by David Healy
Several years before this, in 2004, documents demonstrating that GlaxoSmithKline intended to hide the parts of Study 329 that didn't suit them led New York State to take a legal action that resulted in an agreement by the company to post the results of all its clinical trials on the web. Other companies agreed to do something similar and many journals and academics expressed the view that we had turned a corner, unaware that companies intended to post study reports rather than data.
In fact, neither all of the paroxetine trials nor the underlying data from these trials has been posted as of February 2011.63 What is posted is a set of summaries of studies, authored by company personnel. Despite repeated statements by companies that they now post their studies on the web, this has not been the case. In creating a situation where companies can convey the impression they have come clean with their data, the action by New York State paradoxically may have made things worse by enabling companies to claim they are being transparent.
There are great risks that those attempting to develop policies in these domains, if they don't fully address the deeper sources of the problems involved, will come up with answers that, like Kefauver's 1962 amendments to the Food and Drugs Act, compound rather than remedy our problems. The failure to require access to trial data in particular plays straight into the hands of pharmaceutical companies.
Within medicine, trust in “scientific” data to set us free remains strong. But by failing to challenge companies about their data, we have handed over the instrument of our freedom to company marketers and risk ending up increasingly trussed in company-generated evidence. These are the “data” that tell insurance companies that the latest blockbusters work better than anything that came before and that they come with few if any serious hazards, so that their use will lead to ever shorter hospital stays. Embodied in guidelines, these “data” increasingly limit the abilities of doctors to exercise any discretion in caring for the person in front of them. These “data” should improve the quality of medical care but instead are leading to Pharmageddon.
5
Trussed in Guidelines
Bill was in his seventies, tall and relatively fit for his age if slightly overweight. His wife was petite. She gave every impression of having been dependent on a physically and behaviorally imposing husband, although their roles were now reversed. Bill had had a stroke a month earlier and Sally was distraught. She was sure that he could recover and concerned that the medical team had not referred him for active rehabilitation. He had shown no signs of recovery of function after his stroke, though. In their opinion there was nothing to build on, but they had asked me to assess if there might be psychological factors or a depression holding him back.
When I saw Bill he had no language. He appeared to be trying desperately to communicate, however, almost like Jean-Dominique Bauby in his diving bell. He had something between a cough and an effort to clear his throat that apparently had persisted for weeks. He hacked every few minutes, then fixed me with what appeared to be a pleading look in his eyes. Some patients after a stroke cannot clear saliva pooled in their esophagus, but Bill’s hacking was different in its quality and persistence. “Why can’t they do something about it,” Sally said, “surely they can make him more comfortable.”
Bill was on a cholesterol-lowering statin and an ACE inhibitor to lower his blood pressure, I discovered. The attending doctor had said it was in line with current international guidelines to put everyone who had a stroke on a statin and an ACE inhibitor.
I recommended stopping both treatments: either of these two drugs could have been holding Bill back from making some progress. The statins can cause muscle pain and weakness, which he could not now complain about. If the medical team was planning no more active intervention on his behalf, why not stop the treatment and see? All that guideline-based treatment could do, at best, would be to prolong an agonizing life. As for the ACE inhibitor, it was almost certainly causing his hacking cough—this was an unusual but known side effect of this group of drugs.
But guidelines were guidelines, and the medical team was unlikely to go against them. There was no point telling Sally that the hacking could be sorted out. Why set her against the doctors even more than she already was if they were so set in their ways? Generating hostility on their part at betrayal by a medical colleague’s interference wouldn’t do Bill any good, but it might compromise my ability to make a difference for someone else.
Aside from the horror of this case, few physicians would see anything remarkable in it. An unfortunate medical error in the treatment of this particular patient perhaps, but it’s impossible to practice medicine without errors. Better a few have grim outcomes like this than have more lives lost because of a failure to manage patients properly after a stroke.
The problem is, the distress Bill had to put up with so that his doctors could feel comfortable and comply with ostensibly the best available evidence as embodied in the latest authoritative guideline is fast becoming the clinical norm rather than an exception. There was once only a small number of exceptions, as for instance in the case of vaccination, where medicine was prepared to inflict vaccine-induced injuries on some in the hope that a much greater number would benefit, but this ethos is changing rapidly.
Take Sheila, who had had anxiety and agoraphobia during the 1960s—she would likely have been diagnosed with panic attacks today. She had been caught up for years in what was an almost automatic prescription of benzodiazepines for anxiety and became physically dependent on them. The combination of anxiety and benzodiazepine dependence made her agoraphobic. She was scared to venture out of the house.
When her husband died two decades later, everyone feared she would become dependent on her children. Remarkably, she instead struck out on her own. She bought an apartment, some distance from any of her children, and began socializing in a way she had not done for over two decades.
