Back in 2000, researchers led by Pharmacia's Stephen Geis reported on the outcome of a randomized, controlled trial of celecoxib, the COX-2 inhibitor better known as Celebrex. After six months of treatment, patients who took celecoxib had lower incidence of ulcers and other toxic effects compared to those who took ibuprofen or diclofenac, the researchers reported in the Journal of the American Medical Association. However, as letters to JAMA later pointed out, that data appeared to be cherry-picked. According to trial data submitted to the US Food and Drug Administration, at 12 months and 16 months the opposite was true of the drugs' effect — there was no advantage to taking celecoxib. The authors' competing interests in developing a better drug appeared to have gotten the better of them.
While such cases of conflicting interests affecting what is reported in a journal are not common in genomics, as the field moves toward translational medicine and develops more assays and devices destined for the clinic, disclosure and mitigation of conflicts of interest will become more important. "Over the past few years and going forward, there will be more opportunities for genomics researchers to find themselves with a competing interest — people who are finding themselves moving from the basic research into translational research," says Rebecca Furlong, executive editor of Genome Medicine.
Howard Brody, a bioethicist at the University of Texas Medical Branch at Galveston, says that in the medical field, physicians' primary duty is to their patients and their health. Likewise in scientific research, researchers have a duty to the science — to do their work with integrity and to report accurate and reliable findings. A conflict of interest is anything, he says, that could tempt physicians or researchers to give their primary interest a backseat in favor of other priorities. "To have a conflict of interest means that you have entered into some kind of arrangement that would make you at least be tempted ... to serve some conflicting -interest — some other interest that takes you in a different direction than what your primary obligation requires," Brody says. Serving that secondary interest would include actions like revising or downplaying negative or unwanted results, as the researchers in the celecoxib study appeared to have done.
"If an author has a competing interest, it could, in theory, influence their interpretation of the data. That's why the declaration is important," Furlong adds.
To address this, most journals, including those that publish genomics and related research, have policies requiring authors to disclose any competing interests they may have. Both Genome Medicine and Genome Research require authors to fill out and sign conflict-of-interest statements prior to the publication of an article.
"We absolutely require that the statement has to be completed prior to publication and, obviously, if they have no competing interests, then this must still be explicitly declared," Furlong says. "In terms of making sure that people are disclosing everything properly and not leaving things out, obviously editors have a responsibility. We always check in house, first of all that the statement is present and, whether it is present or not, if there is anything obvious in the article that might perhaps be a competing interest. That's part of triage process of the submission of an article."
Hillary Sussman, executive editor of Genome Research, says that for her journal, "every author is required to declare their conflicts of interest or competing financial interests on a form after acceptance and before any production work is done on the manuscript." She also says that the conflict of interest policy at Genome Research is being revisited. As articles in the field tend to have a large number of authors, getting each one to fill out a conflict of interest statement can be cumbersome. The journal is considering having the corresponding author fill out the form on behalf of all of the authors.
These disclosures then become the statements at the end of the journal article — either listing authors' competing interests or saying that they have no conflicts to report. These statements are ethical "red flags," Brody says. As a flag, a disclosure of a potential competing interest does not mean that the researcher gave into an ethical temptation, Brody says, just that the temptation exists. It points out that "this is a situation where people with normal human psychology, in the past, put in those kinds of situations are frequently tempted to not pursue their primary obligation that they should really be pursuing and to be distracted by their secondary interests," Brody says.
The hope is that readers see these flags and take that information into consideration when assessing the article. "I read it. As a reader, and looking at articles in other journals, I look and see, is this a pharmaceutical company publishing the work or related to a company?" Sussman says. "I always look to see where the conflicts are and the relationship to the authors."
Researchers, of course, are savvy and trained to be skeptics. "They don't just take the authors' interpretation of the data for granted," Genome Medicine's Furlong says. "Readers will assess [competing interests] in coming to a conclusion about the data the same way that they would assess all of the rest of the manuscript — their prior knowledge of the subject, anything they already know about the authors, and their own opinions of the methods that are being used. … If you read that there's a conflict of interest, but everything in the manuscript is done in a fair and scientifically valid way, then that's just part of the whole."
In addition, journal articles undergo peer review. "I would think that [readers] would look at any conflicts — that if we had a paper from Helicos or 454, I would assume that when they read those conflicts, they still assume that paper's undergone a rigorous peer review," Sussman says. "As long as it's disclosed, and people take the work for what it is, they know that it has been peer-reviewed under high standards at Genome Research."
Disclosing conflicting interests, though, does not eradicate them or the problems they present, says Brody, who thinks that disclosure is only part of the solution. "If I tell my wife that I have a conflict of interest because I am actually having an affair on the side, but I am purely moral and ethical because I told you about it, I don't think that goes very far," he says. "Just merely disclosing something in and of itself does not resolve the problem."
Having a large number of conflicts of interest appear from a group in a journal can also erode public trust, he says, regardless of whether wrongdoing has occurred. "It may be that we know that each individual has such high integrity and is so able to withstand temptation that none of them step over the line and do the wrong thing. Can we expect the general public to have the same confidence that that is so?" he asks. In light of that, Brody advocates that conflicts of interest be less tolerated.
There are conflicts of interest, particularly in medicine, that physicians or researchers don't need to open themselves up to. Some, like grants to do research are reasonable — funding is hard to come by these days — but others are less so. "If the company offers you a nice grant, OK so you take their grant, that's understandable. But to be on the speaker's bureau, to be a consultant, to have stock options, to do some of the other stuff that the high rollers do, no that's not necessary," he says. "Sorry, you can get by on what most physicians get paid, even in academics."
As for the conflicts that are unavoidable, there are ways to mitigate their effect. "You don't say the person can't publish, that this person can't be the PI; you acknowledge that there is a serious conflict and so you address it because it really is unavoidable," Brody says. For example, when a trial of a patented medical device is to be done, it's difficult to do that work without the person who developed the device as the PI, even though that person has a conflict of interest. An advisory or review board could be set up to watch over the trial, keep the researchers on their toes, and provide an extra layer of scrutiny. In the end, it's about professionalism and trust, Brody adds.