|
Activity Number:
|
481
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #305865 |
|
Title:
|
Statistical Approches To Analyze Censored Data with Multiple Detection Limits
|
|
Author(s):
|
Wei Zhong*+ and Linda Levin and Paul Succop and Rakesh Shukla and Jeffrey Welge
|
|
Companies:
|
ICON Clinical Research and University of Cincinnati and University of Cincinnati and University of Cincinnati and University of Cincinnati
|
|
Address:
|
1700 Pennbrook Parkway, North Wales, PA, 19454,
|
|
Keywords:
|
multiply censored data ; multiple detection limits ; EM algorithm ; meta-analysis method ; lognormal distribution
|
|
Abstract:
|
Censored data with multiple lower detection limits frequently arise when data are measured by different procedures or are combined from multiple laboratories. The substitution method which replaces nondetects with certain functions of the detection limit is a common approach to compute summary statistics for censored data. However, this method lacks a theoretical basis and results differ depending on the substituted value. MLE with the EM algorithm integrated method and the meta-analysis method were introduced for censored data with lognormal distribution and their properties were evaluated through simulation studies. Compared to the substitution methods, simulation results consistently showed that the proposed methods provided the most accurate and efficient summary statistics for multiply censored data and they were able to incorporate the sample collection process into the estimation.
|
