Skip to Navigation
University of Pittsburgh
Print This Page Print this pages

April 5, 2007

Senate plenary: Protecting science from bias

The passage in 1980 of the Bayh-Dole Act, which allowed universities and researchers to benefit from the commercialization of publicly funded research, plunged academia into a new world of forging partnerships with industry, forming startup companies and other aspects of technology transfer.

In a follow-up to last year’s University Senate spring plenary event, which looked at technology transfer trends, the flip side was examined at this year’s plenary session, “Protecting Science from Bias by Private Interests.”

Citing the University’s position as a major research institution that ranks No. 7 in annual National Institutes of Health support and No. 12 in total federal science and engineering research and development grants, Chancellor Mark A. Nordenberg welcomed keynote speakers Sheldon Krimsky of Tufts University, whose recent research has focused on the sometimes uncomfortable relationship between academic research and the industries that sponsor it, and Pitt trustee Catherine DeAngelis, editor of the Journal of the American Medical Association (JAMA).

Noting that the University expends more than $600 million annually in research grants, Nordenberg said, “Obviously, that is a wonderful position to be in, but it also carries with it an important responsibility to ensure that high standards of integrity are maintained in the research conducted by the people of our University.”

In his opening remarks, plenary session organizer and co-chair Herbert Needleman, a professor of psychiatry and pediatrics, noted, “Science is a powerful tool. It is also a fragile construction. It’s often vexed by bias. Scientists expend a great deal of effort and energy trying to reduce or eliminate the many different biases that impinge on decision-making.

“We use control groups, double blinding of investigators, complex mathematical techniques to adjust for confounding variants and many others. But there are other sources of bias and the most apparent and perhaps important is the bias from industrial sources.”

Krimsky called for caution on the part of universities in entering partnerships that may undermine the freedom of their scientists or ownership of research data. He also advocated reform in drug testing, calling for a separation between corporate interests and those who test drugs and medical devices.

While outside influences on academic science aren’t new, Krimsky noted that profound differences began to arise in the late 1970s as fears of federal research funding cuts linked to the passage of the Bayh-Dole Act “created a kind of incentive for universities to become academic enterprise zones.” Now, citing the trend toward academic capitalism, “The production and marketing of knowledge are connected,” he said.

The change brings with it new ethical questions. “I guess the big question is: Does corporate funding distort the values in science and medical practice?” he asked. “Does conflict of interest impair scientific integrity? How is the corporate influence on science played out in drug and environmental research? And, is there a funding effect in science? In other words, does it really matter for the quality of the science where the funding came from?” he said.

“How do American corporations view their patronage relationships with science? How do corporations influence the outcome of academic research? How do corporations suppress undesirable research results? … How do corporations fund academic science to undermine credible scientific studies by manufacturing uncertainty? And after we’ve diagnosed the problem, if university-industry partnerships have become a problem, what can be done about it?” he asked, noting that universities now depend upon funding from such sources.

Krimsky cited several examples in which he said outside interests can undermine scientific integrity. Among them are ghostwriting, in which scientists are paid to “author” papers that are written by corporate interests; the funding of studies that yield positive findings about a company’s product or discredit critics of it, and companies retaining control over the publication of study results in exchange for funding or for research materials such as drug samples.

“Corporations view science not as a generator of truth, but as one of many inputs into production,” he said.

Journal editors have been threatened with lawsuits if unflattering articles are published, and data may be suppressed when conclusions inconsistent with corporate financial interests result from company-funded research, he said.

“Corporations want an advocate scientist,” he said. “They don’t want one who paints the full picture.”

Krimsky noted that a Nature magazine survey of scientists asked whether they’d ever changed the results, designs or methodology of a study in response to a sponsor’s pressure. In response, 20.5 of mid-career scientists and 9.5 percent of early-career scientists said they had, Krimsky said.

Citing a 2003 JAMA study, Krimsky said, “Evidence suggests that financial ties that intertwine industry investigators and academic institutions can influence the research process. … Strong and consistent research shows industry sponsored research tends to draw pro-industry conclusions.

“Whenever a drug company studies head-to-head its drug with respect to some other company’s drug, chances are more than 75 percent of the time, their drug is better. Whenever they study old versus new drug, new drug is always better. Why?” Because the new drugs may be protected under patent while ones on which the patent has expired are less lucrative, he said.

Citing the case of University of California professor of clinical pharmacy Betty Dong, Krimsky decried the practice of giving corporate interests veto power over publication of scientific results in exchange for funding. Dong’s research compared the thyroid drug Synthroid with generic products and found they were bioequivalent. But because the maker of Synthroid had sponsored the study, it fought to suppress the results when Dong submitted them for publication in JAMA.

“The university made a mistake. They signed a contract which gave the company the rights over publication,” Krimsky said. Although the article Dong submitted was well reviewed, “The company didn’t want to hear that,” and she was forced to withdraw the article under threat of a lawsuit, he said. “Not only did the company force her to pull it from JAMA, they took her data, they re-analyzed it and published it in a third-rate journal declaring they were the winner.” Krimsky said Dong later was able to publish the data and the company was sued for false marketing practices.

