"Evidence-based medicine" has become the mantra of the medical profession. When patients "present" to their physicians, the physicians are supposed to set aside their intuitions and prejudices and, instead, treat the patients' problems according to the best and most rigorous scientific data available in the leading journals.

Let us suppose that you are one of those patients fortunate enough to have a family doctor who actually keeps abreast of the medical literature, rather than making her prescription decisions based upon what was popular when she graduated from medical school 20 years ago or, worse still, what was left to her as a pile of free samples by that handsome young drug company "rep" who took her to lunch last week at Rae and Jerry's steak house. (If the doctor is male, then substitute pretty young woman for handsome young man in the above example.)

You have been diagnosed with hypertension. That is, you have high blood pressure. Your doctor has consulted the journals and has found a high percentage of articles in the leading journals which favour treating hypertensive disease with drugs known as "calcium-channel antagonists or blockers." You then begin your career as what could turn out to be a lifetime consumer of this very expensive medication.

Is the Evidence on which Your Doctor Is Relying Truly Reliable?

There is a good deal of recent research to suggest that the correct answer to this question is: "Probably not." The principal reason for doubting the scientific reliability of the published studies, even though they have been carried out by eminent physicians and appear in leading journals, is that they have mostly been funded by the drug industry itself.

One of the most influential studies of how researchers' objectivity might be compromised by drug industry sponsorship appeared in The New England Journal of Medicine, in January 1998. Stelfox and colleagues set out to examine published articles on the safety of calcium-channel antagonists. Their goal was to answer the question: To what extent does industry support of medical research influence the research findings of investigators? For purposes of this study, Stelfox divided authors according to their relationships with pharmaceutical companies and then, independently, classified their research findings on the safety issue as "supportive," "critical" or "neutral." The conclusion reached by Stelfox et al must be of serious concern to every supporter of industry-university partnerships: "Our results demonstrate a strong association between authors' published positions on the safety of calcium-channel antagonists and their financial relationships with pharmaceutical manufacturers."¹

It may be worth spelling out just how influential drug company sponsorship appears to have been: "Ninety-six per cent of supportive authors had financial relationships with the manufacturers of calcium-channel antagonists, as compared with 60 per cent of the neutral authors and 37 per cent of the critical authors." A caveat is required here. The Stelfox study needs to be interpreted with care. Were authors first funded by companies making calcium- channel blockers, after which they wrote favourably about the product, or did they first write favourably about the product and only then receive financial support from the companies? Since the Stelfox authors were unable to determine the time line, this question cannot be answered conclusively. Interestingly, even researchers who had financial ties with manufacturers of competing products were significantly less critical of the drugs being tested than authors who had no ties to industry. In other words, scientists who are funded by the pharmaceutical industry produce studies which tend to be more favourable to new drugs than those whose funding is industry-independent even when the new drug being tested is a product of a rival company.

In a more recent study, Canadian doctor Joel Lexchin and colleagues did a comprehensive meta-analysis of what is currently known about the alleged tendency of drug company sponsorship to produce biased research results. They concluded that there is a systematic bias to the outcome of published research funded by the pharmaceutical industry. When research is funded by the pharmaceutical industry, it is significantly more likely to favour the industry's products than when the funding is industry-independent. The results apply across a wide range of disease states, drugs and drug classes, over at least two decades and regardless of the type of research being assessed.

Although it seems intuitively obvious that "he who pays the piper calls the tune," there are competing hypotheses to explain why it is the case that when a pharmaceutical company pays for the clinical trial of its new drug, the results are more likely to be favourable to that drug than when the funding is industry-independent. The most plausible explanation is encapsulated by the phrase "publication bias." This phrase is used to describe a tendency on the part of researchers not to submit their findings for publication when those findings are negative. Publication bias is sometimes referred to, colloquially, as "the file drawer effect." Suppose that there are 20 studies done of some new drug, and suppose that of those 20 studies, six are positive and 14 are negative. Suppose, further, that as a direct or indirect result of company influence, 12 of the negative studies are not published (that is, they are banished to the file drawer), while every positive study is published, celebrated even. Those physicians who then attempt conscientiously to review the literature would find six positive but only two negative studies.

