Conflicted Science

How to ensure integrity in science


The left-leaning Center for Science in the Public Interest (CSPI) held a fascinating day-long conference last week entitled "Conflicted $cience: Corporate Influence on Scientific Research and Science-Based Policy." Needless to say, most of the 240 or so participants regarded themselves as selfless, public-spirited guardians of the truth. While I have criticized CSPI in the past (as has my colleague Jacob Sullum), CSPI is on to something about conflicts of interest in scientific research.

CSPI executive director Michael Jacobson, in his introductory remarks, emphasized that "We do not contend that industry-sponsored research is always bad science, or that companies should be prohibited from providing input to government agencies." Much of the conference nevertheless had a distinctly anti-corporate tone, but setting aside that reflexive hostility to industry, let's take a look at the conflict-of-interest problem.

As other sources of authority (e.g., aristocrats and religious figures) were eroded away by the Enlightenment, science has become the chief arbiter of truth for our secular society. Consequently, every interest group, government agency, and corporation is desperate to claim the mantle of science for his or her pet project, program or product. What these groups overlook is that policies don't just derive automatically from a set of scientific facts. For example, even though it is scientifically established that smoking cigarettes causes cancer, you may well oppose banning them because you believe that people have a right to choose to enjoy the pleasures of nicotine at their own risk.

The CSPI conferees also overlooked the public choice problems that come with government funding of research. For example, an Environmental Protection Agency bureaucracy in charge of regulating climate change might have a subtle or not-so-subtle bias when it comes to determining the funding of a climatologist whose project challenges the notion of catastrophic global warming. One must keep in mind that, a surprising amount of the time, research aimed at identifying a problem finds that problem.

Still, it is more difficult to make good choices and good policies if the science is no good, and conflicts of interest can bias results. Scientists, like the rest of us, are people who have financial concerns, political viewpoints, and career goals. Of course, the process of peer review is supposed to sift out the dross of these potential biases and leave behind only the gleaming nugget of truth.

Peer review is one of the great inventions of the Enlightenment. A scientist submits her results to colleagues, who then try to tear them apart. If the results aren't shot down, they are provisionally added to humanity's store of scientific knowledge. "Provisionally," because all scientific results are subject to change in the face of new information. Peer review is not a one-shot deal, either. Scientific controversies can break out when dueling studies both initially appear to be valid, but eventually, as more research is conducted, the results are conformed into a coherent explanation.

So is conflict of interest, specifically the conflict of interest when corporations fund scientific research, a problem? Sheldon Krimsky, a professor at Tufts University and author of the upcoming Science in the Private Interest: Has the Lure of Profits Corrupted Biomedical Research?, noted that "Most scientists would deny that funding would compromise their findings." Nevertheless, Krimsky said, "a number of research papers have discovered a 'funding effect,' which biases findings toward industry." Krimsky specifically pointed to studies funded by the tobacco, lead and pesticide industries that produced results favorable to their interests.

Speaker Lisa Bero, a professor of clinical pharmacy at the University of California, San Francisco, looked at contested areas of public health research, such as the effects of tobacco smoke. In these cases, research results are typically subject to review by the lawyers of the companies that paid for the work. Furthermore, the companies often insist on the right to forbid a researcher from publishing the results. That's all very contractual to be sure, but it should call into severe question the credibility of the research.

To handle conflicts of interest, the CSPI offered several guiding principles in a "Scientists' Declaration of Independence." The most important of these is a requirement for researcher transparency; that is, "scientists should voluntarily disclose—in papers, speeches, and elsewhere—their relevant funding sources." A second very important principle is that "academic scientists should not enter into agreements that cede control of research or its publication to sponsors." The Declaration further urges, "scientific and medical journals should require public disclosure of relevant financial affiliations (including research funding, consulting arrangements, and ownership interests) of authors, editorial board members, and the organizational sponsors of journals." Finally, journalists "should routinely ask sources if they have any financial or other ties to companies, trade associations, or other organizations related to their statements." By the way, all of the journalists on the media panel said that they already did that, but pointed out that they have no way of forcing scientists to disclose any of their own potential conflicts of interest.

Since the most contested areas of research generally involve the epidemiology of environmental health effects or clinical trials, a recommendation by Jeffrey Short, a chemist who works at the U.S. Department of Energy on the after-effects of the Exxon Valdez oil spill, seems valid to me: "Supporting data of scientific papers, regardless of the source of research funding, should be publicly available. All scientists … having announced [their] conclusions to the public … should be required to make their supporting data available for public scrutiny."

Drummond Rennie, who works as deputy editor for the Journal of the American Medical Association, and who has been conducting many studies on peer review, had some of the best advice. "The public is advised to be skeptical of favorable results of studies of medical products when the studies have been sponsored by the manufacturers—the ones who have the most to gain from such publication," Rennie wrote in a conference presentation.

CSPI did require that conference speakers disclose their possible conflicts. It turns out that four of the 18 speakers earn a substantial portion of their incomes by being plaintiff's witnesses in product liability suits, while two of the four moderators work for activist groups. Hmmm? None of the four journalists on the media panel at the conference had anything to disclose other than that they are journalists?

Despite the alarms, science is not broken, nor even in really bad repair. However, that doesn't mean that peer review and the process of managing conflicts of interest can't be improved. In a sense, the CSPI conference is just part of a larger societal peer review process to keep one of our most important institutions, the scientific enterprise, honest.