American Statistical Association
Over the last decade, mental health researchers have made widespread use of hierarchical designs (i.e., clustered and/or longitudinal) in randomized clinical trials as well as observational studies to investigate associations between treatment and outcome. Although the statistical methods to analyze the hierarchical data are in routine use, but the design of such studies often rely on sample size calculations which are inadequate. In this talk, I will explore recent statistical results to provide guidance on sample size determination for both continuous and binary longitudinal data in order to achieve the target power while strictly maintaining the Type I error rate to its nominal value (e.g. 5%). The complexity of the problem of sample size determination depends on 1) the hierarchy of design, 2) type of proposed model, 3) hypothesis testing procedure, 4) level of randomization and 5) differential attrition rates. I will also discuss the software RMASS which computes sample size for both two and three-level mixed-effects linear regression models for the analysis of longitudinal data. I will end this talk with some of the practical issues in designing the studies and provide few examples from psychiatric research.
Kush is currently working at Cooperative Studies Program Coordinating Center, Hines Veteran Affairs hospital as a Biostatistician. He graduated with a doctoral degree in Biostatistics from the University of Illinois at Chicago in July 2010. He research focus during past 7 years has been in the areas of sample size determination problems for hierarchical models, hypothesis testing problems for skewed data, and estimation, modeling and testing problems relating to Functional Magnetic Resonance Imaging (fMRI) field. His publications have appeared in notable journals in both physical and health sciences fields such as Technometrics, Statistics in Medicine, Statistical Methodology, Archives of General Psychiatry, Communications in Statistics and Archives of Dermatology.
|Date:||Tuesday, November 20,2012|
|Time:||3:30 - 4:30 P.M.|
New York State Psychiatric Institute
1051 Riverside Drive
Pardes Building, Multipurpose Room 6602
New York, New York