In fact drug and device makers sometimes suppress research results that are unfavorable to their products. Sometimes they publish favorable data from the same patients several times, creating redundant research reports that make it hard to know just how many studies were really done and how many patients were studied.
Companies provide some research results to the regulatory agencies in order to gain approval for new products, but they aren’t required to publish these results where other scientists and doctors can evaluate them. For example, a recent systematic review on nonsteroidal anti-inflammatory drugs ibuprofen, naproxen, Celebrex, Vioxx, and others-found that only one of thirty-seven studies reported in FDA reviews had been published. (See MacLean CH, Morton SC, Of man, Roth EA, Shekelle PG. How useful are unpublished data from the Food and Drug Administration in meta-analysis? Clin Epidemiol 2003; 56: 44-5l.)
And the problem with unpublished research goes even farther. Many randomized trials get started but are never finished, for a variety of reasons. These rarely get reported, but they may offer important insights to doctors and patients. (See Psaty BM, Rennie D. Stopping medical research to save money: A broken pact with researchers and patients.]AMA, 289, 2003: 2128-213l.)
Even when trials are completed, the ones that turn out negative-that is, showing no advantage for a new treatment-are less likely to get published than the ones that are positive. (See Krzyzanowska MK, Pintilie M, Tannock IF. Factors associated with failure to publish large randomized trials presented at an oncology meeting. ]AMA 2003; 290: 495-50l.)
Some clinical trials get reported only verbally at research meetings or in meeting abstracts; some never get published at all. (Dickersin K, Rennie D. Registering clinical trials. ]AMA 2003; 290: 516-523.)
The reasons are complex, but it appears likely that when the results of their research are negative, researchers often lose their enthusiasm and fail to write up the results and submit them for publication. Less often, journal editors or reviewers may be biased against publishing negative studies.. Even if all the trials on a topic do get published, the ones with positive results tend to get published faster than the ones with negative results. (See Stern ]M, Simes R]. Publication bias: Evidence of delayed publication in a cohort study of clinical research projects. Br Med] 1997; 315: 640-645.)
Therefore, the good news about a treatment often appears well before the bad news. The media coverage, in turn, tends to be unrealistically favorable. The greater likelihood for favorable studies than for negative studies to get published has been labeled publication bias. The bias it creates in adopting medical advances may actually harm patients. As an example, one study of drugs for suppressing abnormal heart rhythms was completed in 1980, but was not published until 1993. (Cowley A], Skene A Stainer K, Hampton]R. The effect of lorcainide on arrhythmias and survival in patients with acute myocardial infarction: An example of publication bias. Int] Cardiol1993; 40: 161-166.) The study showed higher mortality in the group with the active drug than with placebo, consistent with the later studies of encainide and flecainide. Had these negative results been widely known, some experts believe that the use of dangerous drugs might have decreased sooner, saving many lives. Besides distorting the evidence for what works, failure to publish leads to waste and duplication in research efforts.
Consequently, many scientists, journal editors, expert panels, and even presidential committees have argued for a system that would register all clinical trials.15 The NIH and some other groups have made steps in this direction, but they’ve been only partly successful. Even though the FDA has mandated registration of clinical trials evaluating new drugs for serious or life-threatening diseases, it has no mechanism for enforcement, and compliance appears to be poor. The barriers have been reluctance of the drug industry to participate, lack of funding, lack of any enforcement authority, and lack of publicity or awareness.
Once patients understand how little is known about most medical problems and treatments, perhaps their interest in being participants in medical research will increase. Some experts believe that patients who are enrolled in clinical trials get better medical care than those outside clinical trials, regardless of whether they are receiving an experimental treatment or standard care. (See Braunholtz DA, Edwards S]L, Lilford R]. Are randomized clinical trials good for us (in the short term)? Evidence for a “trial effect.” J. Clin Epidemiol 2001; 54: 217-224.) And participating in research is the best way to get better answers about what really works.
One aspect of informed consent that should be assured is that patients understand the researchers’ conflicts of interest. If the researchers hold stock in the company that makes the products they’re testing, patients should know that. If the researchers have a grant from the manufacturer to do the research, patients should know that. If the researchers receive a special reimbursement for every subject they enroll, patients should know that. And if the researchers simply get consulting fees from the corporate sponsor of the research, patients should know that, too.
