Predicting Low Accrual in Cooperative Group Oncology Trials

• Activity Number: 377
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 11:15 AM
• Sponsor: Health Policy Statistics Section
• Author(s): Caroline Bennette* and Scott D. Ramsey and Cara L. McDermott and Josh J. Carlson and Anirban Basu and David L. Veenstra
• Companies: University of Washington and Fred Hutchinson Cancer Research Center and University of Washington and University of Washington and University of Washington and University of Washington
• Keywords: prediction modeling ; accrual ; clinical trials ; oncology

Abstract

Background: Whether and how trial-level factors differentially influence accrual to individual trials has not been comprehensively studied. We sought to study the empirical relationship and predictive properties of putative risk factors for low accrual among NCI's Clinical Trials Cooperative Group (CTCG).

Methodology: Data from 787 phase II/III adult CTCG trials that opened 2000-2011 were used to develop a logistic regression model to predict low accrual, defined as trials that closed with or were accruing at < 50% of target; 46 trials opened 2012-2013 were used for validation. Model performance was evaluated by calibration and discrimination via the area under the curve (AUC).

Results: The final model included 12 risk factors and exhibited good calibration and discrimination to predict trials with low accrual (AUC in trials launched in 2012-2013: 0.730; 95% CI: 0.584, 0.877).

Conclusions: Prospectively, a predictive model based on multiple trial-level risk factors could help quantify accrual feasibility and thereby aid future trial selection and prioritization decisions by targeting limited resources to support accrual where it is most needed.