How Good Are New Drugs?

HOW     DO     WE     KNOW     PRESCRIPTION drugs are any good? You might answer that doc­tors wouldn't use them if they weren't. Doctors, you might say, know what works by experience, and so do their patients. But experience can be highly misleading. The assumption that a drug works if a patient gets better does not allow for natural variations in the illness, for the placebo effect (the tendency of both doctors and patients to imagine a drug is working), for all the other times when the drug might fail, or for the possi­bility that another drug might have worked bet­ter. That is why clinical trials are required by the Food and Drug Administration (FDA). Only by testing drugs in large numbers of people under rigorously controlled conditions can we really know what works and how well it works.

Okay, you might say, I'll buy that. Still, we know drugs work because otherwise the FDA would not approve them. After all, drug companies can't bring new drugs to market until they have carried out clinical trials to show they're safe and effective. But that raises another problem. Can we believe those trials? After all, that crucial last stage of research and development is usually sponsored by the company that makes the drug, even if the early research was done elsewhere. Is there some way companies can rig clinical trials to make their drugs look better than they are? Unfortunately, the answer is yes. Trials can be rigged in a dozen ways, and it happens all the time.

A Wake-Up Call

Let's start by looking at one of the few recent drug trials not sponsored by a drug company. Called ALLHAT (short for Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial), this was a mammoth trial of the treatment of high blood pressure (hypertension).¹ Although it received some support from Pfizer, it was mainly supported and orga­nized by the National Heart, Lung, and Blood Institute—a part of the National Institutes of Health (NIH). The ALLHAT study was eight years long and involved over 42,000 people at more than six hundred clinics, the largest clinical trial of the treatment of high blood pressure ever done. It compared four types of drugs: (1) a calcium channel blocker—sold by Pfizer as Norvasc, the fifth-best-selling drug in the world in 2002;² (2) an alpha-adrenergic blocker—sold by Pfizer as Cardura, and also sold generically as doxazosin; (3) an angiotensin-converting-enzyme (ACE) inhibitor—sold by AstraZeneca as Zestril and by Merck as Prinivil, and also sold generically as lisinopril; and (4) a generic diuretic ("water pill") of a type that has been on the market for over fifty years.

The results, reported in 2002 in The Journal of the Ameri­can Medical Association, were startling. To nearly everyone's surprise, the old-time diuretic turned out to be just as good for lowering blood pressure, and actually better for preventing some of the devastating complications of high blood pressure— mainly heart disease and strokes. Participants treated with the diuretic were much less likely to develop heart failure than those treated with Norvasc. And they were less likely to develop heart failure, strokes, and a number of other complications than those treated with the ACE inhibitor. As for Cardura, that part of the trial had to be stopped early, because so many people who re­ceived that drug developed-heart failure. The director of the Na­tional Heart, Lung, and Blood Institute was unequivocal in his conclusion: "ALLHAT shows that diuretics are the best choice to treat hypertension, both medically and economically."³

Yet over the years the newer drugs had largely supplanted diuretics as treatment for high blood pressure. Diuretics were not promoted because generic manufacturers don't usually spend money on marketing. In contrast, when the new drugs came on the market, they were promoted incessantly. In 1996, for example, Norvasc was the most heavily advertised drug in The New England Journal of Medicine, but there was not a single ad for diuretics.⁴ As you might guess, the use of diuret­ics plummeted. Whereas in 1982 they accounted for 56 per­cent of prescriptions written for high blood pressure, ten years later, after ACE inhibitors and calcium channel blockers hit the market, they accounted for only 27 percent. In general, the newer the drug, the better it sells. If you look at the top fifty drugs used by senior citizens in 2001, Norvasc was number two. Three brand-name ACE inhibitors were also in the top fifty. But diuretics like the one that proved superior in ALL-HAT appeared nowhere on that list.⁵

