Photo by Rhoda Baer
Researchers in the lab
Interim results from randomized trials may be misleading at times, according to research published in JAMA.
Researchers compared interim and final publications on randomized trials and found that, 21% of the time, results changed significantly.
“Changes between interim and final publication matter because clinicians and the public could have been misled about whether an intervention was beneficial, harmful, or ineffective,” said study author Lisa Schwartz, MD, of Dartmouth Institute for Health Policy and Clinical Practice in Lebanon, New Hampshire.
She and her colleagues searched PubMed for randomized trials from 2006 to 2015 with “interim,” “not mature,” or “immature” in the title or abstract.
To identify final publications, they searched PubMed, ClinicalTrials.gov, and Web of Science through 2016. The team emailed authors of interim reports when no final publication was identified.
For interim and final publications reporting the same efficacy and or safety outcome, the researchers compared trial characteristics and prominence. They also categorized abstract conclusions (not different, beneficial, or harmful) and compared changes between interim and final publications.
Findings
Interim results were reported in 613 of 1267 screened publications.
Of those publications, 72% reported on trials stopped early (for benefit, harm, futility, or other problems). The remaining 171 ongoing trials (mostly in oncology, surgery, or cardiology) reported interim efficacy or safety results.
Forty-one percent of the publications stated that the interim analysis was specified in the protocol, but half provided no reason for the interim publication.
Final results were published for 98 of the 160 trials (61%) that were more than 1 year beyond the completion date.
The researchers compared 73 interim and final publications reporting the same efficacy or safety outcome. And they found that interim and final publications had similar prominence.
In most cases (79%), the abstract conclusions did not change from the interim publication to the final publication. However, for 21% of trials, there were significant changes.
For 4 trials, the conclusions changed from “no difference” between randomized to treatments to the study treatment being “beneficial.” In 3 trials, the conclusions changed from “not different” to “harmful or possibly harmful.”
In 6 trials, the conclusions changed from “beneficial” to “not different.” One trial changed from “beneficial” to “harmful,” and another changed from “inconclusive” to “noninferior.”
Dr Schwartz and her colleagues concluded that while most interim and final publications reached similar conclusions, frequent non-publication of final results can lead to confusion or unfounded assumptions with true treatment effects remaining unknown.
To safeguard against any such confusion, the researchers recommended routinely adding the word “interim” in the title and justifying the reason in the publication. (Many interim publications reported analyses without any justification.)
“Most importantly, journals, authors, and funders should commit to making final results accessible by linking interim publications to final reports whenever available,” said study author Steven Woloshin, MD, of Dartmouth Institute for Health Policy and Clinical Practice.
