Peer-reviewed articles on imaging diagnostic text accuracy that have positive conclusions are more likely to be published faster, potentially flooding the literature with positive findings.
When it comes to radiology peer-reviewed literature, if an investigator is hoping for a fast publication time, he or she needs to have positive conclusions.
A literature review of 516 studies published between Jan. 1, 2015, and Dec. 31, 2015, revealed that putting a positive spin – that is, a title – on a study does not necessarily correlate to a truncated publication timeline. Instead, the speedier time-to-press is awarded to research articles that do not report negative conclusions.
A multi-institutional team of investigators led by Ottawa Hospital Civic campus published their discovery in the American Journal of Roentgenology on Nov. 13, pointing to the likelihood that this abundance of positive-conclusion research is tipping the scales in the imaging diagnostic test accuracy (DTA) literature.
“[This study] shows that published imaging DTA studies with positive conclusions (but not positive titles) are published faster than those with neutral or negative conclusions,” said the team led by Lee M. Treanor from Ottawa Hospital. “An inflated perception of test performance ultimately may adversely influence clinical decision-making and patient care.”
For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-Newsletter here.
The team identified 516 studies that provided the time frame from study completion to publication, and they examined the positivity of each study’s title, as well as conclusions. What they found was a significant difference on publication time frames for studies that offer positive conclusions.
Specifically, for the 413 studies with positive conclusions, the average time to publication was 18 months, and it was 23 months for the 63 studies with neutral outcomes. But, the process was slightly longer – 25 months – for the 40 studies with negative conclusions.
“Our results align with prior findings of studies in the clinical trial literature, which have shown that studies with positive findings (i.e., those with a positive conclusion or statistically significant results) are published 1.5-to-2 times faster than those with non-positive findings,” the team said. “Our results confirm that time-lab bias is associated with DTA estimates.”
When comparing the impact of the title, the team found there was no significant association between title positivity and time-to-publication, though. For the studies with positive titles, the average time-to-publication was 18 months. It was 19 months for those with neutral titles and 30 months for negatively titled articles.
It is possible, the team said, that imaging DTA research could have an over-abundance of publications with positive conclusions because authors are frequently free to draw subjective conclusions with no standard positivity threshold. This is important, though, they said, because this over-representation of positive results could adversely affect clinical decision-making and patient care.
To push back against that possibility, the team suggested a two-step review process that would allow authors to submit their protocols for review and feedback, allowing journals to reject or provisionally accept a study based on its quality. Reviewers and editors would be blind to any investigation results and, the team said, the method could encourage study authors with non-positive results to write articles detailing their studies in reasonable time frames.