Advancing Education, Research, and Quality of Care for the Head and Neck oncology patient.
Introduction: Clinical trials are a critical component of evidence-based medicine, yet many remain unpublished, contributing to bias and research waste. This study sought to determine the rates of and contributing factors to nonpublication of head and neck cancer clinical trials registered on ClinicalTrials.gov.
Methods: Head and neck cancer trials were identified via a Boolean search. Data from 100 head and neck cancer trials listed on ClinicalTrials.gov performed within the last 20 years were included in the study. Trial variables (title, phase, intervention, etc.) were compiled in SPSS. Corresponding publications were found using PubMed, Google Scholar, or by emailing principal investigators.
Results: Of the trials examined, 56 (56%) were never published in a peer-reviewed journal. A robust Poisson regression was performed in SPSS to calculate relative risk of nonpublication between trial variables. Absence of statistically significant results (i.e., p < 0.05), incomplete trial status, and absence of results reported on ClinicalTrials.gov were most strongly associated with nonpublication. Furthermore, a disproportionate number of trials originating from China (both mainland and Taiwan) did not have an associated publication in comparison to trials originating from other countries (82% vs 51% nonpublication rates, respectively). Conjunctive consolidation was performed with all statistically significant variables to generate a gradient of risk of nonpublication based on trial characteristics. This gradient was used to create a model to help predict risk of nonpublication for head and neck cancer trials on ClinicalTrials.gov.
Conclusions: These findings demonstrate that nonpublication of clinical trials, especially those with nonsignificant results, remains a problem and undermines the ability to conduct evidence-based medicine.