Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients (45 page)

How do we deal with this problem? In the most extreme view, anyone who has a conflict of interest in a particular area should be barred from expressing any view on it. Radio DJs, after all, were supposedly forbidden from accepting ‘payola’ from record companies, and their world didn’t collapse (though I’m sure there are other jollies for radio DJs).

A straight ban raises interesting problems, however. Firstly, in some areas of medicine you might struggle to find any experts at all who’ve never done any work with industry. Here we should pause for a moment to remind ourselves what we really think about the drug industry, and the people who work in it. Although this book is about problems, my goal is that pharma should be adequately regulated and transparent, to the extent that academics can feel positive and enthusiastic about collaboration with it. There is no medicine without medicines; companies can produce great products; and working with people who are focused on completing a project for profit, however distasteful you might find some aspects of that world, can be very exciting.

It’s also odd to take our frustrations out on individual doctors and academics, when they’re simply doing what governments have told them to do over the past three decades: get out there and work with industry. From the 1980 Bayh-Dole Act in the USA, which helped academics register patents on their ideas, to the Thatcher drive for ‘university entrepreneurs’, academics have consistently been told they must engage with industry, and find commercial applications for their output. Discounting all of these academics, having pushed them to engage with industry, and successfully convinced some of our finest minds to do so, would be bizarre.

There are other problems with a straight ban. Even if you can find experts with no conflicts, sometimes the people you most want to hear from are those from industry: they might have special inside knowledge of the processes that have shaped new medicines, for example. And once you start to go down the path of listening to their commercial insights, you run up against a new problem. Sometimes, though this is a
very
tricky area, it might be useful to allow industry people with huge conflicts of interest to speak discreetly, without attribution, on something like a medicines regulatory committee.

Journalists know that deep background, off the record, from a source inside a story they are trying to understand can be extremely valuable. Sometimes an industry person will speak more candidly, but unattributably, to a drug approval committee that doesn’t publish its minutes. I was told one story of an honorary professor of medicine, now working full time in drug development, standing up on an approvals committee to say, ‘Honestly? Everyone knows that drug is rubbish, it won’t last two years, and you’d save me some fuss if you killed it now.’ I’m not telling you this to persuade you that we should permit secrecy in regulation: I don’t think we should. I’m telling you so you can know that you’ve thought it through fairly.

On occasion, some journals have taken the view that industry is simply not to be trusted, even with declarations, and have then made rules accordingly.
JAMA
, for example, decided a few years ago that it would no longer accept industry-funded studies unless they had an independent statistician analysing the results rather than an industry one. It’s an interesting stipulation – it implies that the analysis is where the black magic happens – and it caused an interesting fuss. Stephen Evans is an eminent statistician who works in the same building as me: he is upright, an expert on fraud detection, and a movingly compassionate Christian (truly) in the way he talks about the dishonest academics he has exposed. He argues that we cannot simply discount the work of individual professionals out of hand, on the grounds that there is an observed association between working for industry and producing biased results:

    Suppose that a biomedical journal invoked a new policy requiring that all authors based in western Europe or North America would receive ordinary peer review, but authors from other countries would receive a peer review with additional hurdles. This policy may seem unfair, but suppose the journal claimed that research has shown that there is a greater prevalence of fraud, bias, and sloppy work among papers coming from these other countries.
¹¹³

I think he’s probably right, and that we should judge each paper on its merits, although I do slightly wish he was wrong. It’s also interesting to note that after
JAMA
brought in its ‘independent statistician’ rule, the number of industry-funded trials published in its pages dropped significantly.
¹¹⁴

In general, the most common approach to conflict of interest is that it should be declared, rather than outlawed, and there are two reasons for pursuing this policy. Firstly, we hope it will allow the reader to decide whether someone is biased; and secondly, it is hoped that it might change behaviour. When I suggest that doctors should be forced to tell their patients, with prominent notices in their waiting rooms and on their desks, exactly what companies they’ve accepted money or services from, and exactly what drugs those companies manufacture, it’s partly because I think it might elicit a small amount of shame. Sunlight is a very powerful disinfectant, and has been proven to be so in many different areas. In Los Angeles, the simple act of posting every restaurant’s kitchen hygiene score in its window improved standards, and statistics on car safety led consumers to demand safer cars.

With medicine, however, declaration is more complicated than a single hygiene rating, or safety score, because it’s not always clear what should be declared. Conflict of interest, after all, goes well beyond simple drug industry payments to individual doctors. In the US – this will sound strange to readers in the UK – oncologists get more money if they treat their patients with intravenous drugs rather than simple pills: over half a community oncologist’s income comes from giving chemotherapy, so there is room for a conflict of interest. Similar problems may arise in the UK, with GPs managing the budget for their area, and taking a profit from some of the services they provide. And similar problems can arise when people are writing about a treatment they provide, even where there is no corporate involvement, simply through a vague sense of professional allegiance.

One study, for example, looked at whether academic papers said radiotherapy was a good idea for patients who’d had a particular kind of tumour removed, but where the stage of the cancer was not known: twenty-one out of twenty-nine radio-therapists thought it should be given, compared with five out of thirty-four clinicians from other specialities.
¹¹⁵
A similar bias has been shown for surgeons and coronary bypass operations, surgeons and surgery for a bleeding ulcer, and so on. There has been startlingly bad behaviour from advocates of breast cancer screening, who have overstated the benefits and underplayed the harms (such as the medical risks from unnecessary procedures in wrongly diagnosed women) simply because they were passionately wedded to the procedure.

