Extrapolating Survival Curves in Clinical Trials Beyond Follow-Up Periods

• Activity Number: 530
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Jerry Cheng*+ and Javier Cabrera and John Kostis and David Madigan
• Companies: Robert Wood Johnson Medical School, Rutgers University and Dept of Statistics, Rutgers University and Robert Wood Johnson Medical School, Rutgers University and Columbia University
• Keywords: Survival Curve ; Clinical Trials ; Extrapolation

Abstract

In clinical trials, participants routinely go to clinics for check-ups and tests so that important clinical data are collected for analyses. Due to budget limitations, most clinical trials have a fixed length of follow-up periods, beyond which vital information especially the mortality data are not available. Generally, long follow-up periods allow estimation of the fraction of long-term survivors and increase the precision of the estimated survival curve. However, gains in precision may be offset by the cost and time associated with additional follow-up and accrual periods. As a result, a difficult question with clinical trial design arises as to how long to continue following patients and whether to terminate accrual. In this presentation, we develop comprehensive and novel statistical approaches to more accurately predict long-term survival of the participants in clinical trials beyond follow-up.