Activity Number:
|
538
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Biopharmaceutical Section
|
Abstract - #308933 |
Title:
|
Maximum Likelihood Estimation of the Difference between Correlated Poisson Cough Counts
|
Author(s):
|
Xin Yan*+ and Thomas Bradstreet
|
Companies:
|
University of Missouri-Kansas City and Merck & Co., Inc.
|
Address:
|
206 Haag Hall UMKC, Kansas City, MO, 64110,
|
Keywords:
|
Left truncated Poisson distribution ; MLE ; Correlated Poisson data
|
Abstract:
|
A statistical method is proposed to assess cough-modifying agents using the difference between cough counts before and after treatment. This clinical endpoint provides a direct assessment of efficacy in cough challenge studies. The conventional clinical endpoint is indirect, specifically the change from baseline in the concentration (either C2 or C5) of inhaled food-derived extract that triggers a predetermined number (either 2 or 5) of consecutive coughs in a given subject. Maximum likelihood estimation of the difference between cough counts before and after treatment is proposed assuming a left truncated correlated Poisson distribution. Estimation of the difference using simulated data is presented to illustrate the method. Two S-Plus functions are provided to facilitate calculating the maximum likelihood estimate of the difference and the corresponding confidence interval.
|