Oncology trial reports written by medical writers are more likely to have surrogate endpoints as the primary endpoint, according to a study published in JAMA Network Open.
Researchers also found that articles written by medical writers were more likely to report favorable outcomes from oncology trials. However, that was no longer the case when the researchers adjusted for trial funding and other variables.
For this study, researchers evaluated articles from The Lancet, The Lancet Oncology, JAMA, JAMA Oncology, The New England Journal of Medicine, and the Journal of Clinical Oncology. The researchers manually collected articles published between May 1, 2021, and May 1, 2022.
The articles were required to report on a cancer trial and contain original research. The trials were required to include a tumor-targeting treatment. A study was excluded if it had a primary endpoint of noninferiority or equivalence, was a meta-analysis or retrospective analysis, or was an observational study.
A total of 270 articles met the criteria for inclusion. About half of the publications (52.2%, n=141) used at least 1 medical writer.
For studies that used a medical writer, the most common endpoint was progression-free survival (PFS), occurring in 32 studies (22.7%). When no medical writer was involved, the most common endpoint was disease-free survival (DFS) or event-free survival (EFS), occurring in 29 studies (22.5%).
Studies for which a medical writer was used were less likely to have a primary endpoint of overall survival (OS) than studies without a medical writer (10.6% and 13.2%, respectively).
Studies with a medical writer were also less likely to have a primary endpoint of DFS or EFS (11.3% and 22.5%, respectively) but more likely to have a primary endpoint of PFS (22.7% and 13.2%, respectively) or response rate (22.7% and 17.8%, respectively).
In an unadjusted analysis, studies with a medical writer were more likely than studies without a medical writer to meet the primary endpoint (58.9% and 49.6%, respectively; odds ratio [OR], 1.45; 95% CI, 0.90-2.36).
Studies with a medical writer were also more likely than studies without a medical writer to have positive conclusions (80.1% and 69.0%, respectively; OR, 1.81; 95% CI, 1.04-3.19).
However, after the researchers adjusted for study funding, randomization, trial phase, and being a primary report of the study, the use of a medical writer was not associated with meeting the primary endpoint (OR, 1.04; 95% CI, 0.60-1.81) or positive conclusions (OR, 1.84; 95% CI, 0.92-3.72).
The researchers noted that, among the studies with medical writers, 17.7% had writers who worked for the company sponsoring the study, 70.9% had writers who worked for a medical writing company, and 5.7% had writers who worked for a hospital or other group.
“These findings suggest that journals need to give more scrutiny to studies with medical writers, and that authorship needs to be properly acknowledged,” the researchers wrote.
Disclosures: This research was supported by Arnold Ventures LLC. The study authors disclosed relationships with Arnold Ventures and other companies. Please see the original reference for a full list of disclosures.
Buck E, Haslam A, Tuia J, Prasad V. Frequency and characteristics of trials using medical writer support in high-impact oncology journals. JAMA Netw Open. Published online February 1, 2023. doi:10.1001/jamanetworkopen.2022.54405
This article originally appeared on Cancer Therapy Advisor