Pharmageddon
Page 21
Here’s how it happened in the case of bipolar disorder. The first step was to run short-term trials involving patients with much less severe conditions, and less certain diagnoses, using rating scales that may have reflected little more than how highly sedating were the drugs being tested. A strong sedative will always produce a “signal” on a rating scale for mania—the patient will be less active, less talkative, less disinhibited while under the influence of the drug. This is all it takes to get FDA approval for company claims their drug is antimanic. As a result of these trials all guidelines from the first formulated by TMAP in 1998 to a set of NICE guidelines in 2006 recommend the use of on-patent antipsychotics—but not any of the older antipsychotics.46
The second step was to run debatably ethical trials elsewhere, such as one Janssen ran in 2003 and 2004 on Risperdal for mania in India.47 This trial became the subject of a BBC investigation into the ethics of studies for Western treatments outsourced to India. Did patients know they were involved in research? Did they consent to it? Did they know that once their participation was over they would be removed from drugs that might have been helping them? And it wasn’t just ethics that was at issue. The correspondence in the columns of the British Journal of Psychiatry on the validity of this study was more extensive than for any other study the journal has published.48
In this newly globalizing clinical trial world, everyone faces a future in which the bulk of the evidence that dictates clinical practice when it comes to the use of statins, antihypertensives, painkillers, antibiotics, and practically everything else from mental health to respirology will come from settings that are very different from those in which the treatment will be given. There are likely to be many consequences for clinical practice, not least because both efficacy and side effects of different medicines may vary markedly in different ethnic groups.
As a result of the trials undertaken in India and elsewhere, only the new antipsychotics had randomized controlled trial support. The older agents hadn’t. The straitjacket of current notions of evidence-based medicine, as applied by guideline bodies like TMAP or NICE, places published evidence from controlled trials above everything else—almost to the absurd point of not using a parachute until a study is undertaken to indicate formally its usefulness. What’s more, the marketing departments of companies depend on our fascination with the supposed ironclad science of controlled trials and use it to capture the process of developing guidelines.
The third step involves something close to checkmate. In the case of manic-depressive illness the only agent with an established prophylactic effect is lithium. But modern guidelines also variably recommend Zyprexa, Depakote, and other new antipsychotics or anticonvulsants even though these are not licensed for this purpose. This has essentially happened because Abbott heavily advertized Depakote as a mood stabilizer in the first instance and the companies with follow- up anticonvulsants and antipsychotics followed suit. The term “mood stabilizer” generates expectations of a prophylactic effect even though none has been shown. Claiming Depakote was prophylactic would have been illegal—but there was no need for Abbott or other companies to tempt the law when a prestigious guideline recommends Depakote for a use the regulator would not let the company claim. In this case, bound by the law, the FDA is a lot more stringent than the guideline makers. This is advertising that’s hard to beat.
The final step involves the use of guidelines to create new disorders. Over a century of clinical opinion has unanimously held that bipolar disorder can occasionally start in adolescence but usually has a later onset. The guidelines makers are trapped into mentioning pediatric bipolar disorder by the simple fact that companies have published a number of trials giving sedative drugs to unruly children, labeled as suffering from bipolar disorder. Being value neutral, because such trials had been run, in their 2006 guideline NICE had to mention pediatric bipolar disorder. In so doing they breached a century of worldwide clinical consensus, and all but endorsed the disorder, pushing Europe down a route the United States had already traveled.49
When it comes to bipolar disorder, American medicine is in the grip of an enthusiasm reminiscent of the seventeenth century Dutch tulip mania. Children as young as one year of age are being put on antipsychotics, and some clinicians even contemplate the possibility of making in-utero diagnoses. Guidelines have been a significant factor in this infection. In recent years a series of pediatric bipolar consensus conferences were organized in the United States, such as one organized by Best Practices, a marketing firm specializing in central nervous system drugs.50 This conference was supported by all the major pharmaceutical companies, and its final recommendations were ghostwritten, but even if such meetings weren’t financially supported in this way with carefully sculpted ghostwritten recommendations, the result would likely have been the same. Running trials of sedative drugs in overactive or disruptive children, who are labeled bipolar, will produce an apparent benefit. That clinical trial result in effect pulls a guideline into existence, and if there’s a guideline, the condition is assumed to be real. All the marketing company need do is ensure the guideline making process happens in a timely fashion, with a consensus statement for publication and dissemination.
Once the participants agree that the guideline has to be based on clinical trial evidence, the guideline all but writes itself before the participants sit around the table. The guidelines produced by TMAP in 1998 for the treatment of bipolar disorder are essentially indistinguishable from those produced by NICE in 2006. Where in 2004 NICE were saved by a television program from contributing to making children depressed, nothing saved them in 2006. The reasons for NICE’s failure to distinguish itself from TMAP in 2006 do a great deal to help explain our current healthcare problems.
