Statistical Methods for Observational Studies Seminar Series announcement!The Ohio State University Center for Clinical and Translational Science (CCTS) is pleased to announce an upcoming seminar in the Statistical Methods for Observational Studies Seminar Series.
Presenter: Bo Lu, PhD Presented by: Design, Biostatistics and Ethics Support and Training (DBEST) Friday, May 15, 2009 Randomized experiments are generally considered the gold standard in estimating causal effects of treatments. However, it is not always possible to conduct randomized study due to a variety of reasons. Therefore, knowledge from observational studies contributes significantly in terms of estimating causal effects. In observational studies, the assignment of patients into different treatment groups is not determined by a randomization mechanism. Consequently, treatment effects estimation are erroneous without properly adjusting for selection bias. Propensity score based approach has been increasingly popular in dealing with selection bias in observational studies. Under the potential outcome framework, adjustment based on propensity score is shown to remove the selection bias due the observed covariates. Analytically, propensity scores could be used in matching, stratification or as sampling weights. All three methods are reviewed and an illustrative example of an Italian smoking cessation study is presented. Dr. Bo Lu is an assistant professor in the division of Biostatistics, College of Public Health at The Ohio State University. He received his doctoral degree in Statistics from the University of Pennsylvania. Dr. Lu’s major research interest is causal inference in observational studies with propensity score based approach. He has extended propensity score matching into complicated study designs, including multiple treatment groups, time-varying treatment assignment, etc. His other research interests include design and analysis of large public survey and statistical analysis with missing data.
|
 |