Surprisingly, this information, and even the identity of the research sponsor, isn’t currently a required part of the informed consent process at many institutions. We think that such disclosures should be a routine aspect of the ethical review process and a part of every research consent form.
So for example in 2002, James Quinn died in a Philadelphia intensive care unit. He had lived nine months with a new AbioCor artificial heart. For seven of those months, he had been in the hospital, where he suffered two strokes and was on a ventilator much of the time. Before he died, Quinn said, “This is nothing like I thought it would be. If I had to do it over again, I wouldn’t do it. 1
In 2001, Bayer removed its cholesterol-lowering drug Baycol from the market. Between 1997 and 2001, manufacturers pulled fourteen drugs off the market because of unexpected toxicity and related deaths. The makers had marketed these drugs aggressively; they were taken by millions of patients and collectively are suspected in over a thousand deaths.2
In 2002, researchers reported that hormone replacement therapy for postmenopausal women may often cause more harm than benefit, although it’s been used for decades.3 The same week, we learned that a popular arthroscopic operation for knee arthritis doesn’t work.4
In 2001, twenty-five-year-old Jennifer Rufer of Seattle discovered that she didn’t have cancer after all. Blood tests had repeatedly suggested a rare form of uterine cancer, and she had endured an unnecessary hysterectomy, removal of part of a lung, and months of chemotherapy.5 Rufer had wanted a big family, but the false positive tests destroyed that dream.
In 2000, the New England Journal of Medicine reported that an aggressive new treatment for breast cancer was no better than standard care, although it was more toxic and cost twice as much.6 Over 42,000 women had received the new treatment, at a total cost of $3.4 billion.7
In 2002, researchers reported the largest clinical trial ever conducted for high blood pressure, involving 33,000 patients. An old fashioned diuretic drug proved more effective than newer drugs in preventing the complications of high blood pressure.8 But the use of diuretics had fallen steadily for years before the trial, thanks to aggressive marketing of newer and higher-priced medicines.9
As doctors and manufacturers introduce medical advances, good money too often trumps good science. Vested interests, marketing, politics, and media hype often have more influence on how new medical advances get used than the best scientific evidence.
New tests, devices, or drugs are routinely introduced as “breakthroughs”-when, in reality, many are marginally effective, useless, or even harmful.
The best scientific evidence may get pushed to the background, suppressed, manipulated, or ignored. Good science sometimes wins in the end, but often only after avoidable illness, discomfort, costs, and delays.
Doctors sometimes make use of medical advances, at some risk and substantial cost, for the flimsiest of reasons.
A treatment for patients who had been referred for cardiac catheterization once involved threading a long tube through the aorta to the heart and injecting dye into the arteries that supply the heart. It’s usually safe, but serious complications (like a heart attack) sometimes occured.
With the availability of balloon angioplasty, which opens narrowed arteries by inflating a tiny balloon inside the narrowed area, cardiac catheterization has become not just a diagnostic test, but also a treatment opportunity. And with the addition of coronary stents small wire tubes that prop open narrowed arteries), the treatment opportunities and costs have expanded further.
Unfortunately, these treatments are increasingly being used for patients who don’t even have chest pain or other symptoms-a situation in which the risks are clear, but the benefits are speculative.
In some cases, the medical needs are more compelling, but using the latest treatments can still be hard to justify because the benefits are minuscule or nonexistent. An example would be aggressive treatment of an elderly patient who is near death from an incurable chronic disease. Most treatments have smaller benefits in this situation than in others, and the risks are usually greater.
The pressure from the public and from industry to develop and disseminate new treatments is intense. In some cases, patients have no effective treatments for their diseases and are willing to pay-or at least have their insurance pay-for anything new that comes along. But most innovations aren’t treatments for fatal diseases. Instead, they’re aimed at discomforts and inconveniences, many of which already have effective treatments. And some new technologies ultimately prove ineffective or harmful, although this may become clear only after many patients have received them. Also, in most cases, “new” seems to be synonymous with outrageously expensive.