FACTS ON THE GROUND
Across medicine, however misleadingly certain academic papers may be written, with a few exceptions, no studies allow claims that one drug is superior to another. Even so, a series of guidelines in different areas of medicine advocate newer, more expensive drugs over older ones. However well-meaning these may be, in these cases there should be suspicions that the guideline has been captured by pharmaceutical companies.
Capture is engineered by a combination of smart publication strategies and targeting trials at illnesses where there have been no trials before, whether restless legs syndrome, female sexual dysfunction (FSD), or osteoporosis. In these ways, companies can make diseases fashionable, can engineer the appearances of comparative efficacy, and can enlist academic advocates for particular treatment options. By these means, too, they have been able to control the content of guidelines and transform even independent guidelines into something close to an extension of company marketing departments.
This dynamic plays a key role in the selling of diseases from FSD to PTSD, overactive bladder, osteoporosis, and osteopenia.51 Getting a drug licensed for FSD or osteoporosis does not mean that physicians are thereafter enabled to treat women in a more effective way than they had been able to do before. Rather, it means that Pfizer, Lilly, and GlaxoSmithKline are enabled to start marketing these disorders and in the process to convert the vicissitudes of intimate life on the one hand or the changes of middle years on the other hand into illnesses. Guidelines achieve even more for a company—they make it appropriate, indeed almost necessary, to detect and treat these illnesses. Company-sponsored and ghostwritten “scientific” papers, along with selectively presented trial results, offer the raw material out of which clinical consensus will later be manufactured. When it comes to annexing territory, this clinical consensus in the form of guidelines establishes facts on the ground.
Consider what happens when a guideline is published. For managers running healthcare institutions, there need to be standards against which the organization can be held accountable. Whether or not the current guidelines are wrong is immaterial. If attempting to implement them produces no health gain, this still makes no difference to managers, at least in their strictly institutional role. The
key point is adherence.
A celebrated episode from the fourth series of the American TV medical drama House shows Dr. Foreman grappling with a patient’s life-threatening problem. He ultimately finds an unorthodox answer to it that saves his patient’s life but gets him the sack. As his boss tells him, it may have been good medicine but it was bad practice. Dr. Foreman is not alone; clinicians worldwide are increasingly faced with managers enquiring about their compliance with guidelines and more and more are getting the sack. What’s a manager going to do if a doctor retorts that these ostensibly evidence-based guidelines amount to pharmaceutical marketing by proxy?
The accountants in the finance department of a healthcare organization who see the figures on newer and more costly drugs also find themselves faced with guidelines supporting the use of these drugs, issued by independent academic bodies whose stated brief is, in part, to secure cost effectiveness. The promise is that the organization will save money in the longer run by being “evidence based,” as this will lead to better outcomes for diseases treated this way and to savings on not doing what works less well. The beans line up for both the accountants and the executives. Truth does not.
Articles by guideline proponents, and even the guidelines themselves, state that clinicians do not always need to adhere to the guidelines—this is guidance rather than a diktat.52 But medico-legal articles suggest that any deviation from guidelines needs to be justified. Where a clinician wouldn’t have to justify guideline-sanctioned treatment in the medical record, they are advised to justify everything that is “unorthodox.” Doing anything different, then, adds to the bureaucracy, and increases the sense of risk.
An element of coercion has also emerged in many medical settings where reimbursement has been tied to guideline adherence. The element of coercion increases further if one considers that current evidence is framed within settings in which pharmaceutical companies advertise (in the United States) and set up patient groups who lobby for new treatments even though there is no evidence to believe these are any better than older treatments.
The proponents of guidelines put them forward as guidance and believed that they could only lead to improvements in the outcomes of treatment for all conditions. But a series of studies have shown that the outcomes, on average, are in general no different whether or not clinicians adhere to guidelines.53 Soon after guidelines began to appear in the 1980s, opposition to them emerged and grew steadily.
Clinical concerns that guidelines risk becoming coercive are often met with a cynical response—of course clinicians will be worried if their autonomy is being curtailed. While not untrue, this misses an essential point. If a treatment really works, both because they want to help their patients but also for reasons of compelling self-interest, few clinicians are likely to fail to prescribe it whatever the guidelines may say. Who would not give penicillin to a patient with pneumonia or an antipsychotic to a floridly manic patient?
The problem guidelines might pose was outlined first in 1956 long before anyone had heard of them. Following the discovery of the first antipsychotic, chlorpromazine, the National Academy of Sciences (NAS) and National Institute of Mental Health (NIMH) convened a meeting to work out how to build on this discovery. Ed Evarts from the NIMH, one of the leading lights of the day, put it to his colleagues that but for an accident of history they could now be discussing the use of the new antipsychotics for the treatment of dementia paralytica (tertiary syphilis) rather than dementia praecox (schizophrenia). Tertiary syphilis had looked identical to schizophrenia and chlorpromazine would have produced a distinct benefit on this state because it controlled the hallucinations and delusions that went with the disorder, although likely at a cost of increasing mortality—but this increase in mortality would not have shown up in the short-term clinical trials that demonstrated a benefit.