And look at the costs. Diuretics were priced at about $37 a year in 2002 (among the cheapest drugs on the market), com­pared with $715 for Norvasc and $230 for a generic ACE in­hibitor.⁶ So people with high blood pressure who use Norvasc get to pay nineteen times more for the privilege of taking a drug that is no better, and probably worse, than a diuretic. The cost in health may be worse. High blood pressure is extremely com­mon—about 24 million Americans are now being treated for it. If the ALLHAT findings are correct, a great many people may have suffered serious complications that could have been avoided if they had been treated with diuretics. As Dr. Curt Furberg, the principal author of ALLHAT, put it, "We find out now that we've wasted a lot of money. In addition, [the current practice] has probably caused harm to patients."⁷

Why didn't we learn a long time ago that the new drugs were not as good as the old ones? Well, to begin with, no one tried to find out. The last thing drug companies want is a head-to-head comparison with older drugs. So the new drugs were approved mainly because—in accord with the minimal FDA re­quirements—they were shown to be better than placebos. With few exceptions, no one wanted to know how they compared with a diuretic—or with one another, for that matter. New drugs came to market because they were better than nothing, then were promoted as though they were great advances in med­icine. Since ALLHAT, industry apologists have protested that most people with high blood pressure need more than one drug, so new agents are indeed important. That is true but disingenu­ous. Drug companies tested and promoted their blood pressure drugs not as supplements but as first-line treatments.

Studies like ALLHAT are rare. The NIH does not ordinarily conduct clinical trials of drugs. It focuses on basic research into the underlying mechanisms of disease and leaves drug testing to the companies that manufacture the drugs. But sometimes it makes exceptions. The ALLHAT study was begun in 1994 be­cause people were increasingly unhappy with the fact that no one knew which of the one hundred blood pressure drugs be­longing to seven drug classes was best for the first-line treatment of hypertension. Of course, no single study can be considered the last word on this question. In fact, a subsequent, smaller Australian study comparing Merck's Prinivil with a diuretic in­dicated that Prinivil was slightly better.⁸ But the ALLHAT study was a wake-up call. It raised the possibility that the "miracles" big pharma is so proud of really aren't miracles at all.  [PhRMA, Pharmaceutical Research and Manufacturers of America, the principle industry defense organization.  Involve in public relations organization, lobbying, litigation, and like--jk]. Maybe many new drugs are actually worse than old ones. Unless we test the new against the old, head to head, there is simply no way to know.

The Research Enterprise

How do doctors decide which drugs to use for their pa­tients? Some of them, unfortunately, depend on drug company marketing—the subject of the next chapter. But most doctors depend, at least in part, on a number of supposedly unbiased sources of information. They read medical journals to find out about new research and how to interpret it; they use textbooks to see what conclusions the expert authors draw from the over­all body of scientific evidence; and they go to meetings and con­tinuing medical education courses to hear from these experts (called "thought leaders") firsthand. These last two sources are, in fact, derivatives of the first. Textbooks and the thoughts of thought leaders are no better than the evidence on which they are based. And that evidence comes from research reports in medical journals. So it is crucial that those reports be unbiased. Are they?

Increasingly, the answer is no. As I have said, most clinical research on drugs is sponsored by the companies that make them. By itself, industry sponsorship does not mean the research is biased. But in addition, drug companies now have consider­able control over the way the research is carried out and re­ported. That is new. Until the 1980s, researchers were largely independent of the companies that sponsored their work. Drug companies would give a grant to an academic medical center, then step back and wait for faculty researchers to produce the results. They hoped their product would look good, but they had no way of knowing for sure. They certainly did not attempt to tell the researchers how to run their clinical trials. Now, how­ever, companies are involved in every detail of the research— from design of the study through analysis of the data to the decision whether to publish the results. That involvement has made bias not only possible but extremely likely. Researchers don't control clinical trials anymore; sponsors do. 

