TOPLINE:

Nearly half (45%) of head and neck cancer clinical trials registered from 2005 to 2021 were never published. Trials without statistically significant results and those that were considered incomplete were more likely to remain unpublished.

METHODOLOGY:

• Publication bias, the failure to publish research based on the direction or strength of the findings, is a longstanding issue that risks patient safety, wastes resources, and fosters redundant research.
• To determine the rates of and factors related to nonpublication, researchers analyzed 100 head and neck cancer clinical trials listed on ClinicalTrials.gov with start dates from January 1, 2005, to December 31, 2021. Phase 1 trials were excluded.
• Publications were identified by searching the National Clinical Trial identifier numbers or investigator names on PubMed and Google Scholar and contacting investigators through emails.
• Trials were classified as incomplete if their status was terminated, withdrawn, or unknown with no published manuscript. Statistical significance of the results was assessed from authors’ conclusions in publications or ClinicalTrials.gov tables.

TAKEAWAY:

• Overall, 45% of trials (45 of 100) went unpublished. Of the 66 trials without statistically significant results, 65% were unpublished. Studies without statistically significant findings were 11 times more likely to remain unpublished than those with significant findings (relative risk [RR], 11.08).
• Incomplete vs completed trials (n = 37) were 3.8 times more likely to remain unpublished (nonpublication rate, 83.8% vs 22.2%; RR, 3.77).
• Although the relative risks were not significant, a larger proportion of nonrandomized vs randomized trials went unpublished (nonpublication rate, 62.5% vs 39.5%); phase 2 and 4 trials had higher nonpublication rates than phase 3 trials (46.2% and 60.0%, respectively, vs 35.3%).
• When the researchers looked at trials by funding source, they observed that pharmaceutical-sponsored trials had a 51.5% nonpublication rate compared with 28.6% for government-sponsored trials and 43.3% for university-sponsored trials.

IN PRACTICE:

“Failure to report experimental results is unethical, contributes to research waste, and leads to overestimations of the benefits of potentially suboptimal drugs and therapies,” the authors wrote. More stringent criteria should be in place to reduce early trial termination, and the scientific community “should work to minimize incentives to publish only positive results,” the authors added.

SOURCE:

The study, led by Ethan Lewis, Washington University School of Medicine, St. Louis, Missouri, was published online in JAMA Otolaryngology-Head & Neck Surgery.

LIMITATIONS:

The use of ClinicalTrials.gov as the primary database led to an overrepresentation of US trials and underrepresentation of international studies. Investigators for seven trials could not be contacted to verify unpublished results. Additionally, the sample size may not be fully representative of all head and neck cancer clinical trials.

DISCLOSURES:

The authors did not disclose any funding information. Two authors reported receiving personal fees or compensation from JAMA Otolaryngology-Head & Neck Surgery outside the submitted work and royalties from Washington University. No other disclosures were reported.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.