Evarts pointed out to his audience that none of the rating scales, clinical trial methods, or animal models that were then being put in place as the engines of progress that would move the new psychopharmacology field forward would have helped doctors to work out that penicillin rather than chlorpromazine or psychotherapy was the right answer to dementia paralytica. What made the difference was understanding that tertiary syphilis was a microbial infection. He predicted that the proposed scaffolding of clinical trials, although eminently sensible, would create an academic and industrial complex that would slowly stifle progress in therapeutics.54
No one paid heed to Evarts. He came to the conference as a leading figure within the psychiatry of his day but vanished from the radar afterward—leaving a set of predictions that have been right on the money. Fifty years later, compulsory detentions for mental illness have risen three-fold, admissions for serious mental illness have risen sevenfold, admissions overall have risen fifteen-fold,55 suicide rates in schizophrenia have gone up twenty-fold,56 and diseases such as diabetes have increased exponentially among the mentally ill.57 There has been a dramatic drop in life expectancy for serious mental illness in the United States—with a fall of up to two decades compared with the rest of the population.58 The same has been found wherever else these things have been measured,59 with increases in mortality correlated with the numbers of psychotropic drugs given.60
We have focused on mental health in this chapter, but the same is happening in other areas of medicine where there are blockbuster drugs. The interaction between the first of the blockbusters, Zantac, and the treatment of ulcers, outlined in chapter 2, bears out Evarts’s prediction better than anything else. Many doctors had been using antibiotics for ulcers before Barry Marshall demonstrated that ulcers were often caused by the helicobacter pylori bacillus. Had there been guidelines for the treatment of ulcers then, any doctors prescribing antibiotics would have been at greater risk of a lawsuit than they might have been before the guidelines were formulated.
Current cardiovascular guidelines all mandate lowering low-density lipoprotein (LDL). Company marketing took advantage of this with Merck and Schering Plough suggesting that Vytorin (a combination of ezetimibe and simvastatin) would lower LDL cholesterol further than would treatment with a statin alone. The thrust of the guideline played into the marketing of Vytorin—until the clinical trial evidence finally demonstrated that prescribing Vytorin produced no benefit in terms of mortality.61 Hormone replacement therapy entered guidelines as a means of lowering cholesterol, but it is now clear this increases death rates.62 Cardiovascular guidelines also call for optimal control of blood pressure, and company marketing has suggested adding angiotensin receptor antagonists (ARBs) to ACE inhibitors as one way to do this, but the clinical trial evidence now suggests that this also increases mortality.63
For the treatment of diabetes, guidelines recommend tight glucose control. GlaxoSmithKline’s Avandia (rosiglitazone) was promoted as doing just this, making the company billions of dollars annually in the process, until it was withdrawn following evidence that Avandia increases rates of heart attacks and death by up to five hundred cases per month above what might have been expected had other agents been used.64 The question of whether GlaxoSmithKline knew about these risks and hid clinical trial data—just as they did with Paxil in both children and adults—became the subject of a US Senate investigation, discussed in chapter 7.65 More generally, large-scale studies have shown that adhering to these diabetes guidelines have led to higher death rates and more hypoglyemic episodes than found in patients treated with less emphasis on tight glucose control.66
The country that consumes the greatest amount of on-patent medications and the greatest amount of medications attested to by the most authoritative guidelines is the United States, but over the past decade, American life expectancies have progressively fallen behind other developed countries.67 Over the same period of time spending on health has escalated in the United States beyond elsewhere, rising from less than 1 percent before World War II per annum to over 17 percent of GDP now. This is not what happens when treatments work. It is not what happened to the clinics and bed
s used to treat patients with tertiary syphilis after the discovery of penicillin, or tuberculosis after the development of streptomycin—when the patients vanished, the beds were closed down, the staff redeployed, and money was saved. The promise of the guideline makers—that if only policymakers follow the evidence (such as it is), health will improve while costs come down—has not held up.
When faced with evidence that guideline-mandated treatment with statins, antidepressants, or drugs for osteoporosis fails to make a difference, guideline makers sometimes attribute this failure to a delay in instituting treatment. In July 2008 the American Association of Pediatricians issued a new guideline on the health of children. It recommended screening children as young as eight years old for raised cholesterol levels, and then possibly instituting treatment with a statin.68 The promise is held out that catching people ever earlier will make a difference. Similarly, advocates of mood stabilizers commonly attribute the failure of their drugs to make a difference to the delay in starting the drugs, and they suggest catching and treating ever and ever younger children. Once the disease takes hold it is supposedly more resistant to treatment.
It seems strikingly difficult for clinicians and others to ask whether robust independent assessment of drugs can be undertaken in a world where data is privately held. The reviewers for the NICE guidelines on pediatric depression teetered on the brink of making this point but backed down. The point was finally made in January 2011 by the Cochrane Center reviewers of Tamiflu, who made it clear that in the current circumstances we have little option but to recognize that independent assessment of drugs is not possible.69 There has so far been a deafening silence from Western governments, all of whom have handed over billions of dollars to stockpile a remedy little better than one of the proprietary nostrums from the nineteenth century.