Timeline Prediction for Major Cardiovascular Trials

• Activity Number: 458
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Hui Quan*+ and Xuezhou Mao and Yujun Wu
• Companies: Sanofi and Sanofi and Sanofi
• Keywords: hazard rate ; conditional probability ; study duration ; intent-to-treat

Abstract

Cardiovascular (CV) event trials are long term, large scale and costly trials. Since the study power for an event trial depends on the total number of events, it is important to use all information including blinded information acquired during the study to continuously predict the timeline of reaching the target number of events at different study stages. The sponsor thus can timely optimize resources and modify the new drug development strategy. In this paper, we consider the methodology for such a timeline prediction. The background hazard rates of different time intervals used for the prediction are based on the observed internal blinded rates and the external published rates. The conditional probability of having an event by a specific time point or the common end of the study given the observed information of a patient can be derived. With the total sample size for the study, the total number of events as the sum of the conditional probabilities across all patients is an increasing function of the overall study duration. Hence, the timeline for reaching the target number of events can be predicted. A CV trial example is used to illustrate the application of the methodology.