Why the change? It comes down to the immense growth in industry wealth and influence since the watershed year of 1980. As they became richer, more powerful, and more profit-driven, drug companies became less willing to sit back and wait for academic researchers to produce their results. For one thing, the trials ate into the patent life of drugs, and for another, just wait­ing was too uncertain. The research findings could go against them. So instead of relying on academic centers to test their drugs, drug companies turned to the new for-profit research in­dustry that grew up to serve them—the contract research orga­nizations I described in Chapter 2. As you will remember, these companies contract with private doctors to collect data on pa­tients in their offices according to company instructions. The doctors are not themselves trained researchers, so they simply do what they are told—or risk losing their lucrative deals with the contractors. The contract research organizations, in turn, answer only to big pharma. That means the drug companies have nearly total control over these trials.

The academic medical centers were unhappy to lose drug company contracts—even though they were a small fraction of their research income. Whereas in 1990, about 80 percent of industry-sponsored trials were done at academic institutions, within a decade that share had dropped to less than 40 percent. The loss came at a time when many medical schools and teach­ing hospitals were in financial trouble because of shrinking reim­bursements from patient care and reduced support for medical education. So they began to compete with contract research or­ganizations the only way they could—by becoming more accom­modating to pharmaceutical sponsors. When companies insisted on controlling the way clinical trials were done, they met re­markably little resistance.

Furthermore, the whole context of academic-industry rela­tions had changed. With the 1980 Bayh-Dole legislation, the traditional boundaries between academia and industry were blurred. Academic medical centers now saw themselves as "partners" of the pharmaceutical industry in common endeav­ors—and junior partners, at that. Consider some of the aca­demic-industrial arrangements at Harvard University.⁹ The Dana-Farber Cancer Institute, a Harvard hospital, has a deal that gives Novartis rights to discoveries that lead to new cancer drugs. The Japanese cosmetic maker Shiseido gave Harvard's Massachusetts General Hospital $180 million over ten years for first rights to discoveries by faculty dermatologists. Merck is building a twelve-story research facility next door to Harvard Medical School. Both sides expect a close, multifaceted collabo­ration, although the terms have not yet been announced. Part­ners HealthCare, a pair of Harvard teaching hospitals, issued a request for faculty applications to be part of a "Partners Faculty Exchange Program" with Millennium Pharmaceuticals. It promised that, as part of Partners' collaboration with Millen­nium, "interested faculty will be integrated into Millennium project teams." Harvard is not unique. One recent survey showed that two-thirds of academic medical centers hold equity in start-up companies that sponsor some of their research.¹⁰ The drug companies, for their part, are generous to the medical schools. In Harvard Medical School's Dean's Report for 2003-4, for instance, the list of benefactors included about a dozen of the largest drug companies. So it was just part of the new climate to do clinical .trials pretty much the way the phar­maceutical industry wanted them done.

As a result, drug companies now design clinical trials to be carried out by researchers who are little more than hired hands— whether the trials are in academic centers or in physicians' offices. Sponsoring companies keep the data, and in multicenter trials, they may not even let the researchers themselves see all of it. They also analyze and interpret the results, and decide what, if any­thing, should be published. The authors of a recent survey of aca­demic policies on the matter concluded, "We found that academic institutions rarely ensure that their investigators have full partici­pation in the design of the trials, unimpeded access to trial data, and the right to publish their findings."¹¹ All of this makes a mockery of the traditional role of researchers as independent and impartial scientists. Academic institutions and their faculties vary in how much control they are willing to cede to sponsors, but in general, they cede much more than they should. Contract re­search organizations and their networks of private doctors cede most of it.

If faculty researchers have lost much of their independence, they've gained in other ways. Many of them have lucrative fi­nancial arrangements with drug company sponsors that would have been impossible twenty years ago. Researchers serve as consultants to companies whose products they are studying, be­come paid members of advisory boards and speakers' bureaus, enter into patent and royalty arrangements together with their institutions, promote drugs and devices at company-sponsored symposiums, and allow themselves to be plied with expensive gifts and trips to luxurious settings. Many also have equity in­terest in the companies. These kinds of deals can add signifi­cantly to their salaries. The head of the Department of Psychiatry at Brown University Medical School, for instance, was reported by The Boston Globe to have made over $500,000 in consulting fees in 1998.¹² It's hard to believe that close and remunerative personal ties with drug companies do not add to the strong pro-industry bias in medical research and education. Big pharma not only controls the details of the way clinical trials are performed, but as backup, it also seeks to win the hearts and minds of researchers.