The new drug would be hailed as a medical breakthrough and would rapidly become part of standard therapy. This is not science, however, so much as marketing through censorship or self-censorship. If the much-touted movement towards "evidence-based medicine" is to mean anything, then physicians need unbiased data on the clinical effectiveness, toxicity, convenience and cost of new drugs compared with available alternatives. Because of the phenomenon of publication bias, what passes for good scientific evidence may be simply a mirage.

When a university or hospital researcher's career depends upon attracting and keeping drug industry funding, the desire to please the company (and protect one's career) can easily lead to ethical violations. When the researcher's university is beholden to the same company for huge donations of equipment and buildings, the incentive not to kill the goose that's laying these golden eggs will sometimes be irresistible. Moreover, when researchers discover potentially serious harms being caused by the drugs they are testing on their patients, and when they attempt to inform their patients of such newly-discovered risks, they may find themselves threatened by the company with ruinous legal action, as happened to Dr. Nancy Olivieri at Sick Kids Hospital, in Toronto. Neither her hospital nor her university (University of Toronto) was willing to support her effectively in this struggle. When the public discovered that the company had promised a $20 million donation to the university and more millions to the hospital, those of a cynical bent suspected the defence of academic freedom may have been less than vigorous because of a desire, conscious or unconscious, to please a wealthy donor.

To make matters worse, sometimes the research design itself is loaded in favour of the company's drug. For example, the new drug is compared to a drug already proven to be ineffective rather than to one already proven effective. Or it might be compared to an effective older drug, but the older drug is given in an inappropriate dose, which misleadingly makes the new drug seem better than it is. Moreover, even well-designed studies can be poorly conducted, and biases that favor the sponsor's product can be introduced by protocol violations, failure to keep proper records or failure to submit accurate data to a journal for publication.

Evidence such as that cited above calls into question the integrity of drug company-funded research. Since company-funded research has become preponderant, in both Canada and the United States, the clear implication is that the integrity problem is both systemic and serious.

Why Drug-Industry Funding of University Research Should Be Banned

The evidence now seems very strong, indeed. When medical research is sponsored by drug companies, it tends to be biased in favour of the companies' products. Given the major role of the pharmaceutical industry in funding research, it is not altogether surprising that published drug studies are often of poor quality.

Data showing that a new drug may be dangerous for patients is sometimes "lost" or "reinterpreted" or simply not submitted for publication. Scientists who try to warn patients about potentially serious drug side effects sometimes find that their university or hospital appointment has been terminated. This happened to Dr. David Healy after he issued public warnings about his concern that Prozac was being overprescribed. Healy was not intimidated, but many researchers are understandably concerned to protect their future funding and their jobs.

Since doctors rely on published studies when deciding the best treatment for their patients, anything which threatens the objectivity of medical research is bad news for patients. Partly this is a matter of patient safety, but it is also a matter of economics. New drugs, which are vastly more expensive than older drugs but often no more effective, sometimes achieve enormous popularity based upon flawed research.

Celebrex, for example, became a "blockbuster" drug for the treatment of arthritic pain, largely based upon a now-notorious study appearing in The Journal of the American Medical Association (JAMA) in the summer of 2000. The published data, covering six months, indicated that the new drug was associated with lower rates of stomach and intestinal ulcers than two older drugs. Only after publication was it revealed that the full year's data - which had not been disclosed - did not support the claim that Celebrex was safer than its generic competitors.

Although the company, Pharmacia, continues to claim that Celebrex has a superior safety profile, the U.S. Government Food and Drug Administration's arthritis advisory committee concluded, based upon the full year's data, that Celebrex offers no proven safety advantage in reducing the risk of ulcer complications. JAMA's editor laments: "I am disheartened to hear that they had those data at the time they submitted (the manuscript) to us."

It should be noted that half of the Celebrex study's 16 authors were medical faculty at eight different universities, acting as paid consultants of the company. The other eight were company employees.

In such an environment, can the public trust in the objectivity of university research?

Cases such as this one have generated a powerful reform movement within the medical research community. In an article published recently in The Canadian Medical Association Journal, the acceptance of drug-company funding by university researchers is compared to "dancing with the porcupine." Those who wish to dance with porcupines must exercise great care if they are to avoid painful skin punctures. By analogy, it's okay for scientists to partner with industry, so long as precautions are taken. What kind of precautions?

