Presentation and subsequent publication rates of phase I oncology clinical trials.

BACKGROUND

Many trials submitted to scientific meetings are not reported in peer-reviewed journals. Results may vary substantially and the lack of final publication constitutes a form of reporting bias. The authors sought to determine the presentation and publication rates of Phase I trials submitted to a major oncology meeting and the factors impeding their subsequent publication.

METHODS

The authors identified all Phase I studies submitted to the annual meeting of the American Society of Clinical Oncology in 1997, categorizing them as novel (agents not approved by the Food and Drug Administration at the time of submission) or nonnovel (at least one agent approved), and as industry sponsored or not. MEDLINE searches and an E-mailed questionnaire confirmed publication in peer-reviewed literature and the reasons for nonpublication.

RESULTS

Approximately 54% of the 275 Phase I studies were selected for presentation. Abstracts reporting novel agents were more likely selected for presentation than those reporting nonnovel compounds (68% vs. 38%; P < 0.0001). Seventy-two percent of the presented abstracts were subsequently published, compared with 62% of those not presented (P = 0.08). The overall publication rate was 67% at 7.5 years. Presentation status was associated with time to publication (P = 0.01), with abstracts chosen for presentation being published sooner. The median time from presentation to publication was found to be 3.4 years. Lack of time and author relocation were the major obstacles to publication.

CONCLUSIONS

The underreporting of final results of Phase I oncology trials remains a serious problem. In the future, investigators must commit to the publication of final results in a timely manner. Journals should provide mechanisms for rapid reporting of Phase I trials.