Date of Award

6-1-2009

Degree Name

Doctor of Philosophy

Department

Statistics

First Advisor

Dr. Michael R. Stoline

Second Advisor

Dr. Joseph W. McKean

Third Advisor

Dr. Rajib Paul

Fourth Advisor

Dr. Hong Liu-Seifert

Abstract

Previous statistical analyses of patient discontinuation in clinical trials have used discontinuation status as the response of interest. These analyses assume that the risks of discontinuation for specific reasons (lack of efficacy, adverse events, other reasons) are independent of each other and that significant risk factors for patient discontinuation have the same effect on the different causes of discontinuation. However, it is possible that the underlying risks of discontinuation for specific reasons could be related and that risk factors for one type of discontinuation could have a very different effect on another type of discontinuation. The competing risks methodology can be applied to test for significant differences between the different risks of discontinuation and determine significant predictors of discontinuation specific to each type of discontinuation. These competing risks methods are applied to real-world clinical trial data in multiple disease states, which have not been applied previously in the analysis of patient discontinuation.

Access Setting

Dissertation-Open Access

Share

COinS