|
Activity Number:
|
72
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306414 |
|
Title:
|
Some Issues in Fitting Clinical Count Data with Poisson Regression Model
|
|
Author(s):
|
Abdul Sankoh*+
|
|
Companies:
|
sanofi-aventis
|
|
Address:
|
200 Crossing Blvd., Bridgewater, NJ, 08807,
|
|
Keywords:
|
clinical count data ; goodness of fit ; Poisson regression ; type I error rate
|
|
Abstract:
|
Poisson regression is routinely used to model response rates or clinical count data as a function of covariate levels. However, regulatory agencies are generally concerned about the goodness of fit of Poisson distribution to clinical data. Even when there is a good fit, regulators may still be concerned about the possibility of false-positive findings resulting from under-estimation of standard error. We examine in this presentation the appropriateness of the regular Poisson regression model for the analysis of clinical count data from randomized clinical trials by comparing its performance to other models including the generalized Poisson and negative binomial models regarding goodness of fit, type I error rate control, and power. Simulation results show comparable performance for the regular Poisson model regarding type I error rate control under reasonable sample size.
|
- 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 2006 program |