One of the more sobering indications of the extent to which big pharma has compromised the research community is its ex­tensive inroads into the NIH itself. This agency funds the lion's share of basic medical research throughout the country, using taxpayer dollars. It is supposed to award grants entirely on the basis of scientific merit, and conduct its own research, including the selection of industry collaborators, entirely in the public in­terest—free from commercial considerations. But a 2003 piece of investigative reporting by David Willman in the Los Angeles Times called that picture into serious question.¹³ Willman found that senior NIH scientists (who are among the highest paid em­ployees in the government) routinely supplement their income by accepting large consulting fees and stock options from drug com­panies that have dealings with the institutes. At one time, most of these kinds of connections would have been prohibited, but in 1995, the then director of the institutes, Harold Varmus, with a stroke of the pen, lifted the restrictions. After that, the NIH placed no limits on the amount of money its scientists could earn from outside work or the time they could devote to it.

According to Willman, senior scientists with financial ties to industry included the director of the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the director of the NIH Clinical Center (the main site for human subjects re­search), the former director of the division of diabetes, en­docrine, and metabolic diseases at the National Institute of Diabetes and Digestive and Kidney Diseases, and the former di­rector of the National Human Genome Research Institute. Some NIH scientists made hundreds of thousands of dollars in consulting fees. The deputy director of the Laboratory of Im­munology, for instance, whose salary was $179,000 in 2003, was reported to have collected more than $1.4 million in con­sulting fees over eleven years and received stock options valued at $865,000.

It is impossible to know to what extent these financial deals influenced NIH judgments about grants, research priorities, or the interpretation of results, but they certainly are cause for con­cern. Outside activities were said to be approved by supervisors, and scientists supposedly excused themselves from direct in­volvement in decisions that affected their outside clients, but Willman reported instances in which even those minimal re­strictions seem to have been ignored. Moreover, the NIH did not even require most senior scientists to file public disclosures of their outside income. (That was accomplished by jiggering the pay scale in such a way that highly paid scientists were in­cluded with those with lower salaries, who were not required to disclose conflicts of interest.) The result was that, as of 2003, more than 94 percent of the agency's 2259 top scientists did not have to reveal their outside consulting income.

In an editorial in the Los Angeles Times that accompanied the Willman revelations, the paper got it exactly right: "The pharmaceutical industry is everywhere in Washington, all but writing the Medicare prescription drug bill, fielding more lob­byists than there are members of Congress, flinging gifts and trips at doctors and trying to prevent double-blind drug trials that pit one drug against another, instead of against a placebo." It concluded, "Willman's story, shocking as it is, is just one piece of an unwholesome picture. Congress helped make this system and can help unmake it. Start with high-level hearings. Repeal the most destructive portions of the Bayh-Dole Act. Above all, restore the integrity of the National Institutes of Health."¹⁴ In January 2004, the Senate Appropriations Subcommittee on Labor, Health and Human Services, and Education had begun to hold hearings on the matter, and the Department of Health and Human Services inspector general and the U.S. General Ac­counting Office launched their own inquiries. As might be an­ticipated under these circumstances, the NIH director appointed a Blue Ribbon Panel on Conflict of Interest Policies.

Bias—And Lots of It

Not surprisingly, bias is now rampant in drug trials.¹⁵ A re­cent survey found that industry-sponsored research was nearly four times as likely to be favorable to the company's product as NIH-sponsored research (despite the Willman revelations).¹⁶ That is in accord with a large body of evidence showing that researchers with industry connections are far more likely to favor company products. In the case of calcium channel block-ers like Norvasc, for instance, one survey of seventy articles about their safety found that 96 percent of authors who were supportive of the drugs had financial ties to the companies that made them, whereas only 37 percent of authors who were crit­ical had such ties.¹⁷

