WASHINGTON—Drummond Rennie is a self-professed “fraudy”— his term for members of the coterie of journal editors, university administrators, science lobbyists, and government officials who are called on to affer testimony, give lectures, and attend meetings on science fraud. But that doesn’t mean that Rennie, deputy editor of the Journal of the American Medical Association, enjoys the title. In fact, he doesn’t think that fraud is very common within the research community, and he positively resents having to air what he calls “biomedicine’s dirty linen.”
But Rennie is also a realist. He recognizes the growing concern in Congress that scarce public funds are being squandered by fradulent scientists, and he understands the damage that a few well-publicized cases of serious misconduct can do to the reputation of science. He also knows that the first step in finding the answer to any problem including fraud, is to collect data.
That’s why Rennie has challenged his colleagues to support a random audit of the raw data that form the basis for future submissions to a handful of journals. He made his provocative suggestion at a two-day conference last month on the ethics of scientific publications, a meeting funded by the National Science Foundation and sponsored by the Council of Biology Editors. He reminded his audience that the current debate about scientific misconduct lacks any re liable information on the prevalence of the problem. And he pleaded with his fellow “fraudies” to act quickly.
“Auditing is going to become a part of our medical and scientific lives, whether we like it or not,” he declares. “We need to seize the initiative and gain some control over something that is going to have an enormous impact on us.”
Rennie admits that his proposal isn’t a cure, or even a treatment, for searcher Robert Slutsky. “Once you have someone drawing a salary to investigate fraud, they’ll find something, I promise you.”
Other skeptics argued that scientists bent on committing fraud could easily circumvent any audit. And the most vocal critics, notably editor-in-chief Arnold Relman and executive editor Marcia Angell of the New England Journal of Medicine, declared that they want no part of such a process. “I do not share the enthusiasm some have shown for the idea of an audit,” said Relman. “No matter how simple you make it, it will still produce a nightmare of confusion and misunderstanding. Moreover, I think it would erode the fabric of trust upon which science is based.”
Rennie acknowledges that his proposal is still fuzzy on such issues as protecting the confidentiality of the data, defining what would be audited, and devising a mechanism that would be practical and unobtrusive. But finding a sponsor, a perennial problem for researchers, doesn’t appear to be a major obstacle.
“We’d be happy to consider a proposal relating to peer review and ethical conduct,” says Rachelle Hollander, whose Ethics and Values Studies program within NSF provided a $55,000 grant to fund the conference. And she offered Rennie and his supporters some free advice: “The key to conducting a fraud audit is to keep it very simple, and decide ahead of time what you are trying to accomplish. I don’t think anybody’s talking about prevention, or going out there with the intention of tracking down cases of misconduct.”
Janet Newburgh, NIH misconduct policy officer, was equally supportive of the idea. She said that NIH has some $50 million available each year for evaluation studies and that Rennie’s proposal would probably fit within the broad range of projects the money was intended to support. But she cautioned that several questions need to be answered before any such proposal could be funded. ‘The big question is its feasibility and whether you would get something useful out of such an audit. If it’s crude, you might miss a lot of magnificant misconduct.”
Rennie said he plans to seek support for his idea among his editorial colleagues and to inform certain members of Congress that scientists are prepared to tackle the issue of fraud. But the most appealing aspect of his proposal, he says, is its neutrality.
“If we find that 30% of the papers are based on evidence that doesn’t exist,” he explains, “we’ll be brought up short and be forced to realize the seriousness of the problem. But if it’s 0%, we can say, ‘It’s not a problem. Let’s stop talking about fraud and do something more useful.’ “And that would allow Rennie to escape his worst nightmare: a career spent as a “fraudy.”