|
Activity Number:
|
404
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #301667 |
|
Title:
|
The Minimum Required Number of Clusters for Cluster Randomized Trials
|
|
Author(s):
|
Zhiying You*+ and Gary Cutter
|
|
Companies:
|
The University of Alabama at Birmingham and The University of Alabama at Birmingham
|
|
Address:
|
, , ,
|
|
Keywords:
|
cluster randomized trial ; sample size ; statistical power ; feasibility
|
|
Abstract:
|
There are two sample size considerations in cluster randomized trials (CRTs): the number of clusters per intervention group and the number of individuals in each cluster (called cluster size). We show the unequal contribution of the number of clusters and cluster size to study power. The former is the main determinant though the later plays an important role. We also show the existence of an upper bound of study power for a given number of clusters per group when the cluster size tends to infinity, and thus the existence of a minimum number of clusters per group for a prespecified study power. The minimum required number of clusters per group actually serves as a criterion for the feasibility of CRTs. This is very important especially for trials with small number of clusters of large cluster size. We develop methods to calculate the sample size requirements and the upper bound of power.
|
