|
Activity Number:
|
110
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305803 |
|
Title:
|
Testing Noninferiority for Clustered Matched Pair Studies
|
|
Author(s):
|
Jun-mo Nam*+
|
|
Companies:
|
National Cancer Institute
|
|
Address:
|
, , ,
|
|
Keywords:
|
adjusted score test, ; adjusted Wald-type test ; clustered matched pair ; moments estimate ; non-inferiority ;
|
|
Abstract:
|
Non-inferiority tests and sample size formulae are well established for matched pair studies where pairs are independent. When matched pairs are clustered, these methods are not applicable since the pairs in a cluster are likely positively correlated. Durkalski et al presented an appropriate non-inferiority test statistic based on moments estimate for clustered matched pair data. In this paper, we propose alternative non-inferiority test statistics, and their performance in Type 1 error rate and power are assessed by using simulations. We examined a relative importance of a number of clusters and cluster size for designing a clustered matched pair study. A total number of pairs required for a specific power in clustered pair study is greater than that one in completely independent pair study design and an increase is positively related to an intra-class correlation.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