When Do New Treatments Stop Being Experimental?
It may be surprising to learn that there are no standard definitions of experimental and standard care. There’s no “Good Housekeeping Seal of Approval” that new operations, devices, tests, or drugs must receive before they become standard care. There’s not even agreement about what evidence of effectiveness should be required if there were to be such a seal.
Many people imagine in the USA that FDA (at times stricter then European authorities) approval constitutes such a seal. It’s true that for drugs there must be a reasonable demonstration of safety and efficacy, but the FDA makes no claims concerning value for money The FDA doesn’t require comparative studies to determine which of several competing drugs is most effective. As a result, products with trivial benefits often receive regulatory approval, regardless of their cost. An expensive new drug may offer no advantages over a cheap old one, but if it works at all, the FDA generally must approve it. For medical devices, the FDAs approval process is less rigorous than it is for drugs, and for new surgical procedures, there’s no approval process at all. The FDA also gives little scrutiny to nutritional supplements-a multibillion-dollar industry.
There’s no standard for judging when new medical advances are worth the cost, or when insurance should cover them. Attempts by researchers to establish such limits have never reached consensus. Every insurance plan or health-care organization makes this determination on its own, with wide variability in the process. Insurers often cover new treatments because others do, because effective marketing creates demand, or because lawsuits are threatened-not because of convincing evidence that the new treatments work.
In fact the evolution of clinical practices seem to share a great deal with the evolution of public policy-it’s the result of competing interests, rather than a linear translation of scientific knowledge into practical application. Students of human behavior, economics, and political science may think it naïve to expect the development and diffusion of medical treatments to follow a rational process. Doctors are sometimes surprised when they see marketing or media hype triumph over scientific knowledge. But no outrage materializes because most of the time, neither patients nor doctors even recognize them.
The Backlash Against Evidence-Based Medicine
Evidence-based medicine is a hot topic right now among doctors, insurance companies, planners, and the public. There are frequent workshops on how to practice and teach it, as well as workshops on its role in insurance coverage and policy making. Medical schools and residency programs are figuring out whether and how to incorporate it into their curricula. New centers and new medical journals have been launched with this as a central theme.1/2
But there’s also been a backlash against the idea. Many doctors argue that we’ve always practiced evidence-based medicine, that medicine has always been based on science. But “science” has often meant inferences drawn from what we know of human physiology and disease behavior, information from studies of animals, or expert opinions. It’s sometimes based on the observation that “I tried treatment X and my patients got better.” Even lawyers recognize the fallacy in this line of thinking, embodied in the phrase “post hoc, ergo propter hoc.” Just because patients get better, doesn’t mean that they are getting better because of the treatment applied. As we saw earlier, they may have improved because of natural healing, regression of symptoms to their mean, or placebo effects. It’s easy for both doctors and patients to be fooled about the real benefits of treatment just because someone gets better after the treatment is given.
The new model argues that these older methods are often a good basis for designing and testing a treatment, but they aren’t enough. They’re necessary, but not sufficient. The biological inferences, animal models, opinions, and clinical impressions must generally be subjected to the purifying heat of randomized trials to reveal the true benefits and risks of a new treatment.
Any hope that doctors are already consistent about practicing evidence based medicine is dashed by the repeated observation of huge variations in practice from doctor to doctor, hospital to hospital, and place to place. Unless all treatments and clinical strategies are equally effective, the different approaches can’t all be right. If we were all practicing evidence-based medicine, the litany of ineffective treatments that began this chapter wouldn’t have been used for years or decades. And even with the best of intentions, it’s almost impossible for any doctor to keep up with all the latest scientific advances in any specialty.
Other doctors complain that evidence-based medicine amounts to “cookbook” medicine. Some complain that it’s “a dangerous innovation, perpetrated by the arrogant to serve cost cutters and suppress clinical freedom. They argue that it denies the value of clinical judgment, specialized expertise, and patients’ unique personal preferences.
