Activity Number:
|
213
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 13, 2002 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Biometrics Section*
|
Abstract - #301321 |
Title:
|
Looking for Treatment Effect in Genetic Sub-studies of Randomized Trials: Do We Need to Genotype Controls?
|
Author(s):
|
Samuel Wieand*+ and Mary King and Thomas Richards and Joseph Costantino
|
Affiliation(s):
|
University of Pittsburgh and University of Washington and University of Pittsburgh and University of Pittsburgh
|
Address:
|
Suite 325 - Sterling Bldg - 201 N. Craig Street, Pittsburgh, Pennsylvania, 15213-1567, USA
|
Keywords:
|
imbedded case-control ; genetic substudies
|
Abstract:
|
When randomized trials are conducted to determine whether an experimental treatment is effective in preventing a disease (such as cancer), tissues and/or blood samples are often collected to allow for subsequent genetic testing of the study participants. When the trial is complete, secondary, retrospective genomic analyses are performed to determine whether the experimental treatment is effective in a subset of participants with specified genotypes (the target subgroup). For large studies involving extensive genomic analysis of every test sample, practical issues may limit the ability to carry out genetic testing in participants who did not develop the disease (controls). We show that to develop valid statistical tests for evaluating the treatment effect in the target subgroup, one need genotype only the patients with disease (cases). The loss of efficiency associated with failing to genotype controls is shown to be small in most practical situations.
|