UCLA researchers report that although the majority of medical journals have conflict of interest policies in place for study authors, less than half require such policies for editors or peer-reviewers. In addition, many journals do not inform readers about those potential conflicts that have been disclosed to them.
The study, published in the December issue of the Journal of General Internal Medicine, is one of the largest studies of its kind, and evaluated conflict of interest policies and practices for a wide range of major biomedical journals, including general and specialty publications.
Conflict of interest occurs, for example, when an author reporting on a new drug also has ties to the pharmaceutical company that manufactures the drug. Media attention and public scrutiny about conflict of interest in biomedical research has increased, and journals have stressed the importance of such disclosures.
“While it is healthy that doctors, patients and journals are paying more attention to conflict of interest, it is not meaningful if the disclosures are not made public, or if the disclosure policies do not pertain to those individuals who decide what gets published,” said Dr. Jerome Hoffman, study author and professor of emergency medicine at the David Geffen School of Medicine at UCLA.
Study researchers queried editors of 135 peer-reviewed biomedical journals on specific conflict of interest policies; restrictions based on conflict of interest and the public availability of these disclosures.
Of the 91 journals that responded to the survey, 85 journals or 93 percent reported having a conflict of interest policy for authors. While 77 percent reported collecting conflict of interest on all author submissions, only 57 percent published all author disclosures.
“We found it striking how many author disclosures were not made public in these prominent journals,” said Dr. Richelle Cooper, study author and assistant professor of emergency medicine, David Geffen School of Medicine at UCLA.
Researchers note that absence of disclosure appeared most common for narrative reviews, editorials, policy statements and guidelines — the types of publications reflecting opinions of the authors, and that may well have the greatest influence on readers.
A minority of journals reported having a specific conflict of interest policy for peer-reviewers (46 percent) or editors (40 percent).
“Peer-reviewers and editors influence decisions regarding publication, as well as article content,” said Hoffman.
Of the journals reporting a conflict of interest policy for peer-reviewers and editors, the UCLA researchers found that recusal policies were not common, with only 25 percent and 31 percent respectively requiring peer-reviewers or editors to recuse themselves from reviewing an article if they report a conflict of interest.
Only three percent of the responding journals publish conflict of interest disclosures of peer-reviewers, and 12 percent publish editor conflict of interest disclosures, while 11 and 24 percent respectively, reported that this information is available upon request.
“We encourage more transparency in conflict of interest reporting,” said Cooper.
The next step, according to the researchers, is to examine further the extent to which conflict of interest affects the biomedical literature, and how physician prescribing behavior is affected by decisions made by those with a conflict of interest.