The general public would likely be amazed to learn that, until very recently, researchers were not required publicly to disclose who was funding their research. Worse, it was - and still is - common for medical researchers to attach their names to scholarly articles that they have neither designed nor written. That is, the experiments are designed by the drug companies and ghostwritten by drug company employees, but appear under the names of prominent doctors (who duly reap status from their colleagues, and cold cash from the company).

Important reforms, adopted in recent years by many universities, have included the requirement that investigators disclose the sources of their funding, that they take responsibility for the design of their experiments and the interpretation of the data, and that they (not the companies) have the final say as to whether to publish.

Disappointingly, however, a recent (2002) survey shows that these reforms are routinely ignored by academic institutions. It seems to be the case that universities and researchers are competing so fiercely for drug-industry funding that the reforms have been sidelined in a race to the ethical bottom.

Is it likely that any reformist approach will succeed in restoring integrity to medical research? Keep in mind that the fundamental duty of universities is "to seek truth." By contrast, the duty of pharmaceutical companies is "to make money for their shareholders." Thus, from the university's perspective, "partnership" with the drug industry may be characterized as "an unholy alliance" - more akin to swimming with sharks than to dancing with porcupines. To avoid becoming shark bait, universities might be wise to decline the swimming invitation altogether.

Universities and university researchers have, as their prime duty, the obligation to put the health and safety of the public above all else. When they accept drug company donations and funding, they are beholden to the companies. Disclosing this conflict of interest is scarcely an adequate solution to the problem. Would we permit our judges to own shares in for-profit prisons on condition that they disclose their shareholdings? Would we allow judges to accept payment from some of their corporate litigants? The answer is obvious. Our judicial system would lose all credibility if such practices were permitted, even with full disclosure.

Analogously, if biomedical researchers and their universities are to retain public trust, then they should simply not be permitted to put themselves into situations of financial conflict of interest. If the community values public science in the public interest, then it will have to be paid for by public tax dollars.

This does not mean that university-originated discoveries should never be commercialized. In western marketplace societies, many of the discoveries of fundamental research will be developed and marketed commercially. It should become the job of governments to develop new mechanisms so that a fair share of the resulting profit would be captured for the benefit of universities and hospitals. There is no need for our researchers to become handmaidens of business, nor is it desirable for universities to become adjuncts of large corporations. University research and university researchers must be sequestered from the process of commercialization if we want to avoid the kinds of damaging conflicts of interest described earlier.

How might such a sequestration be achieved?

One practical possibility might be to require of any drug company which desires to bring a new drug to market that it provide to an independent institute all the funding necessary for the design and performance of a clinical trial of its drug. The institute would then allocate to qualified university and hospital researchers the task of conducting the necessary clinical trials. The independence and objectivity of clinical research would thereby be protected.

Another promising solution might lie in changes to the tax system. If industry profitably exploits the public's investment in scientific research, it could not legitimately complain when a fair share of its profits from such research is recaptured through special taxation. The government could then provide research support through special taxes raised from corporations which make use of discoveries originating from university science.

Moreover, if drug research were publicly funded there might actually be a net saving, because drug costs would be significantly lower. At present, we waste fortunes on ineffective or positively harmful treatments, not to mention the cost of lives lost or blighted.

The stakes are certainly high. To continue on our present course is to risk losing the one commodity which, for physicians, universities and hospitals, should be viewed as beyond price: the public trust.

Arthur Schafer is Director of the Centre for Professional and Applied Ethics at the University of Manitoba.

¹ Stelfox HT, Chua G, O'Rourke K, Detsky AS. Conflict of Interest in the Debate Over Calcium-Channel Antagonists. New England Journal of Medicine; 1998; 338:101-6.

A more detailed discussion of the issues raised in this article can be found in Arthur Schafer's paper "Biomedical Conflicts of Interest: A Defence of the Sequestration Thesis - Learning from the cases of Nancy Olivieri and David Healy," forthcoming publication in the Journal of Medical Ehtics. The JME has posted the article in advance on its website http://jme.bmjjournals.com/misc/advanced.shtml.

The views expressed are those of the author and not necessarily those of CAUT.