In late 2003 the Los Angeles Times triggered a major scandal by publishing information on the conflicts of interest of several staff members of the National Institutes of Health (NIH). Together, these NIH staff members had received several million dollars of external funding since 1995 (1,2).

Following these revelations, the Center for Science in the Public Interest (CSPI), a US consumer organisation (a), examined the frequency with which conflicts of interest failed to be disclosed in biomedical research publications. Undeclared conflicts of interest CSPI surveyed four biomedical and scientific research journals with strict policies on conflicts of interest (b)(3). They focused on primary publications, and tried to identify possible conflicts of interest among the first and last authors, both in ad hoc databases and in various freely accessible information sources.

Among a total of 176 articles published between December 2003 and February 2004, no conflicts of interest were declared by the first or last author in 163 articles. Yet the CSPI researchers found a clear conflict of interest for authors of 13 of these articles (8%). If a less restricted definition of conflict of interest had been adopted, the authors of another 11 articles (7%) would have had to declare conflicts of interest.

Most cases involved financial links, ranging from payments made by companies directly concerned by the study’s outcome, to patents held by an author for a technology that was evaluated in the study, or whose sales might be boosted by the study. CSPI published the names of the authors who did not disclose their conflicts of interest and described the nature of the links they uncovered (3).

CSPI did not examine editorials, comments or review articles, and research into financial links was based on publicly accessible data. It is therefore likely that the true frequency of undeclared conflicts of interest was far higher.

A freely accessible database Since 2001, CSPI has maintained a database, available through its website, describing the financial relationships of many American scientists, learned societies, universities, etc., with industries in the health and food sectors (4). When available, the financial transactions are quantified every year, backed up by a list of information sources.

Thus, in September 2005, CSPI revealed that, in a meeting of the Endocrinologic and Metabolic Drugs Advisory Committee of the Food and Drug Administration (FDA), three of the nine committee members who were asked to judge the risk-benefit balance of an inhaled insulin product had financial links to Pfizer, the manufacturer (c)(5).

In January 2006, the Korean scientist Hwang Woo-Suk published falsified cloning data in Science and Nature, even though he held several patents relating to these results. This led CSPI to ask the two journals to adopt stricter rules and to refuse, for a 3-year period, manuscripts submitted by authors who have not declared their financial conflicts of interest (6).

Despite the obvious difficulties inherent in detecting financial ties, the Center for Science in the Public Interest is focusing public attention on the issue of conflicts of interest in the United States.

A general search and information source for conflicts of interest would be welcome elsewhere.

©Prescrire 2006

Prescrire Int 2006; 15 (84) 148-149.

Notes:
a- This association is mainly funded by subscriptions to its bulletin Nutrition Action Healthletter, and receives 5% to 10% of its income from various foundations (ref 7). Its objectives are to provide the public and decision-makers with useful and objective information; to conduct specific research on technological and scientific aspects of food, alcohol, health, the environment, etc.; and to represent citizens before the regulatory, judicial and legislative authorities (ref 8). Its website address is http://www.cspinet.org
b- The journals concerned were the New England Journal of Medicine, JAMA, Environmental Health Perspectives, and Toxicology and Applied Pharmacology (ref 3). c-One of the members of this committee, who usually chaired it, participated (without voting) despite receiving yearly payments of $5000 to $10 000 from Pfizer (ref 5).

References:
1- Hopkins Tanne J "National Institutes of Health criticised for not preventing conflicts of interest" BMJ 2004; 329: 10.
2- Labrador D "Damage control. A crackdown to prevent conflicts of interest at the NIH" Scientific American 2004. Website http://www.sciam.com accessed 12 October 2005.
3- Goozner M "Unrevealed: non-disclosure of conflicts of interest in four leading medical and scientific journals" Center for Science in the Public Interest - Integrity in Science, Washington 2004: 16 pages.
4- Center for Science in the Public Interest "The Integrity in Science Database". Website http://www. cspinet.org accessed 12 October 2005.
5- Center for Science in the Public Interest "FDA lets scientists with drug company ties evaluate new insulin product". Website http://www.cspinet.org accessed 12 October 2005.
6- Center for Science in the Public Interest "CSPI Calls on Journals to Strengthen Disclosure of Conflicts" Website http://www.cspinet.org accessed 12 February 2006.
7- Center for Science in the Public Interest "Our funding. CSPI's funding sources". Website http://www. cspinet.org accessed 12 October 2005.
8- Center for Science in the Public Interest "Mission statement". Website http://www.cspinet.org accessed 12 October 2005.