Addressing Pitt administrators in the audience, Krimsky said gravely, “Chancellor, Provost, make sure you don’t sign any contract at your university that takes away the autonomy of your researcher to publish federal science.”

Provost and Senior Vice Chancellor James V. Maher later noted that the University refuses to accept research grants that restrict researchers’ ability to publish and also refuses material transfer agreements that restrict a faculty member’s right to pursue research based on using the material. In addition, Arthur S. Levine, senior vice chancellor for Health Sciences and dean of the School of Medicine, noted that Pitt has a University-wide policy of refusing tobacco company sponsorship of research.

Krimsky proposed that the roles of those who produce knowledge in academia and the stakeholders who have a financial interest in that knowledge be kept separate and distinct. “You want to have one sector do the innovation and another sector do the assessment and evaluation,” he said. “To mix them is to get into all of these conflicts of interest.” Merely disclosing financial interests, often thought to be a safeguard, isn’t necessarily enough, he said.

Krimsky also called for a reform of the system of drug and chemical testing, including a public database listing the results of all clinical trials. “Ideally we should have a firewall between the drug manufacturers and those who evaluate the safety and efficacy of drugs. We should have something like a national institute of drug testing” which would choose who does the testing. “And of course we should have full disclosure and of course employees of the FDA should be paid by the government,” he said.

In universities, “No contract should be accepted that restricts the scientists or medical researchers from having full autonomy over research and ownership of the data,” he said, adding that full conflict-of-interest disclosures should be required and that researchers with financial interests in a drug or device should be kept from testing those devices or their competitors’ in clinical trials.

Echoing Krimsky’s remarks was Pitt medical school alumna Catherine DeAngelis, editor of JAMA.

DeAngelis noted that conflicts of interest are a fact of life in medical research. Among the temptations for researchers and authors are career advancement, peer recognition, competing research interests, competition for grants, garnering high profile publications, intellectual biases or passions and financial considerations.

For editors, there are other potential conflicts: promoting themselves and their journals, improving their impact factor, increasing subscriptions and profitability and decreasing stress, hostility and harassment all can come into play.

At JAMA, DeAngelis said she aims to avoid conflicts by ensuring that the journal’s advertising is sold by the publishers, not by the editorial staff, and that the publishers don’t know the editorial content of the issue when selling the ads. In addition, she holds veto power over what advertising is accepted, DeAngelis said. “My publishers tell me that I cost them at least a million and a half dollars a year in ads I absolutely refuse to publish,” she said, adding that even if a product is FDA approved, it’s not necessarily enough to convince her the ad should appear.

She said JAMA is laid out in a way that the scientific articles are separate from the advertising. Purists can pull the publication apart, “so you can have an intact real journal with all the scientific articles in it and there’s not an ad anywhere in there,” she said. “That’s the best I can do and still be able to know that we can afford to put out a journal. But I still don’t like it,” she said, uncomfortable that accepting advertisements is a financial necessity.

Financial conflicts of interest can take many forms and disclosure of interests isn’t without its own problems, she said. Conflicts might be current, recent, remote or future, DeAngelis said, noting that one of many troublesome loopholes is the future.

Even if a deal is in the works, if there’s no conflicting relationship at present, it might not be disclosed.

DeAngelis noted the increase in articles on the topic of conflict of interest: They’ve gone from two or three listed on the Medline database in the mid-1970s to 515 in 2005.

“Why are we all worried about this?” she asked. Drug companies lead in lobbying government officials and pour money into campaign coffers, she said. Two-thirds of all money for all kinds of biomedical research comes from for-profit companies, DeAngelis said. And, she noted, for clinical studies and trials about 90 percent comes from pharmaceutical companies and device makers.

“If it takes $800 million to develop a new drug, who’s going to be able to put that kind of money in it?” she asked, noting that it’s probably appropriate that the companies who will benefit from the outcomes should be paying for the studies. But she takes issue with how that’s done, arguing that the duty should be placed into the hands of a trustworthy organization that would direct the studies.

“That’s not what happens now,” she said, adding that those who control the studies can control the outcomes, again citing the 2003 study Krimsky mentioned that showed industry-affiliated researchers are four times as likely to report pro-industry research findings.

DeAngelis recounted several examples from her own experience at JAMA, including one that came in her first year as editor. A clinical research trial examining the addition of HIV immunogen to patients’ treatment found it to be ineffective. The manufacturer sued the research institution in an attempt to prevent publication of the finding, and DeAngelis herself was threatened with a suit. The researcher had 85 percent of the data, but the company would not release the remainder, although the researcher’s portion of data was sufficient to show the vaccine did not work.

DeAngelis said she found it unconscionable that the attempts to block publication continued while people with HIV infection were continuing to receive the ineffective treatment, she said.