When the alarm was raised in the 1980s about benzodiazepine dependence, many primary care physicians changed their patients to low doses of antipsychotics or antidepressants instead, and Sheila’s new doctor was no exception.
I first met Sheila around the time that the selective serotonin reuptake inhibiting (SSRI) antidepressants came on stream. Her doctor referred her to me for review of her medications. Rather than the combination of an antipsychotic and antidepressant she was on, I started her on an SSRI. At first, she did much better. But shortly afterward she began grinding her teeth. We changed her to another SSRI. Again she initially did well, but then the grinding and restlessness commenced again. The same happened with each of the four SSRIs then available.
Sheila’s teeth grinding was so intensely painful that she had to remove her dentures. With her teeth out she became more self-conscious and grew more reclusive. She was slipping back into the shell in which she had lived for over twenty years. I opted to go back to one of the older anti- depressants which, in a low dose, made her less anxious without causing teeth grinding and restlessness. We met regularly thereafter for close to ten years during which time she maintained a delicate equilibrium.
Then, at the age of eighty, she had a “turn” and was brought to hospital. There were some inconclusive changes on her cardiogram, possibly indicative of a minor heart attack, and Sheila came out of hospital on both an SSRI and a statin. The SSRI had been prescribed by the hospital medical team because it was supposedly safer for the cardiovascular system than her older antidepressant, the statin because international guidelines now recommend statins for everyone who has had a cardiac event—regardless of whether the person’s cholesterol levels are high.
Sheila now developed two sets of problems. Her teeth began to grind again and her legs became so weak and painful that she fell when she least expected it, so she couldn’t go out to the shops or to see friends. I was asked to see her. I suggested stopping the statin as it was probably causing the weakness and pain in her legs, and switching her back to her previous antidepressant. Her primary care doctor was faced with a cho
ice between my advice and the input from the medical team. He opted to continue the statin and SSRI prescriptions. The calculation he apparently felt called on to make was whether it was better to keep her alive, although disabled by treatment, or give her a better quality of life but risk her dying earlier than she would otherwise have done.
For over twenty years I have copied my patients on all correspondence that concerns them. Sheila had the letters to her doctor recommending that he stop her SSRI and statin. Although she told me that she was sure I was right about the drugs, she didn’t demand her doctor do what I recommended. She was nervous that in her current frail state she might suddenly have a medical emergency and would be critically dependent on him. She had a niggling doubt that he might not be as quick to help her if she were a difficult patient. She was a hostage—as many patients are.
Her doctor was finally persuaded that SSRIs were no better for the cardiovascular system than her older antidepressant and might actually increase her risk of a stroke especially when combined with the aspirin she was on. Switching antidepressants improved her tooth grinding and her restlessness. But Sheila never recovered. The statin was still prescribed and her leg pains and weakness remained. Unhappy and lonely, she ended up in a residential home.
Sheila’s doctor never let pharmaceutical company sales reps into his practice. He had no dealings with industry. Yet here he had been doing exactly what industry would have wanted—and seemingly oblivious to the obvious difficulties his patient was having in front of his eyes. The problem he had and the biggest problems Sheila and Bill faced had a common source: sets of guidelines produced by medical organizations, in both the United States and Europe, in the hope that these guidelines might improve medical care and provide a bulwark against company marketing. But these same guidelines have instead too often become a means to harness the medical impulse to give the best possible care to the delivery of the latest drugs, even when these offer less benefit and more harm than older treatments.
THE END OF DISCRETION
As 2009 closed a controversy erupted across the pages of medical journals concerning Tamiflu (oseltamivir), an antiviral drug produced by Roche, which had had several years of good fortune as Western governments stockpiled the drug, fearing a pandemic first of avian flu and then of swine flu. The published evidence appeared to indicate that Tamiflu reduced the likelihood of a full-blown influenza, reduced the length of a flu episode that developed, and reduced the secondary complications of influenza such as pneumonia or other respiratory disorders that might lead to hospitalization and even death in vulnerable groups. This led national governments throughout the Western world, and agencies like the Center for Disease Control in the United States, to a set of recommendations to doctors on the management of flu that hinged on a widespread use of Tamiflu. The trouble was, no one could access the data on which these recommendations were made. Furthermore, it became increasingly clear that only a fraction of the trials that had been undertaken were published, and of those published, ghostwriters had played a significant role in what was reported.1 The more material leaked into the public domain, the less effective Tamiflu looked and the more dangerous using it began to seem—it appeared to induce neurological problems in a subgroup of patients and to make others suicidal. But a further dilemma came into view—governments had spent billions on this drug. Would they admit they had spent billions on a drug for which they had seen only a portion of the evidence and that might not work as designed? Would they pressure Roche to release all the data on the drug?2
Sequestering data violates a basic norm of science even if it is overlooked by law. Today when public policy at many levels is or aspires to be based on scientific data, such violations have ever greater ramifications, from the individual treatment our doctor gives us to decisions about national and international health care. To see how hiding medical data directly affects the doctors we consult and the quality of medical care we receive, we need to explore two aspects of everyday medical practice: the increasing use of guidelines and what they are based on (the subject of this chapter) and the interests behind the measurement technologies to which practitioners like Dr. N, discussed in the introduction, turn (the subject of the next chapter).
