insisted in the aftermath of the 9/11 attacks that it should be allowed to buy large amounts of cheap ciprofloxacin to treat anthrax, when it was worried that spores were being sent to politicians by terrorists.
So in January 2007 the Thai government announced that it was going to copy
Abbott’s drug, only for the country’s poor, to save lives. Abbott’s response was interesting: it retaliated by completely withdrawing its new heat-stable version of Kaletra from the Thai market, and six other new drugs for good measure, and then announced that it would not bring these drugs back to the Thai market until the government promised not to use a ‘compulsory licence’ on its drugs again. It’s hard to think of anything less in keeping with the Doha Declaration. If you want more moral context, the World Health Organization estimates that half of HIV transmissions in Thailand come from contact between sex workers and their clients. There are said to be two million women and 800,000 children under eighteen working in the Thai sex trade, much of which is servicing Western men, some of whom you may know personally.
So those are phase 1, 2 and 3 clinical trials: both the science of them and, I hope, some colour about the reality beyond the protocols, in the clinics and on the streets. It may have made you nervous. The story from there on is simple: the medicines regulator, whether it’s the FDA, or the EMA, or some other country’s body, looks at the results from these phase 1, 2 and 3 trials, works out if the drug is effective and the side effects are acceptable, then either asks for some more trials, tells the company to bin the drug, or lets it go on the market for prescription by any willing doctor. That’s the theory, at any rate.
In reality, things are much messier.