|
Activity Number:
|
430
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #303544 |
|
Title:
|
Sample Size and Power Analysis for PG and PGB Trials Using Parametric Tests
|
|
Author(s):
|
Mallikarjuna Rettiganti*+
|
|
Companies:
|
The Ohio State University
|
|
Address:
|
1958 Neil Ave, 404 Cockins Hall, Columbus, OH, 43210,
|
|
Keywords:
|
PG trials ; PGB trials ; Negative Binomial Distribution ; Likelihood Ratio Test
|
|
Abstract:
|
Parallel Group (PG) trials and Parallel Group trials with baseline correction (PGB) are two of the most common clinical trial designs. They are also very popular in studying the disease progression of relapsing remitting multiple sclerosis (RRMS) using MRI brain lesion counts. Sormani et al (1999) showed that the negative binomial distribution provides a good fit for MRI lesion count data in patients with RRMS. Sormani et al (1999) and Sormani et al (2001) have used the negative binomial model to simulate MRI data but used nonparametric tests for comparison of groups. In this paper, we simulate data using the NB model and use a likelihood ratio test (LRT) which makes use of the parametric assumption to the fullest extent. This results in 38--42% smaller sample sizes for PG trials and 70--80% smaller sample sizes for PGB trials when compared with the ones obtained by Sormani et al (2001).
|
