SCOTTSDALE, ARIZ. – Since 2005, only 26% of randomized trials published in the leading headache medicine journals have followed registration reporting guidelines from the International Committee of Medical Journal Editors, a study found.
“The three major headache journals should pursue consensus regarding ICMJE registration requirements and their enforcement,” lead author Dr. Melissa Rayhill said at a symposium sponsored by the American Headache Society.
Instructions for authors should state that the journals require abstracts about randomized trials to report trial registration numbers, said Dr. Rayhill, a neurologist at Brigham and Women’s Hospital, Boston.
She and her coauthors found that paper titles did not consistently identify randomized trials as such, making them less likely to appear in PubMed searches.
“This has implications for meta-analyses and other efforts to synthesize trial information,” she said.
To address the problem of selective reporting of results, since 2005 the ICMJE has required member journals to only publish reports of clinical trials that were prospectively registered in an approved clinical trial registry. Other journals later followed suite, and the journal Headache adopted the policy in 2009. As a result, the number of registered studies rose from less than 4,000 in 2000 to more than 177,000 in 2014, Dr. Rayhill and her associates reported.
To assess the extent to which trials reported in headache medicine journals have complied with ICMJE requirements, Dr. Rayhill and her colleagues manually reviewed all issues of the three leading journals – Headache, Cephalalgia, and the Journal of Headache and Pain – from July 2005 through June 2014. After identifying papers on randomized clinical trials, they determined which manuscripts reported valid trial registration numbers, and then cross-checked the relevant website to see if the trial was registered.
The three journals published 226 reports of randomized trials between 2005 and 2014, Dr. Rayhill and her associates reported. A total of 46% of these papers appeared in Cephalalgia, 38% in Headache, and 16% in the Journal of Headache and Pain, they said. Only 59 (26%) manuscripts reported a trial registration number that linked to corresponding entry on the clinicaltrials.gov website, they added.
The proportion of published trials that reported a registration number “steadily increased” for the journals Headache and Cephalalgia during the study period, but not for the Journal of Headache and Pain, the investigators reported. In fact, only 8% of trials published in the Journal of Headache and Pain reported a registration number, they added.
Authors should know that trials can be registered retrospectively in cases where prospective registration is not required, Dr. Rayhill and her associates said. They declared no relevant conflicts of interest.