the 23nd Annual Workshop in Applied Statistics
Detailed
information, including the registration form, is available here
Analysis of Right Censored Time to Event Data
Scott Emerson, M.D., Ph.D.
Professor of Biostatistics, University of Washington
Saturday, May 1, 2004
8:45-3:30pm
at
California State University, Long Beach
College of Business Administration Building, Room 140
Abstract
The analysis of data measuring time to event is often complicated by
incomplete observations: Some subjects have not yet had an event at the time
of data analysis. A wide variety of statistical methods have been developed
for use in this setting of “right censored data”, including parametric and
semiparametric regression models, as well as a broad array of nonparametric
methods. In this short course, I review the scientific and statistical issues
involved in the analysis of right censored data and provide an overview of
the spectrum of analytic methods available. I then focus on the nonparametric
analysis of censored time to event data in the presence of nonproportional
hazards. In particular, I consider a class of weighted logrank statistics,
and contrast the use of those statistics with the behavior of statistics based
on the weighted difference of survival curves. I then explore the effect that
a varying censoring distribution has on the statistical operating characteristics
of these methods when applied in sequential sampling.
Click here for detailed abstract (pdf).
Professor Emerson
Scott Emerson, Professor of Biostatistics at the University of Washington,
is an active researcher in the design, monitoring and analysis of group sequential
trials. A major focus of his research has been in the use of estimation methods
to examine the robustness of inference in the setting of poorly specified
stopping rules. Currently, much of his research focuses on the use of nonparametric
methods which are robust to model misspecification when used with a sequential
sampling plan, especially in the setting of longitudinal studies. He serves
on a number of government and industry sponsored advisory boards related to
the design and monitoring of clinical trials. Computer programs that he developed
for his research into group sequential methodology now form the backbone
of S+SeqTrial, an S-Plus module for group sequential trial design.
Dr. Emerson is an accomplished teacher, having taught formal courses in
statistical applications, methods, and theory at the Universities of Washington,
Arizona, Florida, and Virginia. Students in those courses cover the range
of majors and nonmajors at both the graduate and undergraduate level. In
1999, Dr. Emerson received the Outstanding Teaching Award from the School
of Public Health and Community Medicine at the University of Washington.
He has taught a number of 1 and 2 day public short courses on group sequential
trial design in the United States, the United Kingdom, Switzerland, and Australia,
as well as a 2 day private course for the U.S. Food and Drug Administration
(FDA).
More
information, including registration forms, schedule, map and directions.