by David Healy
In the fifteen years following the FDA ruling on Risperdal and other new antipsychotics, no independent evidence appeared that any of the newer antipsychotics was superior to the older ones in terms of either safety or efficacy—even though the new treatments cost between fifty and eighty times as much.6 But in the interim the companies managed to convert virtually everyone in the medical community from older to newer antipsychotics, and all of the new drugs made it on to hospital formularies—how? More generally, how do the pharmaceutical companies manage to market newer drugs so successfully when the cost of healthcare is forcing everyone to be aware of costs and in the face of guidelines, which ostensibly based on the evidence might be expected to come to the same conclusion as the FDA?
Part of the answer to this conundrum lies in the medical academics who, as we have seen, are among the key people who influence a doctor’s view of particular drugs. Regulators have no control over what these academics say—academics, often, whom pharmaceutical companies have made into opinion leaders. In the case of antipsychotics like Risperdal, statins like Lipitor, or proton pump inhibitors like Nexium, professors of medicine, psychiatry, pharmacology, or general practice can say what they like in lectures or report what they like in medical journals. Companies can even include statements in their ads claiming that, say, Risperdal is superior to haloperidol, provided it is clear the statement has been made by an academic rather than the company. There will be a footnote in the ad to a medical article in which superiority is claimed—almost certainly a ghostwritten article.
An even more effective marketing technique is to coax support from medical academics who are not hired guns, who may even see themselves as hostile to company marketing and keen to constrain this marketing within a framework of independent treatment guidelines. It is in fact by manipulating the most independent of medical academics through guidelines that companies have been able to make new drugs from Risperdal to Lipitor, Vioxx, Nexium, and Fosamax into the most profitable drugs in the world.
CONSENSUS CONFERENCES
In the 1980s it seemed obvious to many medical academics with no links to the pharmaceutical industry that where there was a dispute about a drug or other medical treatment it made sense to bring representatives of the differing points of view together in an attempt to achieve consensus. This led to the creation of consensus conferences aimed at producing guidelines for clinical practice.7 Initially, these consensus conferences seemed like a way to rein in the excesses of pharmaceutical company marketing departments—if we review all the evidence it will surely be clear that the benefits of a new drug are far less than the marketing hype might suggest. With this in mind, groups across medicine began to convene conferences to produce treatment guidelines on new drugs for conditions ranging from arthritis to schizophrenia.
Initially, the organizers of these consensus conferences were in the business of developing guidance for doctors rather than guidelines to be rigidly adhered to. Twenty years later the guidelines we now have still notionally offer guidance to doctors, but this is the kind of help that once led Ronald Reagan to suggest that the scariest words in the English language were “I’m from the Government and I’m here to help you.”
By the time I was invited to a consensus conference in London convened by Catalyst Healthcare Communications on behalf of Janssen in 1995, drug companies far from feeling constrained by guidelines had begun to embrace them. Other invitees to this London meeting included senior psychiatrists, pharmacists, and economists. No one among the invitees would have been thought of as a friend of the pharmaceutical industry. We were presented with the published results of Janssen’s trials of Risperdal. There was no attempt to stifle debate or to block us from bringing in any relevant material we might have been aware of.
The exercise involved taking the published research on Risperdal and discussing what would happen in real life if the results found in the clinical trials, which had all been reported in the better journals, applied. What effect would it likely have on the rate patients got discharged from acute hospital settings or from longer-term care facilities and on their rates of readmission to a hospital bed? When costing the outcomes, a significantly higher cost was used for Risperdal compared to the older drugs. Nevertheless, use of Risperdal came out as less expensive compared with older drugs in the long run. This result didn’t make sense to me and was at odds with everything I saw about the use of Risperdal in clinical settings, where those taking Risperdal should have clearly been doing better if this “finding” was a real one.
Looking at how Catalyst pulled off this trick, it became clear that companies can almost guarantee an outcome like this. The reason: the bedrock on which guidelines depend is the published evidence from company clinical trials. If a guideline is going to be credible, its proposers should have access to all relevant trial data—exactly what the companies appeared to offer (but didn’t). With this assumption, advocates of evidence-based medicine would think that, based on the data, the individual bias of participants or collective bias of the group or any bias stemming from the fact that these were company trials should have little effect on the final conclusions. A group of radiologists, doctors free of drug company influence, or even hostile to industry but prepared to go by the evidence, should come to much the same conclusion on Risperdal as our consensus group—that switching patients from older drugs to Risperdal would save money. This consensus-group meeting resulted in a publication claiming that treatment with Risperdal offered value for money.8 It was followed over the next few years by publications on results from similar exercises undertaken with Zyprexa and other antipsychotics.9
Slightly over a year later I was invited to another consensus conference, again linked to Janssen and Risperdal. The procedure was the same. We had all been sent a dossier with all the published Risperdal trials as well as trials of other new antipsychotics. Any other information we asked for was forthcoming. Based on this material, we were asked what would be the optimal and cost-effective first line of treatment for patients with schizophrenia in chronic care and other treatment settings. Again based on the clinical trial data, Risperdal looked good and “our findings” were presented under our names at major international psychopharmacology settings.10
Pharmaco-economic evaluations like that of our consensus group were, at least on the surface, aimed at costing medical procedures to determine which offered value for money. A few voices at the time were saying that we in medicine couldn’t do what the economists were purporting to do—that too little was understood about what really goes on in medical care. But it seemed clear the pharmaceutical industry was going to pull this new discipline into existence. Drugs function within healthcare the way automobiles do in the wider economy—they can be costed while the degradation of the environment or of medical care remains unmeasurable and uncostable.
