Purpose: Many phase III trials presented at annual meetings of the American Society of Clinical Oncology (ASCO) remain unpublished. The results of these unpublished trials, if more generally known, might have an impact on clinical practice.
Methods: Abstracts of large phase III trials evaluating systemic cancer treatment were identified from conference proceedings of the 1989 to 2003 ASCO annual meetings. PubMed, Medline, and Embase were searched for corresponding publications. A compendium of unpublished phase III trials was assembled. Clinical significance of nonpublication was determined by disease site-specific oncology experts from two academic cancer centers in Canada.
Results: A total of 709 phase III trials were identified of which 66 (9.3%) remain without a subsequent publication at a minimum of 6.5 years of follow-up and 94 (13%) were published after a delay of ≥ 5 years from their initial presentation. Of the unpublished trials, 48% were presented as oral sessions at an ASCO meeting, and 71% of the abstracts reported negative results. The experts judged that 70% of the unpublished trials addressed an important question and 59% might have had clinical impact if their results had been published promptly.
Conclusion: A substantial number of cancer clinical trials with potential influence on clinical practice remain unpublished and many other trials are published after a substantial delay. Nonpublication of clinical trials breaks an implicit contract with participants, institutional review boards, and sponsors.