Artikel
Sample size calculations for controlled clinical trials with clustered data using the SAS macro GEESIZE
Suche in Medline nach
Autoren
Veröffentlicht: | 14. September 2004 |
---|
Gliederung
Text
Several approaches for sample size and power calculation in studies with correlated response data based on generalized estimating equations (GEE) have been proposed recently since these trials require fewer samples than classical methods. However, most of the proposed methods are very specific and not as general as required for designing clinical trials. Therefore, we have extended the SAS macro GEESIZE. Specifically, we have added option of an independence working correlation matrix. Additionally, we have reformulated the hypotheses in order to use classical coding schemes (for details see the examples on the web page). We exemplify the flexibility of the enhanced SAS program and investigate the calculated sample sizes in several situations. We show that the software is a flexible tool for sample size calculation purposes in clinical trials with correlated data.