But advocates of evidence-based medicine argue just the opposite. They argue that clinical judgment is critical in deciding how to apply evidence that comes from other people to each individual patient. Without that ability to generalize, we’d be lost. Advocates also argue that doctors must always consider patients’ preferences in making clinical decisions. Rather than slavish adherence to top-down dictates, evidence-based practice integrates the best evidence, clinical judgment, and patient preferences in every clinical situation. The hard part for most doctors is finding, evaluating, and interpreting the best evidence.
As for costs, evidence-based practice may in some cases decrease costs, but in other cases it can increase costs. Eliminating useless operations or choosing less expensive and more effective drugs will decrease costs. On the other hand, we know that many types of effective screening and treatment are underused. If we start doing all the screening for colon cancer that’s recommended, or treating every diabetic with the full cocktail of drugs and monitoring that are recommended, costs are likely to increase. The approach isn’t inherently cost saving, though in some situations it may be.
Requiring evidence of the sort advocated here places a burden of proof on the advocates of a new treatment. In areas of medicine where there isn’t a long history of rigorous research design or where research funding is scarce, these criteria are often unwelcome and are seen as an obstacle to innovation. Some opponents of evidence-based principles argue that new treatments should be adopted and paid for unless they’re proven ineffective, reversing the burden of proof. As seen already, both safety and cost considerations make this position untenable. Too many people have been hurt by blind faith that new treatments have to be better.
Leave it to the Aussies to recognize pomposity and pretense, which are sometimes associated with the resistance to evidence-based medicine. With tongue in cheek, two Australian doctors summarized some of the alternatives to evidence-based medicine,3 all of which still operate widely:
Eminence-based medicine: Grey hair and years of experience are the basis for recommendations. The faith of such physicians in their clinical experience has been defined as “making the same mistakes with increasing confidence over an impressive number of years.”
Vehemence-based medicine: The louder and more strident the advice, the easier it may be to browbeat colleagues.
Eloquence-based medicine: Here, a fine suit and verbal eloquence supercede evidence.
Providence-based medicine: When the doctor has no idea what to do, decisions can be left in the hands of the Almighty.
Diffidence-based medicine: Doing nothing from a sense of despair. Of course, just standing there may often be better than just doing something, but it’s unsatisfying.
Nervousness-based medicine: Fear of lawsuits drives excessive testing and overtreatment. “The only bad test is the one you didn’t think of ordering.”
Confidence-based medicine: Especially common among surgeons, of whom it’s sometimes said, “they’re often wrong, but never in doubt.”
What Is the Evidence?
Even for doctors, hospitals, and insurance carriers who buy the idea, evidence-based medicine is surprisingly hard to practice. Simply knowing the evidence is hard, because of the enormous size of the medical literature and the rapid pace of its change. Busy doctors don’t have the time to do even electronic searches of all the relevant literature, identify the most useful articles, read them, evaluate the quality of the results, and integrate the pieces into a coherent whole every time they see a patient. They’d be lucky if they were able to do this once a month.
Fortunately, doctors can get help with this. A growing number of medical journals focus on high-quality, clinically relevant articles and provide quick summaries. There also are a growing number of clinical guidelines, which at their best provide a doctor with summaries of the literature that have been extensively researched and evaluated, to help guide the doctor’s decisions, but not dictate them.
One hazard is that guidelines are sometimes promulgated by organizations with a self-interest in the recommendations or with commercial influences lurking in the near background. For this reason, doctors can’t accept guidelines uncritically. The other major hazard is that guidelines can be cumbersome and time-consuming to construct, so they risk being outdated before they’re released or shortly thereafter. Nonetheless, they’re generally far more current than what doctors learned in medical school, and they’re often both more rigorous and more current than medical textbooks.
Having found the relevant medical research or guidelines, evaluating their quality is the next challenge most doctors face. This requires a strong knowledge of research design and statistical methods, which are rarely taught well in medical schools. If they’re included at all, they usually are treated as side dishes to the main courses of biochemistry, pathology, microbiology, and the like. Although the latter remain integral to training effective doctors, the principles of evidence-based medicine need to be taught with equal rigor, enthusiasm, and high expectations.
