PubMed Asked to Add Conflict-of-Interest Info to Abstracts

April 04, 2016

Believing there can never be too much transparency, a bevy of scientists, physicians, and US senators has asked the National Library of Medicine (NLM) to add conflict-of-interest information to article abstracts available through PubMed, the library's free online search engine.

Betsy Humphreys, the NLM's acting director, says her agency will explore how to provide abstract readers those data in a way that does not create new problems in the process.

Journal articles increasingly include information about conflicts of interest for the authors, but many PubMed users do not go beyond the abstract, 62 scientists and physicians, along with six organizations, said in a March 30 letter to Humphreys and Francis Collins, MD, PhD, the director of the National Institutes of Health. Some of these PubMed users, particularly journalists, do not have immediate access to the full article, whereas many users who do only want to read the abstract, the letter signatories said.

Disclosing conflict-of-interest information is crucial because "numerous studies have documented that industry funded research is far more often than not associated with results supportive of the funders' interests," they said.

The individual scientists and physicians who signed the letter are from the United States, Latin America, Europe, and Australia. The six organizations that added their names are the Center for Science in the Public Interest, the Environmental Working Group, the Humane Society of the United States, the Natural Resources Defense Council, the Union of Concerned Scientists, and the World Public Health Nutrition Association.

Five Democratic senators made the same points in their own March 30 letter to Humphrey and Dr Collins. They singled out industry-funded research that minimized the dangers of harmful chemicals and tried to "obscure the link between soda products and obesity, diabetes and heart disease."

The five senators signing the letter are Sens. Richard Blumenthal (D-CT), Elizabeth Warren (D-MA), Edward Markey (D-MA), Brian Schatz (D-HI), and Dick Durbin (D-IL).

Hyperlink Might Be the Solution

The scientists, physicians, and organizations that petitioned the NLM and National Institutes of Health envision adding a conflict-of-interest statement between the authors' names and the main body of the abstract. This statement would indicate whether the authors reported competing interests, whether the article lacked a disclosure statement, and whether the journal failed to disclose funding sources or the authors' other competing interests.

The NLM's Betsy Humphreys told Medscape Medical News that the best way to supply this information in an abstract might be through a hyperlink, a method her agency already uses when it wants to explain that an article has been retracted or corrected.

Simply adding the conflict of interest verbiage to the PubMed abstract may confuse data mining programs that look for indexed words in an abstract to compile a list of relevant studies. "The conflict-of-interest information may have words, concepts, and phrases that are potentially misleading," said Humphreys. "You might come up with articles that aren't valid for the subject you're researching." The conflict-of-interest information, she added, does not necessarily reflect the scientific content of an article.

The NLM will need to talk to journal publishers and editors to figure out the most practical solution to the problem of missing conflict-of-interest data in PubMed abstracts, Humphreys said. If the solution turns out to be a hyperlink, publishers and editors would have to submit it electronically, as they do the abstract and bibliographic information. The additional data would require changes to their submission routines, which may not happen overnight.

"These changes are likely to be easier for large publishers than for small ones," she said.

Humphreys noted the mismatch between the almost 26 million citations in PubMed, most of which come with an abstract, and 3.8 million full-text articles available for free on NLM's PubMed Central. Researchers who cannot find a full-text article in PubMed Central can turn to the publisher's website, which may charge for it. Some people balk at paying the price and confine themselves to the abstract, missing out on crucial information.

Ever since abstracts first appeared in the MEDLINE bibliographic database, said Humphreys, there have been moans about "people practicing medicine based on abstracts."


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.