Prediction of factors affecting clinical trial accrual at Cancer Therapy and Research Center, San Antonio.

Abstract

e18207 Background: Oncology is a very rapidly emerging science with need for more investigational drugs. It is highly essential to accrue patients into clinical trials to develop new drugs. According to the available data, about 38 % of the CTEP/NCI supporting oncology trials close due to inadequate accrual. We here by present a retrospective analysis of factors which affect clinical trial accrual in our institution. Methods: Eligible studies included treatment and supportive care intervention studies either open or closed to accrual between 09/1999 and 012/2015 at our center. Data abstracted includes coordinating group, sponsor type, presence of competing trials, trial phase, disease category, single institutional study, author, primary purpose of the study, interventional modality, targeted therapy, advanced disease, randomization, presence of placebo, rare cancer category and new investigational agents. Successful clinical trial accrual was defined as accruing 4 or more patients. Statistical analysis was performed in the statistic software R (R Foundation for Statistical Computing, Vienna, Austria). We used Chi-square test and Fisher’s exact test. The significance level was set to alpha = 0.05. Results: Retrospective univariate analysis at our institution showed that the variables significantly associated with clinical accrual are Authored by principal investigator, type of sponsor, phase of the trial and number of interventions in the study. Using a random forest model with the above mentioned predictors, we computed an AUC of 64.1 % . We have also developed a histogram of predictions which has identified atleast 5 studies with an 80% failure probability to accrue. Conclusions: Clinical trial accrual is absolutely necessary for development of investigational drugs and progress of research in Hematology/Oncology especially with the rise in incidence of many malignancies. We have identified a few factors that may impact accrual in our institution, and we propose to prospectively validate it in our future endeavor. The major limitations of our study are its retrospective nature and being a single site analysis.