As Dr. Drummond Rennie, deputy editor of the Journal of the American Medical Association, remarked, “If you are not trained to think statistically, if you are not trained to think epidemiologically, it’s extremely difficult to grasp `risk’ and a whole lot of other things like that. You have to have years of training before you really get into the habit of thinking of risks rationally.”
In other words, doctors need somewhat different training if they are to distinguish high-quality from low-quality evidence for clinical interventions. Incorporating principles of evidence-based medicine into the medical school curriculum may help the next generation of physicians to become more critical consumers of medical research. It needs a higher priority than it often receives.
Rennie also notes that the practice of evidence-based medicine will require constant updating and questioning throughout a doctor’s career. He has prepared journal articles, a book, a CD, and a web version of guides for doctors, because he believes “it’s the only way ahead.”
The Cochrane Collaboration database is another underutilized resource for doctors, patients, and insurance companies. Its organizers named the effort for Archie Cochrane, a British physician and researcher who died in 1988, who argued that society could afford to pay for all effective treatments, if we would just stop paying for the ineffective ones. The Cochrane Collaboration is an international effort to systematically review the medical literature on all treatments, evaluate the quality of individual studies, summarize the results, and pool data from the best randomized trials. The program leaders also intend to update the reviews on a regular basis. All the results are available online, and the number of topics reviewed (already in the thousands) is rapidly growing.4/5 One advocate has argued that the Cochrane Collaboration will prove to be as pivotal in modern medicine as the human genome project.6
Outside the United States, the Cochrane Collaboration has acquired enormous influence among doctors, scientists, and those who pay for medical care. This is true in Canada, Europe, Australia, and most developed countries. In the United States, though, it’s still only modestly familiar and modestly used, perhaps because commercial, political, and media influences often outweigh scientific ones. If we decide that science should play a bigger role, there are resources to help.
Computers in Medicine
Another source of help in practicing evidence-based medicine, one that’s increasingly available to doctors and hospitals, is computer technology. Medical informatics is the term that is applied not only to the use of electronic medical records, but also to so-called decision support programs that can be integrated with electronic medical records. Many doctors recognize the difficulty they have keeping up with the literature, and would welcome authoritative, evidence-based information just at the moment they’re making decisions. The computer makes this increasingly possible.
Suppose the doctor has just admitted an elderly patient from a nursing home into the hospital and has diagnosed pneumonia. She doesn’t yet know exactly which bacteria (or other microbes) are causing the pneumonia, and in many cases it never becomes clear. But she has to choose an antibiotic and a dose based on what infections are commonly acquired in nursing homes, what the local patterns of antibiotic resistance may be, how well the patient’s kidneys are functioning, what drugs the patient is allergic to, how each antibiotic may interact with the other six medicines the patient is taking, the patient’s age and weight, what spectrum of bacteria each drug will kill, the common side effects of each antibiotic, and the particular clinical pattern of this patient’s illness.
In most cases today, she has neither the time nor the knowledge to look up each of these factors and somehow integrate them into a precise solution. Instead, she’ll choose an antibiotic based mostly on habit, local convention, or what she recently heard from a drug rep she likes. She’ll prescribe a standard dose that she uses for most adults. Because there’s a good margin for error in many medical judgments, the patient will usually do just fine-but not always.
Wouldn’t it be nice if this doctor could simply enter her interview and examination findings into the computer and have the computer automatically look up the most common kinds of pneumonia among patients recently admitted to this hospital from nursing homes; consider the antibiotic resistance patterns of bacteria tested by the lab in this hospital over the last six months; check the lab results for the patient’s kidney function; check the nursing notes for the patient’s height and weight; check the pharmacy records to identify the patient’s other drugs; run an automated check for serious interactions with any of those drugs; check the pharmacy, previous hospital records, nursing notes, and doctors’ notes for any record of drug allergies; and make a recommendation for an antibiotic and a dosage based on all these factors plus the best research evidence, embodied in a current clinical guideline. The computer might even list two or three alternatives, with the pluses and minuses of each. The doctor is then free to choose one of the recommendations or to ignore them, perhaps because of other factors that were not considered by the computer.