The evolution of guidelines is best told firsthand. For that reason we will focus on guidelines for the treatment of mental health disorders, but the story that unfolds here parallels developments in other areas of medicine—and it is these developments that ensnared Bill and Sheila. In every area of medicine, doctors increasingly find they have to take into account guidelines or standards that have been established, not infrequently to the detriment of the patient in front of them. It is against such guidelines that medical personnel are ever more likely to be judged by the managers of the service they work in or by the legal system should one of their patients take an action against them.
Our point of entry into the story lies in 1993 when the Janssen pharmaceutical company was hoping to bring their new antipsychotic
Risperdal (risperidone) to the market. An FDA review of this drug prior to its launch stated that “We would consider any advertisement, promotion or labeling for Risperdal false, misleading or lacking fair balance under Section 502 of the Act if there is a presentation of data that conveys the impression that risperidone is superior to haloperidol or any other marketed antipsychotic drug product with regard to safety or effectiveness.”3
All of the antipsychotics developed during the 1990s, from Risperdal to Lilly’s Zyprexa, Astra-Zeneca’s Seroquel, Pfizer’s Geodon, and Bristol-Myers Squibb’s Abilify (among others), had been tested in premarketing trials against haloperidol, one of an earlier generation of now off-patent antipsychotics. In their trials all of the companies used a higher dose of haloperidol than clinically needed.4 The stated rationale for using haloperidol as the comparison drug was that it was supposedly the market leader. The unstated rationale was that given the side-effect profile of the newer drugs they stood their best chance of looking good from a marketing point of view if compared to high-dose haloperidol. This kind of comparison is standard company practice for bringing any new drug to market, whether statins, antihypertensives, painkillers, treatments for osteoporosis, or for gut problems—compare your drug to some formulation of an older compound against which the new drug is already known to look good on some parameter.
On the face of it, the FDA’s cautionary note, repeated for subsequent antipsychotics and in slightly different form for cholesterol lowering statins, proton pump inhibitors for gut disorders, the latest antihypertensives, the Cox-2 inhibiting painkillers, or biphosphonates for osteoporosis, looks like it should produce problems for any company wishing to market new drugs that, like Risperdal, can cost up to fifty times as much as older drugs.
There are lots of ways to get people to take a new drug that may be no more effective than an older one, however. For one thing new drugs come with a hope of superior efficacy built in that older drugs have lost, so we want them. But how to price up this hope—is the right answer twenty-five or fifty times the price of older drugs when the new drug is no better than the old drug for the same malady? Patients trade on such hopes, and one approach companies now take is to set up patient groups to lobby for the new drug. Such groups are all too willing to believe they are being denied access to the latest and best treatment on cost-cutting grounds. And it is difficult for doctors or, more importantly these days, politicians or insurance companies to resist articulate patients who question whether they are being denied the newest and best treatments on the basis of economic rationing. Doctor, what would you give to a member of your own family?
Patient hopes and expectations work in favor of a pharmaceutical company bringing a new drug to the market, but in addition since the 1990s, doctors in many countries, whether they work for a health maintenance organization or in a universal healthcare system, have also had to adhere to drug formularies (lists of approved drugs) which dictate what they can and cannot prescribe. These formular
ies arose in response to perceptions that healthcare costs were escalating uncontrollably and that a key element in this escalation was the price of drugs. The formularies often start with a principle that, where possible, doctors should prescribe cheaper generic rather than higher cost, branded compounds. The guidelines are intended to be both evidence-based and cost-sensitive—with some trade offs, so that if a new drug costs more but could show a real benefit over older agents, for example, it would be included on the approved list. The types of assessments pharmacists and doctors with no links to industry would make in constructing formularies, it was thought, would in general slow the entry of unnecessary new drugs to the market.
When it came to managing costs, from the 1990s onward service managers and others could, at least in theory, also turn to health economics to assess company claims that their new drugs offered good value for the money. And of course if the market really was a free market and several different companies each brought to market new antihypertensives or treatments for osteoporosis, competition should drive the price of the new drugs down—as many from Senator Kevaufer in 1962 onward have argued.5 But this has never happened.