|
Abstract:
|
Competing risks data are frequently encountered in chronic diseases such as cancer. For example, individuals undergoing surgery and subsequent treatment may experience recurrence near the removed tumor, metastatic recurrence at other anatomic sites, occurrence of a second cancer of another type, or death resulting from non-cancer causes that precludes any cancer event. In these cases, conventional survival analysis methods fail to reveal a complete picture of the disease and any effects of treatment. The presence of competing risks also presents challenges in clinical trial design, including choice of target population, appropriate endpoints and effect measures, and attendant statistical tests. In this talk, the use and interpretation of current methods will be reviewed, followed by a proposal to apply an alternative quantity, the restricted mean life extension or loss according to cause-specific failure type, to evaluate benefits and risks of treatment in a clinical trial. Developmental work, simulation studies, and data examples will be presented.
|