Even with this scenario, there are no guarantees the patient will improve, but the chances are probably better. The doctor’s and patient’s autonomy haven’t been usurped. The doctor isn’t practicing cookbook medicine; she’s had to gather the patient’s symptoms and exam findings, make the right diagnosis, decide on hospitalization, and make dozens of other decisions. But she’s grateful for the more complete assessment of antibiotic choices she’s gotten. It’s saved her time, and she’s been able to use that time to talk with the family and reassure the patient. The recommendations aren’t based on commercial interests, personal biases, or local habits.
This scenario isn’t quite reality today, but it may be in the near future. A few hospital systems have decision support programs with several of these elements already. Early indications are that these systems reduce errors in the hospital and can save money at the same time; their potential in other aspects of medicine is enormous.
So the computer is one technology that we think has been underused in medicine. Given the huge amount of information that we now gather on every patient, the complexity of decision making, and the volume and volatility of medical knowledge, it may be our only hope for effectively using the technological progress available to us. We think that the investment in information technology by American medicine should increase substantially, and the Bush administration is moving to support such initiatives.
As with the construction of guidelines, one caution is that these decision aids must be developed in an evidence-based fashion by someone who doesn’t have a financial interest in the clinical recommendations. Also, the decision aids will have to be continuously updated. These tasks, like the hardware, will be expensive, but they may prove to pay for themselves. Ultimately, the real benefits and risks of information technology will need to be evaluated as carefully as those of other medical technology.
Then there are the commercial pressures of various types that have become a major obstacle to practicing evidence-based medicine.
Dr. Rennie believes that practicing evidence-based medicine is against human nature:
It’s against human nature to make decisions based on evidence, and I really believe that-even though no one could be more wedded than I am to the idea. Doctors form cozy relationships with nice people who respect them, or seem to respect them, and give them presents. It’s why those doctors so willingly prescribe the latest brand name drug without ever looking into the evidence whatsoever … because it makes them happy, it makes the patient happy, it makes the drug rep happy.
He goes on to describe the pressures created by advertising aimed directly at patients:
I think a doctor in a sense would be foolhardy or reckless with his own time and energy to spend time trying to go against the directo-consumer advertising. “ I will lose this patient if I don’t prescribe what she or he is demanding right now.”
Can doctors and hospitals resist?
Medicine is one of the three traditional “learned professions”: law, religion, and medicine. These professions acquired this stature because they worked with ubiquitous, profound human needs. Professionals in these fields had to be trusted with the intimate details of a client’s mind, body, and personal life. A professional had to be trusted to use such intimate knowledge for the benefit of the client, not for exploitation. Self-governance, service, quality, autonomy, altruism, self-sacrifice, and a high level of learning are seen as defining characteristics of learned professions. These traits are the basis for the trust most patients still have in their own doctors.7
On the other hand, of course, medicine is a business. It’s the way doctors, nurses, administrators, technicians, maintenance experts, manufacturers, and salespeople make a living. Without income, doctors and hospitals couldn’t stay in business, nor could the drug and device makers. Business standards are different from professional standards, and they mainly focus on return on investment. Medical practice is always a balancing act between the professional and the business sides of practice.
Many leaders of American medicine, including prominent journal editors and leaders of professional societies, have become concerned that the balance between business and professional values has recently tilted dangerously toward the business side. As we have discussed, both doctors and hospitals today often seem to make choices based on a corporate mentality that focuses on financial return rather than on good evidence of benefit to patients. When the choices involve new treatments, the assumption is almost always that more and newer can only be letter. Conveniently, this stance almost always coincides with financial self-interest.
Several professional organizations, including the American College of Physicians and the American Board of Internal Medicine Foundation, have issued recent calls for strengthening professional values among physicians 8/9 Though the initiatives are short on specifics or sanctions, they may serve to refocus attention on this problem and to enhance the inculcation of professional values in medical school and specialty training. In some cases, it seems, we’ve become so preoccupied with technical training that we’ve forgotten to convey the ethical and social values on which the profession is built. Perhaps a combination of better training in evidence based medicine and reinvigorated attention to true professionalism will help to counter the commercial pull against evidence-based practice.
