|
Activity Number:
|
141
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #307270 |
|
Title:
|
The Causal Effect of Finasteride on the Severity of Prostate Cancer
|
|
Author(s):
|
Bryan Shepherd*+ and Mary W. Redman and Donna P. Ankerst
|
|
Companies:
|
Vanderbilt University and Southwest Oncology Group and University of Munich
|
|
Address:
|
1161 21st Ave., S., S2323 MCN, Nashville, TN, 37232-2158,
|
|
Keywords:
|
causal inference ; sensitivity analysis ; counterfactuals ; principal stratification
|
|
Abstract:
|
The Prostate Cancer Prevention Trial found that finasteride lowered the risk of prostate cancer. However, those diagnosed with cancer in the finasteride arm tended to have worse cases of cancer than those with cancer in placebo. Did finasteride cause more severe cancer or was this due to post-randomization selection bias? We answer this question by estimating the mean treatment difference in prostate cancer severity for those who would have had cancer regardless of treatment assignment. We consider assumptions that allow identification of this estimand: conditional on diagnosis of cancer, independence of cancer severity and incidence of cancer if randomized to other arm; monotonicity; and outcomes missing completely at random. We perform sensitivity analyses by systematically relaxing these assumptions using plausible ranges of sensitivity parameters and relevant covariates.
|
