NATURE

Scientists behaving badly

Journal editors reveal researchers' wicked ways.

4 March 2004

JIM GILES

They lie, they cheat and they steal. Judging by the cases described by a group of medical journal editors, scientists are no different from the rest of us.

Last week's annual report¹ of the Committee on Publishing Ethics details the misdemeanours that the group of journal editors grappled with in 2003. Although the number of reported cases - 29 - is tiny compared with the tens of thousands of papers published in medical journals every year, the cases cover a wide range of unethical activity, from attempted bribery to potential medical malpractice.

Many of the tricks will be familiar to schoolchildren. Two complaints concern cases where researchers were accused of copying someone else's work. When editors investigated, they agreed that the papers were almost identical versions of previously published material, and that plagiarism was the most likely explanation.

Confronted with the evidence, researchers behind one paper insisted that their paper contained only 5% overlap with the original. Another author, when eventually reached by mobile phone, admitted some similarities; but at that point the call ended abruptly.

Duplicate publication, where the same paper is printed twice in different journals to boost publication records, is the most common offence, accounting for seven of 29 cases. This fits with previous studies of the practice.

A 2003 survey of ophthalmology journals estimated that at least 1.5% of all papers are duplicates². Some researchers seem to have perfected the art: a study released last month identified two papers that had each been published five times³.

Compulsory action

Conflicts of interest also rear their head in the report. One journal ran a paper on passive smoking from authors who omitted to mention that they had received funding from the tobacco industry. Further probing revealed that the author had

received tobacco company money throughout his career and even lobbied for the industry.

In cases where the misconduct concerns medical treatments, the report becomes more disturbing. The editors discuss several studies where medical procedures were run by researchers who did not have proper ethical clearance.

One paper revealed that blood samples were taken from healthy babies to set up a control group for a study. This was a painful procedure that the paper's authors later said wouldn't normally be sanctioned for research purposes. The nature of their ethical approval for the procedure was never cleared up.

When confronted with such issues, journal editors usually contact the researchers' employers or ethics committees, who may take action. But this is not compulsory.

The publishing committee wants to formalize this course of action in a code of ethical conduct for editors. It has published a draft of such a code alongside its report, and a final version should be ready in the next few months. The committee wants all editors of medical journals, including its 180 or so members, to sign up to the code and agree to be bound by the associated disciplinary procedures.

Such a code should clarify editors' duties. It should also make clear, if it is not already, which activities are inappropriate. The report describes one bid to persuade an editor to accept a manuscript, in which an anonymous caller offered to buy 1000 reprints of the published paper. "And," the caller added, "I will buy you dinner at any restaurant you choose."

References

1.The Cope Report 2003, (2003). |Article|

2.Mojon-Azzi, S. M. et al. Nature, 421, 209, doi:10.1038/421209a (2003). |Article|

3.von Elm, E. et al. J. Am. Med. Assoc, 291, 974 - 980, (2004).