I will not detail all the ways research can be biased.¹⁸ But a few are worth mentioning. Sometimes, bias is just a matter of spin—researchers extol a drug even though the results do not support their enthusiasm. One recent survey showed that au­thors of industry-funded studies were more than five times as likely to recommend the company drug as authors of studies funded by nonprofit organizations—regardless of the actual re­sults.¹⁹ But often, bias is built into the study design, as is the case with placebo-controlled clinical trials. Almost inevitably new me-too drugs seem to be effective. But in fact, as shown by ALLHAT, if they were compared with drugs already on the mar­ket, they might be revealed to be less effective. Researchers even in the most prestigious medical centers go along with loading the dice this way, because sponsors insist on it. The information from such trials may be of very little use to practicing doctors, who usually are not interested in whether a new drug is better than nothing. They want to know whether it is better than what they are already using.

Another way to load the dice is to enroll only young subjects in trials, even if the drugs being tested are meant to be used mainly in older people. Because young people generally experience fewer side effects, drugs will look safer in these trials than they would in practice. Another is to compare the new drug not just with a placebo but with an old drug given at too low a dose. In the last chapter, I described how this was done with the statins. This bias was also present in many trials of nonsteroidal anti-inflammatory drugs (NSAIDS). (These are drugs like Naprosyn that are taken mainly for arthritis.) The new NSAIDS looked better because the comparison drugs were given at lower doses. Or the old drug can be administered incorrectly. That was true of trials comparing fluconazole with an older drug, amphotericin B, to treat fungus infections in patients with AIDS. Amphotericin B was administered orally, which dramati­cally cuts down on its effectiveness. Not surprisingly, the trials were sponsored by the makers of fluconazole. Or trials can be designed to be too brief to be meaningful. That is true of many trials of drugs that need to be taken long term. Blood pressure trials usually last for just a few months, and antidepressant tri­als for a few weeks, yet patients may have to take these kinds of drugs for years. Some treatments look pretty good for a short time but are not effective and may even be harmful for long-term use.

One of the most common ways to bias trials is to present only part of the data—the part that makes the product look good—and ignore the rest. That happened in a clinical trial of the arthritis drug Celebrex. A study sponsored by the drug com­pany that makes it, Pharmacia (since acquired by Pfizer), ostensibly showed that Celebrex caused fewer side effects than two older arthritis drugs. These results were published, along with a favorable editorial, in The Journal of the American Medical As­sociation. Not until after publication did the editors learn that the results were based on only the first six months of a yearlong trial. When the entire trial was analyzed, there was no advan­tage to Celebrex. The editorialist was understandably outraged. The Washington Post quoted him as saying, "I am furious. . . . I wrote the editorial. I looked like a fool. But... all I had avail­able to me was the data presented in the article." And the editor of the journal said, "I am disheartened to hear that they had those data [the second six months] at the time that they submit­ted [the manuscript] to us. We are functioning on a level of trust that was, perhaps, broken."²⁰

Suppressing Things You Don't Like

The most dramatic form of bias is out-and-out suppression of negative results. That is easily done in privately run trials, but it also occurs in trials done at academic centers. There have been several widely publicized cases. It's instructive to look at one of them.²¹ In 1996, a biotechnology company called Immune Re­sponse Corporation contracted with Dr. James O. Kahn of the University of California at San Francisco and Dr. Stephen W. Lagakos of the Harvard School of Public Health to conduct a multicenter trial of its drug Remune. This drug was supposed to slow the progression of AIDS by boosting the immune system, and the company was seeking FDA approval to market it as a "therapeutic vaccine." Kahn and Lagakos undertook the trial, which was carried out in 2500 HIV-infected patients at seventy-seven medical centers. But the company held the data.

After three years, it became clear that Remune was not ef­fective. But the company objected to Kahn and Lagakos pre­senting the results as negative (meaning that the vaccine had no effect). It wanted them to include in their paper an analysis of a subgroup of patients that it said showed positive effects. Kahn and Lagakos refused, saying that the company's analysis was not in accord with scientific standards. Immune Response then threatened to withhold the final 5 to 10 percent of data if the in­vestigators did not agree to include the company's analysis. Fi­nally, after much pulling and tugging, the company agreed to turn over the remaining data, but only on condition that it be given the right to approve the paper. Kahn and Lagakos again refused. On the basis of the data they already had (which was enough), they published a negative report in The Journal of the American Medical Association. Immune Response filed a multi­million-dollar claim against Kahn and his university, alleging harm to its business. (The company eventually lost.)