Before getting involved in any of these consensus conferences, I had committed myself to the position that pharmaco-economics was bogus science in a debate over claims that the first of the new antipsychotics, Clozaril, which had been launched in 1989 with a price tag of roughly $10,000 per year compared with $100 for the older drugs, was nevertheless cost-effective.11 It was clear at the time that Clozaril had set a price benchmark that, if it did not meet significant resistance, would become the price norm for any subsequent new antipsychotics coming on the market, with major economic consequences for individual patients and health systems in general.
As part of company marketing strategies, economic evaluations of antidepressants also began to come onstream in the 1990s. These purported to show that despite a price of $1,000 per year for a drug like Prozac, compared with the $100 (or less) price tag for older drugs, the new drugs represented value for the money.12 Along with colleagues, I had argued that such prices were even more likely to lead to serious adverse financial consequences for the health services than the even bigger markup on antipsychotics because so many more people were prescribed antidepressants.13 This seemed obvious, but
no one else was saying it. In the face of publications in leading journals claiming the SSRIs or other new treatments would produce savings, there was no dissent.
Given my published positions, it is interesting that company personnel felt confident asking me along to a meeting on economic evaluations and consensus guidelines. Ironically, a few years later, when an independent expert for Britain’s National Institute for Health and Clinical Excellence (NICE) suggested consulting me on the antidepressant guidelines NICE were constructing, the idea was shot down on the basis that Healy was too anticorporate when it came to drugs. Too anti-drug-company for NICE, but just perfect, it seems, for pharmaceutical companies.
Why go to meetings like this? It paid. For many outside observers, the repeated endorsement of on-patent products over older drugs at guideline meetings is close to inexplicable. Finding that the participants at these meetings have at some time been paid by a pharmaceutical company seems the only way to account for this. How else can you explain, for instance, the fact that in these guidelines Healy seems to be endorsing things when he has in other places appeared to say the opposite?
Another factor is lots of us want to be where the action is and industry is very good at creating action or at least the appearance of action. A further factor is friendship. Put in rooms for meetings like these, even people who have been on the opposite sides of arguments in print tend to get on. If others seem friendly in the flesh, it’s somehow easier to see where they’re coming from or to find a way to reconcile views. Companies excel at cultivating friendships—remembering details about you and making you feel that you count. Besides, as the taint of working with industry has receded and as more and more people are linked in, there increasingly seem to be fewer and fewer differences between “them” and “us.” This is a world in which conflict of interest becomes a badge of honor, the more links to the greater number of companies the better.
These are all important issues but the conflicting interests of payment, friendship, or boredom do not explain what happens. Here’s a further puzzle—the guidelines emanating from company-sponsored meetings are all but indistinguishable from those produced by committees with no links to the pharmaceutical industry. Whether the game is played by free market rules or within a socialized system, industry wins.
ONE GUIDELINE, ONE VOICE
To bring out how companies manage to win regardless of which way the game is played, let us contrast practices in the United States and Britain, in particular the operations of the British guideline system run by the National Institute for Health and Clinical Excellence (NICE), widely regarded as the most independent guideline system in the world, and the American Texas Medication Algorithm Project (TMAP).14 TMAP was created by industry. NICE was set up in part to contain industry and has the distinction of having been sued by companies for advising against current drug treatments for Alzheimer’s disease. NICE is exactly the kind of system that the Obama administration looked to put in place as part of its healthcare reform package.15
TMAP was set up in 1994, the year after Risperdal was launched in the United States. The project was initially funded by Janssen but soon thereafter all of the other major pharmaceutical companies had signed on as well. TMAP started with a panel of experts convened to produce a consensus on the use of antipsychotics. Later panels were pulled together to consider the use of antidepressants and mood stabilizers. Many of these consultants had prior links to Janssen and other companies operating in the mental health field, but these experts were distinguished psychiatrists and psychopharmacologists, and none have complained about having data withheld from them.
