|
Activity Number:
|
285
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306216 |
|
Title:
|
Determining Sample Size for Proportions When the Information on a Prognostic Factor Is Unavailable
|
|
Author(s):
|
Bong S. Kim*+
|
|
Companies:
|
Neumann College
|
|
Address:
|
One Neumann Drive, Aston, PA, 19014,
|
|
Keywords:
|
sample size ; power ; proportions ; prognostic factor
|
|
Abstract:
|
Sample size and power for proportions are presented when an important prognostic factor is not observed. We first find the conditional variance of the number of patients with the prognostic factor in the control and experimental groups, and then we find the unconditional variance by integrating over the distribution of the prognostic factor. Thus the variance of the test statistic is larger than it would have been had the prognostic factor been observed, and a large sample size will be required. We show how the sample size and the power vary according to the distribution of the prognostic factor. We illustrate this methodology with a sample size calculation for a clinical trial in which the information on a prognostic factor is unavailable.
|