It is interesting to look behind the scenes in this dispute. The contract between the company and the researchers contained the seeds of the problems that developed. While it did not give Immune Response veto power over publication, it did involve the company in every detail of the work.  It set up a five-person committee that included the company's medical director to write the paper; it stipulated that Kahn would keep the com­pany abreast of the progress of the trial; and it provided for the company to see the final paper before submission for publica­tion. When it became clear that the trial was negative, the com­pany asserted the right to do the analysis. The president and CEO of Immune Response later complained, "Just put yourself in my position. I spent over $30 million. I would think I have certain rights."²² He really seemed to think he had a "right" to favorable results.

Kahn and Lagakos showed courage and integrity in sticking to their guns. It is essential that clinical research be conducted with impartiality, which means sponsors must be kept at arm's length. Many researchers simply accept the will of sponsors or cave in under pressure. But the conditions of the contract were in a sense the camel's nose under the tent. By involving the com­pany in all aspects of the trial, even including the medical direc­tor as a coauthor, Kahn and Lagakos were virtually inviting the camel inside. The company had a clear conflict of interest. Yet by today's standards, the contract gave Kahn and Lagakos an unusual amount of independence. Many current contracts give companies far more control.

So What Do We Really Know?

When a drug company applies to the FDA for approval of a new drug, it is required to submit results from every one of the clinical trials it has sponsored. But it is not required to publish them. The FDA may approve the drug on the basis of minimal evidence. For example, the agency usually requires simply that the drug work better than a placebo in two clinical trials, even if it doesn't in other trials. But companies publish only the pos­itive results, not the negative ones. Often, in fact, they publish positive results more than once, in slightly different forms in dif­ferent journals. The FDA has no control over this selective pub­lishing. The practice leads doctors to believe that drugs are much better than they are, and the public comes to share this be­lief, on the basis of media reports. There is a general inflation in the notion of the good that drugs can do (and a deflation in concern about side effects).

Take the case of antidepressants. Among the top ten drugs in 2002 were two antidepressants of the SSRI type—Zoloft and Paxil. It is generally accepted that SSRIs are highly effective drugs. Millions of Americans take them, and many psychiatrists and primary care doctors swear by them. But a recent study throws doubt on the general enthusiasm. Using the Freedom of Information Act (a law that allows citizens to obtain govern­ment documents), the authors obtained FDA reviews of every placebo-controlled clinical trial submitted for initial approval of the six most widely used antidepressant drugs approved be­tween 1987 and 1999—Prozac, Paxil, Zoloft, Celexa, Serzone, and Effexor (all but the last two are SSRIs).²³ As is typical, most of the forty-two clinical trials lasted for just six weeks.

Their findings were sobering. On average, placebos were 80 percent as effective as the drugs. And the difference between drug and placebo was only about two points on the sixty-two-point Hamilton Depression Scale (a measure of the severity of depression). While that is statistically significant, it is very un­likely to be of any clinical importance. The results were much the same for all six drugs. These figures, of course, are averages, and it is possible that certain types of patients have much better (or worse) responses. But the point is that, based on all the evi­dence, not just what the drug companies published, the new an­tidepressants certainly do not look like the miracle drugs we have been led to believe they are. More recently, there have been serious charges that manufacturers of SSRIs have suppressed data indicating that the drugs are not only ineffective but some­times dangerous in children.²⁴

Another recent NIH study is also instructive. For decades, women have been taking estrogen and progesterone hormone replacement therapy, not just to treat the symptoms of menopause but also in the belief that it will prevent heart dis­ease. That belief was based mainly on industry-sponsored stud­ies. But now a large NIH clinical trial indicates that, instead of preventing heart disease, combination hormone replacement ac­tually increases it. This trial underscores the fact that we need to question how reliable publications from industry-sponsored re­search really are.