She also noted the “deliberate lying” by the manufacturer of Celebrex, who attempted to publish claims that the arthritis medication was equivalent to other types of medication in terms of pain relief, but with fewer gastric consequences. Research was submitted to JAMA with six months’ worth of data, but when she noted that the study was in fact 12 months long, DeAngelis was told there was only six months of data available.

“It turned out that they lied. They had the 12 months,” she said, adding that she recently provided a deposition in a case in which those who bought stock in the company based on the study were suing its corporate descendent Pfizer, which acquired Celebrex in its 2003 purchase of Pharmacia Corp. The lying and falsification she said were revealed in the case was “disgusting,” she said. “The only thing I wish I’d brought into that room [when she was deposed] was an emesis basin,” she said.

DeAngelis said she’s become very cynical when reviewing JAMA submissions. “Be very, very careful about what you read,” she said, advising readers also to be very careful about researchers’ disclosures.

“I turn down papers every single week because I don’t like what the disclosure said. It costs us, but so be it,” she said.

Other conflict issues in the news lately include concealing bad news, such as in the case of Vioxx, which was approved by the FDA but later withdrawn from the market by its manufacturer amid concerns of an increased risk of heart attack and stroke associated with its use. The conflict of sheer fabrication, as was found in the work of Korean stem-cell researcher Woo Suk Hwang, also is problematic. “It arouses public concern and threatens the credibility of biomedical research,” DeAngelis said.

“We need to police our own,” she said, criticizing the practice of pharmaceutical company perks for doctors and corporate involvement in providing continuing education for doctors. “No pharmaceutical company is going to spend $7 million a year for dinners, pizza, whatever, if it didn’t think it was going to get some good out of it,” she said, adding that drug companies shouldn’t be using their representatives to teach doctors. “Physicians should be teaching physicians,” she said. In addition, DeAngelis criticized companies’ direct-to-consumers drug marketing, labeling it all a morass of conflict of interest.

Noting that conflicts are ubiquitous, managing them in a way to ensure the public trust is necessary, she said. Partnerships are needed, but they must be partnerships done the right way, DeAngelis concluded.

Levine pointed out that academia and businesses have conflicting values but common interests. Universities seek knowledge for knowledge’s sake; companies seek the management of knowledge for profit, he said.

While conflicts of interest aren’t new, the potential for conflicts has increased with medical researchers’ new ability to make products that are useful, commercializable and that can generate profit, Levine said.

The Bayh-Dole Act, he noted, has introduced the notion of a “duality of interests.” Levine added, “Entrepreneurship has for a long time been the fourth mission of academic health centers.” The impact of decreasing NIH and philanthropic dollars, increases in malpractice costs and the cost of caring for uninsured patients all are enticements for entrepreneurship.

Often, Levine said, collaboration is the way to bring medical advances to the bedside, adding that partnerships “don’t have to be evil.”

Pitt deals with potential conflicts of interest in several ways, he said, including disclosure policies and reviews and oversight at a number of levels.

Levine said the policy in the schools of the health sciences is that “Educational research and clinical decision-making must be free of all industry influences.”

Following questions from the audience, Maher wrapped up the afternoon with his own observations. Acknowledging Krimsky’s assessment of Bayh-Dole as accurate, Maher said he felt it also was incomplete, adding that the legislation “provides universities an incredible — and what in 1980 was an entirely new — challenge to their mission.”

He compared the situation to the days early in his career when the idea of a research university was not firmly established. “There still were people who quite sincerely worried that a good university could not do research without abandoning teaching,” the provost noted. Maher said that challenge has been overcome and research and teaching are now incorporated in a creative way that has made today’s graduates more sophisticated than those of a half-century ago.

More recently, Bayh-Dole forced on universities another mission, Maher said. “They made it a responsibility of any university that accepts public money to do research that that university must try to commercialize the research that emerges from its laboratories. It threw the professors and the administration of the university into a whole new stance that, in order to be compliant with public policy that wants to commercialize this intellectual property for the good of the nation, we must have our people not just do teaching and research but also be cooperative with this responsibility to commercialize.”

Noting that university leaders continue to discuss the implications, he said, “We do not have our arms around it yet. … What you’re hearing today is we’re not there yet, nationally or locally. We’re still struggling with how do we deal with this new part of our mission and retaining for the good of the world the pre-existing parts of our mission to do really good teaching and really good research.”

He added, “We’re really having a bit of difficulty with it. We’re getting there, though.”

Noting that a large fraction of job growth since 1980 has been in small companies, including those that emerged out of university intellectual property, Maher said, “This is public policy with a vengeance. Congress consciously decided to do that and it worked. And we’re having to cope with the problems that have arisen through this enormous expansion of the American economy, not just in the health sciences but in essentially all kinds of technology.

“Every problem that was raised today is real. Every problem that was raised today we’ve got to cope with. But we do have to recognize that it’s worth learning to cope with it because there are so many good things that can come out of it.”

—Kimberly K. Barlow

Leave a Reply