|
Activity Number:
|
75
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306278 |
|
Title:
|
Minimum Sample Size in Control Group When Comparing Efficacy Rate with Several Treatment Groups
|
|
Author(s):
|
Alan Davis*+ and Inder J. Sharma
|
|
Companies:
|
Pharmanet and Sharma Associates, Inc.
|
|
Address:
|
10150 Mallard Creek Road, Charlotte, NC, 28262,
|
|
Keywords:
|
odds ratio ; Dunnett's ; clinical trial ; power ; alpha ; simulation
|
|
Abstract:
|
In evaluating a treatment in a clinical trial, it is often necessary to compare the success rates in several treatment groups against a single group used as a placebo control. Sometimes it is not practical to enroll many patients in a control group, due to the ethical necessity of providing treatment. This can provide a significant loss to information that can be gathered, particularly in consideration of the desirability of evaluating several treatment groups against the control. This study will evaluate the cost in terms of power and alpha of reducing the number of patients in a control group relative to several treatment groups. The method of determining significance will use Dunnett's test extended to the evaluation of odds of success for 3-5 groups (k) where one group is a control or reference group and the outcome is a binomial used to report drug efficacy.
|
