Changes to primary endpoints of phase 3 cancer trials common but ‘markedly underreported’
Click Here to Manage Email Alerts
Key takeaways:
- Approximately 19% of clinical trials evaluated had primary endpoint changes detected.
- More than 70% did not disclose such changes in the manuscript.
Primary endpoints of randomized clinical trials for cancer therapies frequently change after trial initiation, according to findings published in JAMA Network Open.
Nearly 20% of evaluated trials had primary endpoint changes detected, and primary endpoint changes were “markedly underreported” within study manuscripts, researchers wrote.
“We hope [these] data highlight the need for transparency and clarity throughout the clinical trials process and provide evidence supporting complete disclosure of clinical trial protocols with published results,” Ethan B. Ludmir, MD, assistant professor of gastrointestinal radiation oncology at The University of Texas MD Anderson Cancer Center, told Healio. “Protocol disclosure should include a record of changes to the protocol from trial initiation through completion.”
Background and methods
Changes to primary endpoints during active clinical trials may increase risk for outcome reporting bias or call trial quality into question, according to study background.
Ludmir and colleagues conducted a cross-sectional study, evaluating publicly available data for completed oncology phase 3 randomized clinical trials.
The analysis included 755 trials registered in clinicaltrials.gov from inception through February 2020.
Changes between the initial primary endpoint and the final reported primary endpoint served as the main outcome.
Researchers used three methods to determine primary endpoint changes: history of tracked changes in clinicaltrials.gov, self-reported changes noted in the manuscript and changes reported within the protocol, including all available protocol documents.
Key findings
Researchers determined 145 (19.2%) of the evaluated trials had primary endpoint changes.
Of those 145 trials, 102 (70.3%) did not disclose primary endpoint changes within study manuscripts.
Researchers noted significant variability in rates of primary endpoint detection via each method (chi square = 72.1; P < .001).
Investigators reported higher rates of detected primary endpoint changes when they had access to multiple versions of the protocol (31.8%; 47 of 148) than when they had access to one version (16.4%; 22 of 134) or no versions (16.1%; 76 of 473).
Multivariable analysis revealed an association between primary endpoint changes and trial positivity (OR = 1.86; 95% CI, 1.25-2.82).
Next steps
The “significant discrepancies” observed in the rates of identified primary endpoint changes demonstrated a need for greater transparency in reporting changes within trial protocols and study manuscripts, researchers concluded.
“Further efforts are needed at a granular, individual-trial level to understand the rationale and impact of primary endpoint changes on trial outcomes,” Ludmir told Healio.
For more information:
Ethan B. Ludmir, MD, can be reached at The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Unit 1050, Houston, TX 77030; email: ebludmir@mdanderson.org.
Collapse
Click Here to Manage Email Alerts