San Francisco

Many authors continue to keep financial conflicts of interest to themselves, despite ever stricter journal policies requiring full disclosure in published articles. That is the conclusion the Center for Science in the Public Interest (CSPI), a Washington-based pressure group, has reached after reviewing all the articles in four major medical research journals over a three-month period.

The problem is the result of a combination of lax enforcement by editors and loopholes in disclosure policies, says Merrill Goozner, a CSPI researcher and the author of the review. “Our study is a reminder that journals need to be vigilant,” he says.

Concerns that researchers' financial interests might influence their objectivity or methods have been increasing for 20 years as ties between industry and the academic sphere have broadened. Yet research journals have been slow to respond. By 1997, only 16% of the 1,400 top biomedical journals had any competing-interest disclosure policy in place, and enforcement was patchy at best (see Nature 411, 3; 200110.1038/35075210)

Goozner examined articles published from December 2003 to February 2004 in the New England Journal of Medicine, the Journal of the American Medical Association, Environmental Health Perspectives and Toxicology and Applied Pharmacology. He used websites and public databases to investigate the lead authors of each article in which a declaration of no competing interests was made. In a report released on 12 July, Goozner listed 13 studies, out of 163, that failed to comply with the publishing journal's disclosure policy.

In one example, an author applied for a patent on a reagent that his paper suggested would be valuable in cancer research. In another, the authors on a paper about coronary heart disease were paid consultants to more than 20 companies in the heart-disease field, but did not disclose this. According to the CSPI report, the authors confirmed the relationships by e-mail, but said “None of them had anything to do with the paper. Why give them free advertising?”

Goozner says it should be up to the readers to decide. Because journals ask authors to declare only ‘relevant’ or ‘directly relevant’ ties, authors can too easily rationalize their omissions, he claims. Goozner's solution is to require disclosure of all ties to private firms and all patents regardless of relevance, and then let the editors decide which to publish.

The situation may be improving. Although it examined different journals, a 2001 study found that only 1% of papers disclosed a conflict (Sci. Eng. Ethics 7, 205–218; 2001). In the current study, that rate was up to 20% of papers, with those failing to declare an interest making up only 8%. That at least brings the problem into a manageable realm, says bioethicist Mildred Cho of Stanford University, California. “We should be able to deal with this in the future.”