Bayesian Enrollment and Event Predictions in Clinical Trials Leveraging Literature Data

• Activity Number: 66
• Type: Topic Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Social Statistics Section
• Author(s): Aijun Gao*+ and Fanni Natanegara and Govinda Weerakkody
• Companies: InVentiv Health Clinical and Eli Lilly and Company and Eli Lilly and Company
• Keywords: Enrollment predictions ; Bayesian survival models ; Literature data ; SAS ; R

Abstract

Given the high cost of cardiovascular outcomes trials, it is important to accurately predict the number of patients to enroll and duration follow up to reach the targeted number of events. Under-estimation of event rate will result in more patients need to be enrolled with shorter follow up period than anticipated and over-estimation will result in longer follow up, added cost and delay in regulatory submissions and approval. Bayesian survival models will be used in predicting the number of events at different time point based on the event rates from literature and those currently available from the trial itself. Both enrollment data at interim analysis and the time-to-event data in the future are simulated in the predictions. The predicted number of events at different time point and distribution of predicted duration of follow up are presented. The probabilities of reaching targeted number of events at different time points are estimated. The sensitivity of the event rates from the literature on estimated number of events is addressed. All the simulation, analyses and graphs are implemented with SAS and R.