Conspiracy theorists – who are naturally attracted to the problems in medicine – go further, and build vast castles in the clouds, with huge interlocking tales of conflicts of interest. For them, someone is biased for all time, on every topic, because she has a sister who works for the government; or because somewhere in the university where she works, a person she may never have met has a view on a topic that the industry might find favourable. The conspiracy theorists will then announce that these are secrets which have been
deliberately
withheld, when in reality nobody could possibly have anticipated such elaborate and tenuous fantasies.

So for the most part, if only because it’s practical, academics and doctors tend to concentrate on getting declarations of major financial interests, often just within the past three years, and to leave these more exotic and intangible elements alone. Some do go further. The
BMJ
staff often declare their membership of political parties and other organisations – which is great, but when you step away from money, you drift into territory that starts to feel like an intrusion into someone’s personal life; more than that, as things become more tangential, the decisions about what to declare become more arbitrary, and so perhaps even more misleading, in the selection of what is declared and what is not. As younger people worry less and less about their Facebook security settings, perhaps the future will bring radical transparency for everyone.

But we have other fish to fry. Do people take a declared conflict of interest into account when they read someone’s claims? The evidence suggests that they do. In a trial from 2002, three hundred readers were randomly selected from an academic journal’s database and divided into two groups.
¹¹⁶
Both groups were sent a copy of a short report which described how the pain from herpes zoster, or shingles, could have a substantial impact on patients’ daily functioning; but each group got a slightly different version. The readers in group 1 saw a paper with different named authors from the actual ones, and with a declaration of competing interests, stating that they were employees of a fictitious company treating the condition, and potentially held stock options in it. Readers in group 2 were sent the same paper, but instead of the information about its authors’ employment and stock options, it had a statement that the authors had
no
competing interests. The people in each group were then asked to rate the study, on scales of one to five, for interest, importance, relevance, validity and believability. Fifty-nine per cent of the questionnaires came back (which is remarkably high), and the results were clear: people who were told the authors had competing interests thought the study was significantly less interesting, less important, less relevant, less valid and less believable.

So it is clear that people care about conflicts of interest. And for that reason, specific financial relationships with drug companies are usually declared on academic papers. This system seems to operate reasonably well, but even when conflicts are clearly declared, it may only be on the academic paper, and not in the subsequent work derived from it, such as guidelines or review papers. One study from 2011 took a representative sample of meta-analyses – systematic summaries of all the trials in a field – and looked to see if they described the conflicts of interest of the individual trials they summarised. Of twenty-nine meta-analyses reviewed, only two reported the funders of the trials they included.
¹¹⁷
This is clear evidence that the problem is not paid forward, and that meta-analyses – widely-read and influential documents – simply gloss over this important issue.

We should be clear that declaring conflicts is not a final fix, and that like any intervention it can have side effects which should at least be considered alongside the headline benefit. For example, some have argued that forced disclosure of conflict of interest leads doctors to engage in ‘strategic exaggeration’,
¹¹⁸
knowing that their utterances will be discounted if it is believed that they are acting as shills, and there is some evidence for this in the behavioural economics literature, although only from psychology experiments conducted under laboratory conditions.
¹¹⁹
They may also be affected by a sense of ‘moral licensing’: once you’ve declared your interest, you feel free to let rip with biased advice, because you know the recipient has been warned. These are interesting ideas: overall, I would rather have disclosure.

But these are details. I have a strong suspicion that when you see the scale of this problem you might be slightly amazed. A recent survey in the US looked at senior doctors. Sixty per cent of department heads were receiving money from industry to act as consultants, speakers, members of advisory boards, directors and so on.
¹²⁰

ProPublica, the US non-profit investigative journalism foundation, has done an astonishing piece of work with its Dollars for Docs campaign, creating a huge, publicly accessible database of payments made to doctors.
¹²¹
Individual drug companies have been forced to post this pooled information on their websites, mostly after losing various legal cases. ProPublica has now aggregated data on over $750 million in payments from AstraZeneca, Pfizer, GSK, Merck and many more. The latest slice of data includes details of dinners: so I can tell you that a Dr Emert in West Hollywood ate $3,065 worth of food paid for by Pfizer in 2010, to take just one random example.
¹²²
But while for me this is a curiosity, for patients and others in the US this database has produced a remarkable series of insights, showing the power of putting a lot of information together in one place, where it can be searched and indexed. An individual can look up their own doctor, and see how much they pocketed, to the horror and anger of medics across the country. And anyone can look up whole groups of doctors, to see what horrors lie beneath: overall, 17,700 doctors received money, and 384 got more than $100,000.

What’s more, universities around the country seemed to have little idea what was happening on their own premises, until the data was presented to them clearly. When the University of Colorado, Denver, saw that over a dozen of its senior academics were giving paid promotional talks for pharma, it launched a complete overhaul of its conflict-of-interest policies.
¹²³
The Vice Chancellor was unambiguous: ‘We’re going to just have to say we’re not going to be involved with these [CME] speakers’ bureaus, because they’re primarily marketing.’ In some places, university policy was being routinely ignored. Five faculty members at Stanford were shown to be taking money to give industry-sponsored lectures, and had disciplinary cases launched against them.
¹²⁴