The first set of TMAP guidelines concluded that the recently launched antipsychotics—Risperdal, Zyprexa, and Seroquel—were the drugs of choice for schizophrenia. A second set of guidelines concluded that rather than older, cheaper antidepressants, the more recently launched on-patent Prozac, Paxil, and Zoloft were now the drugs of choice for depression. Further guidelines moved on to endorse mood stabilizers such as Depakote over other treatments for bipolar disorder. In each case the guidelines recommended newer drugs as safer, more effective, and better tolerated than older agents. In 1999 TMAP commissioned a set of guidelines for the management of childhood mental disorders, even though at the time no psychotropic drugs had been licensed for use in children or teenagers.16
In a number of states, Texas among them, legislators have the power to rule that guidelines such as TMAP’s must be applied in the care of any patients receiving treatment in public facilities. The logic is that evidence-based guidelines, if they really do reflect reality, can be expected to be cost-effective over time. The legislators in Texas meet infrequently, are poorly paid, and are intensively lobbied. Perhaps because of such lobbying, or because pharmaceutical lobbyists were able to show the legislators position papers endorsed by experts, in 1999 the state of Texas endorsed, with no dissenting views, the TMAP guidelines for schizophrenia, mood disorders and for children, thus requiring state hospital doctors to prescribe the newer drugs first.
The TMAP guidelines were subsequently adopted by executive decision in a large number of other states.17 In this way companies have effectively produced a situation in which a growing number of patients on Medicaid and other programs end up being put on and maintained on the newest and most costly of drugs.
The consequences are worth looking at. In 2004, eighth-graders in Pflugerville, Texas were screened by psychologists. Aliah Gleason, a thirteen-year-old, ticked the box for suicidality on one of the tests—probably a probe such as, Have you ever wished you were dead. Even though she was regarded as a live wire in class, this tick led to a referral to a psychiatrist and removal from her family by the child protection services. She was admitted to Austin State Hospital and within hours she was receiving the very best treatment—and did so for the next nine months. This involved all the latest antipsychotics, antidepressants, and mood stabilizers, as mandated by TMAP, costing a small fortune. These were administered not individually but in cocktails of up to five different medications daily. She gained huge amounts of weight, developed a range of side effects, and showed no evidence of progress. It took nine months for her family to get her back and begin to get her off treatment.18
Between 1997 and 2004, Texas Medicaid spending on antipsychotics rose from $28 million to $175 million. In the months of July and August 2004, over 19,000 adolescents in Texas were given antipsychotics, even though pharmaceutical companies had not applied for licenses to market these drugs for use in minors.
In 2003, Zyprexa pulled in $4.3 billion in sales in the United States, 70 percent of which came from state health insurance and other public health programs. It will probably come as no surprise that within all the major companies there are divisions aimed at maximizing the effectiveness of company marketing in the public sector. And it may be no accident that, in 2009, research revealed that children being treated under Medicaid were four times more likely to get antipsychotics than children not covered by Medicaid.19
Surely nothing similar could happen within Britain’s socialized system of medicine, where the key guidelines are produced by NICE, which had been set up with a brief to make recommendations as to the most cost- effective treatments for both physical and mental illnesses? The panelists framing NICE guidelines, whether for cardiac treatments, arthritis management, or psychiatric conditions, have access to the resources of the Cochrane collaboration, the independent organization set up by Iain Chalmers initially in Britain but now with centers in all Western countries that systematically reviews the published evidence—taking pains to obtain all the published evidence and eliminate all evidence that has been duplicated to inflate artificially the apparent benefits of one drug over another. When assembling guidelines, NICE also ensures that it has a range of nonmedical participants to balance out any bias the doctors involved may have in favor of the latest treatment.
Despite this, the 2002 NICE guidelines for the use of antipsychotics came to the same conclusions as
TMAP: newer agents like Risperdal and Zyprexa should be used before older ones.20 Lilly responded to this news by incorporating symbols of NICE and NICE statements into its ads for Zyprexa, which was now supposedly a medication NICE endorsed. NICE had done for Lilly what we’ve seen the FDA had indicated would be illegal in the United States for the company to do for itself.
How come? The first point is that while NICE had access to all the published evidence through the resources of the Cochrane collaboration, this really didn’t amount to any more than they would have been provided by the pharmaceutical companies had they asked. The Cochrane Center had made it clear that there was a great deal of duplicate publication. The four initial trials of Zyprexa in schizophrenia for instance had given rise to 234 publications of one sort or another—almost entirely company written.21 While whittling the publications down to establish just how many trials there had been did help to qualify the apparent benefits of Zyprexa, it made no difference in NICE’s overall evaluation.
What might have made a difference lay elsewhere in the vast amount of data from the four Zyprexa trials that simply could not be found in any of the 234 publications—there was nothing on suicides, diabetes, or cholesterol and little on weight gain. Not one publication hinted that patients given Zyprexa in these trials for schizophrenia had the highest suicide rate in clinical trials history; suicide was in fact rare in schizophrenia before the advent of the antipsychotics.22 Not one publication mentioned that patients in these trials went on to develop diabetes at a rate triple the background rate in the general population, when diabetes was almost unheard of in schizophrenia before the antipsychotics.23 The publications concealed the extent of weight gain in the patients given Zyprexa, whose weight often ballooned by anything from 20 to 140 lbs. These and subsequent publications also failed to reveal that, regardless of diagnosis, Zyprexa raised cholesterol levels more than almost any other drug in medicine—though Zyprexa had received a patent in part based on company claims that it would be less likely than other antipsychotics to raise cholesterol levels.