I do not want to sound like a nihilist or a Luddite. I know very well that as a result of innovative research and develop­ment—in both academia and industry—we have available to us many drugs of immense importance. No one would want to be without, say, insulin for diabetes, antimicrobial agents to fight infections, vaccines to prevent a host of serious diseases, anti-clotting agents to treat heart attacks, chemotherapy for cancer, a panoply of effective painkillers and anesthetics, and many others. Gleevec is a major advance, as are Epogen and Taxol. Prilosec is important, too, as are statins and ACE inhibitors and many other drugs. All of these agents have extended and greatly improved our lives. I would not have spent my professional life at The New England Journal of Medicine if I did not deeply be­lieve in the value of medical research and innovative treatments.

So I do not mean to suggest that prescription drugs are gen­erally useless or dangerous or some sort of hoax. What I do mean to suggest is that many of them may be, particularly new me-too drugs that have been tested by companies and re­searchers with financial interests in the outcomes. Are new drugs any better than old ones? Are they worse? The worrying answer is that frequently we just don't know. Too often, all we have is bias and hype.

6. How Good Are New Drugs?

1.  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," Journal of the American Medical Association, December 18, 2002, 2981. For detailed media accounts, see Ron Winslow and Scott Hensley, "Study Questions High-Cost Drugs for Hypertension," Wall Street Journal, December 18, 2002, Al; Lawrence K. Altman, "Older Way to Treat Hyperten­sion Found Best," New York Times, December 18, 2002, Al.

2.  The list of the top ten drugs in the world is available from IMS Health, "World Review for 2002," data released February 25, 2003 (www.imshealth.com).

3.  Winslow and Hensley, "Study Questions High-Cost Drugs."

4.  Altman, "Older Way."

5.  FamiliesUSA, "Bitter Pill: The Rising Prices of Prescription Drugs for Older Americans," June 2002 (www.familiesusa.org).

6.  Winslow and Hensley, "Study Questions High-Cost Drugs."

7.  Altman, "Older Way."

8.  Lindon M. H. Wing et al.,  "A Comparison of Outcomes with Angiotensin-Converting-Enzyme Inhibitors and Diuretics for Hyper­tension in the Elderly," New England Journal of Medicine, February 13,2003,583.

9.  I focus on Harvard Medical School and its teaching hospitals only be­cause they are most familiar to me, and I touch on just a few of the arrangements. For media coverage, see Liz Kowalczyk's reports: "Beth Israel Seeks Deal with Drug Company," Boston Globe, Febru­ary 14, 2001, Al; "Harvard to Use Caution with Merck," Boston Globe, August 1, 2001, Al; "Lucrative Licensing Deals with Drug, Biotech Firms Are Raising Ethics Issues  for Hospitals,"  Boston Globe, March 24, 2002, Cl. Also see Raja Mishra, "Harvard May Ease Rules on Faculty Ties to Drug Firms," Boston Globe, June 9, 2003, Al. The Millennium deal was spelled out in a request for ap­plications sent to faculty at Partners.

10.  See Justin E. Bekelman et al., "Scope and Impact of Financial Con­flicts of Interest in Biomedical Research," Journal of the American Medical Association, January 22-29, 2003, 454.

11.  Kevin A. Schulman et al., "A National Survey of Provisions in Clini­cal Trial Agreements Between Medical Schools and Industry Spon­sors," New England Journal of Medicine, October 24, 2002, 1335.

12.  This was reported by Alison Bass, "Drug Companies Enrich Brown Professor," Boston Globe, October 4, 1999, Al; and by Douglas M. Birch and Gary Cohn, "Of Patients and Profits: Standing Up to In­dustry," Baltimore Sun, June 26, 2001, 1A.

13.  David Willman, "Stealth Merger: Drug Companies and Government Medical Research," Los Angeles Times, December 7, 2003, Al.

14.   "Subverting U.S. Health" editorial, Los Angeles Times, December 7, 2003, M4.

