Abstract
PURPOSE: The aim of this study was to evaluate the rate of early trial discontinuation of oncology trials and reasons for early termination, to assess potential trends in rates of oncology trial termination, and to perform a comprehensive analysis of predictors of early termination. This study intends to inform efforts in improving efficiency of the oncology clinical trial enterprise. METHODS: We conducted a cross-sectional study of interventional cancer clinical trials registered in ClinicalTrials.gov database from September 27, 2007 to June 30, 2015, with at least one site listed in the United States. We evaluated predictors of early trial termination using Fisher exact or χ(2) tests and logistic regression. RESULTS: Of 8687 trials, 22.74% (n = 1975) were terminated trials. Rates of early trial termination appeared stable over the study. Statistically significant univariate predictors of early termination for any reason include cancer category, phase, funding source, location, and age. In multivariable analysis, trials spanning multiple cancer categories and international trials were less likely to terminate early whereas phase 2 trials and trials funded by academia/foundation were more likely to terminate early. The most common reason for early termination was "Other, Multiple Reasons, or Unknown" (36.9%), followed by accrual issues (34.5%). In multivariate analysis among all terminated trials, supportive care trials, phase 2 trials, and non-industry funded trials had significantly higher odds of trial discontinuation specifically due to poor accrual. CONCLUSION: In this national sample of cancer clinical trials, early trial discontinuation was common. Many factors influenced early trial termination with poor accrual being a common reason. Specific trial features are associated with differential likelihood of early trial termination for any reason and for early trial termination due to poor accrual.