|
Activity Number:
|
381
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #307467 |
|
Title:
|
Use of Life Tables To Extrapolate Survival from Clinical Trial Data
|
|
Author(s):
|
W. J. Hall*+ and Hongyue Wang
|
|
Companies:
|
University of Rochester Medical Center and University of Rochester Medical Center
|
|
Address:
|
, Rochester, NY, 14642-8630,
|
|
Keywords:
|
device trial
|
|
Abstract:
|
Practical needs may require extrapolation of survival curves beyond the limited time span observed in a clinical trial - e.g., in economic forecasting of the potential impact of a new expensive device or therapy. The observation span may not be sufficient to permit modeling of the inevitable effects of aging on future survival. We use US life tables to develop a parametric model for survival in a subpopulation matched by age and sex to the clinical trial. Using the resulting cumulative hazard function as a time transformation, the subpopulation has exponential survival on an accelerated time scale. This facilitates modeling for the control arm of the trial by assuming proportional hazards to the US; modeling of treatment-arm survival can then be done to reflect assumptions about treatment efficacy on future survival. We illustrate with data from a recent device trial.
|