Finally, more support for primary care medical practice may help deflect some of the pressures to use unproven treatments. Unlike the cardiologist or the orthopedist, primary physicians have little to gain from using new procedures or gadgets (although they may be susceptible to blandishments from the drug companies). They can therefore afford to be more dispassionate about them. The primary care doctor-typically a family doctor, general internist, or pediatrician-can serve the role of providing unbiased information and assisting patients with many difficult medical decisions.
1. Stolberg SG. On Medicine’s frontier: The last journey of James Quinn. The New York Times, October 8, 2002, p. Dl.
2. How a new policy led to seven deadly drugs; Medicine: Once a wary watchdog, the U.S. Food and Drug Administration set out to become a “partner” of the pharmaceutical industry; today, the American public has more remedies, but some are proving lethal. Los Angeles Times, December 20, 2000, p. Al.
3. Writing Group for the Women’s Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women. JAMA 2002; 288: 321-333.
4. Moseley JB, O’Malley K, Petersen NJ, Menke TJ, Brody BA, Kuykendall DH, et al. A controlled trial of arthroscopic surgery for osteoarthritis of the knee. N Engl J Med 2002; 347: 81-88.
5. Bartley N. Cancer mistake ravages a life. Seattle Times, April 25, 2001, p. B1.
6. Stadtmauer EA, O’Neill A, Goldstein LJ, Crilley PA, Mangan KF, Ingle JN, et al. Conventional-dose chemotherapy compared with high-dose chemotherapy plus autologous hematopoietic stem-cell transplantation for metastatic breast cancer. Philadelphia Bone Marrow Transplant Group. N Engl J Med 2000; 342: 1069-1076.
7. Mello MM, Brennan TA. The controversy over high-dose chemotherapy with autologous bone marrow transplant for breast cancer. Health Affairs 2001; 20(5):101-117.
8. The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic: The Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial (ALLHAT). JAMA 2002; 288: 2981-2997.
9. Petersen M. Diuretics’ value drowned out by trumpeting of newer drugs. The New York Times, December 18, 2002, p. A32.
See also: Kleinke JD. Oxymorons: The Myth of a U.S. Health Care System. San Francisco: Jossey-Bass, 2001, p. 79.
Callahan D. False Hopes: Why America’s Quest for Perfect Health Is a Recipe for Failure. New York: Simon and Schuster, 1998. Kolata G. New heart studies question the value of opening arteries. The New York Times, March 21, 2004, sec. 1, p. 1. Steinberg EP, Tunis S, Shapiro D. Insurance coverage for experimental technologies. Health Affairs 1995; 14: 143-158.
The Backlash Against Evidence-Based Medicine
1. Brown D. First, do the trials. Then, do no harm. The Washington Post, August 4, 2002, p. B1.
2. Sackett DL, Rosenberg WMC, Gray JAM, Haynes RB, Richardson WS. Evidencebased medicine: What it is and what it isn’t. Br Med J 1996; 312: 71-72.
3. Isaacs D, Fitzgerald D. Seven alternatives to evidence based medicine. Br Med J 1999; 319: 1618.
4. Chalmers 1. The Cochrane collaboration: Preparing, maintaining, and disseminating systematic reviews of the effects of health care. Ann NY Acad Sci 1993; 703: 156-163.
5. Chalmers I, Haynes B. Reporting, updating, and correcting systematic reviews of the effects of health care. Br Med J 1994; 309: 862-865,
6. Naylor CD. Grey zones of clinical practice: Some limits to evidence-based medicine. Lancet 1995; 345: 840-842.
7. Lundberg GD. Severed Trust: Why American Medicine Hasn’t Been Fixed. New York: Basic Books, 2000, pp. 162-183.
8. Barondess JA. Medicine and professionalism. Arch Intern Med 2003; 163: 145-149.
9. Brennan T, Blank L, Cohen J, et al. for the Medical Professionalism Project. Medical professionalism in the new millennium: A physician charter. Ann Intern Med 2002; 136: 243-246.
10. Delbanco T, Sands DZ. Electrons in flight-e-mail between doctors and patients. N Engl J Med 2004; 350: 1705-1708.