15.  For an excellent overview of bias in clinical research, see Thomas Bo-denheimer, "Uneasy Alliance: Clinical Investigators and the Pharma­ceutical Industry," New England Journal of Medicine, May 18, 2000, 1539.

16.  Bekelman et al., "Scope and Impact."

17.  H. T. Stelfox et al., "Conflict of Interest in the Debate over Calcium-Channel Antagonists," New England Journal of Medicine, January 8, 1998, 101.

18.  The British Medical Journal published an excellent issue on industry sponsorship and bias in clinical research. The papers are accessible at the journal's website, www.bmj.org. See in particular Silvio Garattini et al., "How Can Research Ethics Committees Protect Patients Bet­ter?" British Medical Journal, May 31, 2003, 1199; see also Frank van Kolfschooten, "Can You Believe What You Read?" Nature, March 28, 2002, 360.

19.  Bodil Als-Nielsen et al., "Association of Funding and Conclusions in Randomized Drug Trials," Journal of the American Medical Associa­tion, August 20, 2003, 921.

20.  Susan Okie, "Missing Data on Celebrex: Full Study Altered Picture of Drug," Washington Post, August 5, 2001, All.

21.  This case was fully covered in the press. See, for example, Philip J. Hilts, "Company Tried to Bar Report That HIV Vaccine Failed," New York Times, November 1, 2000, A26; Richard Saltus, "AIDS Drug Researchers Say Firm Pressured Them," Boston Globe, No­vember 1, 2000, A3; Thomas M. Burton, "Unfavorable Drug Study Sparks Battle over Publication of Results," Wall Street Journal, No­vember 1, 2000, Bl; Carol Cruzan Morton, "Company, Researchers Battle over Data Access," Science, November 10, 2000, 1063. I also spoke with one of the authors and had access to some of the docu­ments, including the agreement between the authors and the com­pany.

22.  This was reported in Birch and Cohn, "Of Patients and Profits."

23.  This study of Prozac-type antidepressants was unique in that it ana­lyzed all clinical trials of antidepressant drugs from the FDA, whether published or not. See Irving Kirsch and Thomas J. Moore, "The Em­peror's New Drugs: An Analysis of Antidepressant Medication Data Submitted to the U.S. Food and Drug Administration," Prevention & Treatment, July 15, 2002.

24.  See Wayne Kondro and Barbara Sibbald, "Drug Company Experts Advised Staff to Withhold Data About SSRI Use in Children," Cana­dian Medical Association Journal, March 2, 2004; "Depressing Re­search," editorial, The Lancet, April 24, 2004.

From Department of Social Medicine, Harvard University School of Medicine at http://www.hms.harvard.edu/dsm/WorkFiles/html/people/faculty/MarciaAngel.html

Marcia Angell, M. D., is Senior Lecturer in the Department of Social Medicine at Harvard Medical School. She stepped down as Editor-in-Chief of the New England Journal of Medicine on June 30, 2000. A graduate of Boston University School of Medicine, she trained in both internal medicine and anatomic pathology and is a board-certified pathologist. She joined the editorial staff of the New England Journal of Medicine in 1979, became Executive Editor in 1988, and Editor-in-Chief in 1999. 

Dr. Angell writes frequently in professional journals and the popular media on a wide range of topics, particularly medical ethics, health policy, the nature of medical evidence, the interface of medicine and the law, and care at the end of life. Her critically acclaimed book, Science on Trial: The Clash of Medical Evidence and the Law in the Breast Implant Case, was published in June, 1996, by W. W. Norton & Company. In addition, Dr. Angell is co-author, with Dr. Stanley Robbins and, later, Dr. Vinay Kumar, of the first three editions of the textbook, Basic Pathology. She also wrote chapters in several books dealing with ethical issues.

Dr. Angell is a member of the Association of American Physicians, the Institute of Medicine of the National Academy of the Sciences, the Alpha Omega Alpha National Honor Medical Society, and is a Master of the American College of Physicians. In 1997, Time magazine named Marcia Angell one of the 25 most influential Americans.

For more articles on profits and drugs