Combining Indirect Evidence or Evidence from Independent Sources

• Activity Number: 53
• Type: Invited
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: IMS
• Author(s): Ingram Olkin*+
• Companies: Stanford University
• Keywords: meta-analysis ; evidence-based medicine ; multivariate models

Abstract

The "evidence-based" movement in medicine, education, and other fields is based on an assessment of diverse types of data. An integral part is the synthesis of the results of independently conducted studies designed to determine the effectiveness of an intervention. Meta-analysis is the basis of the assessment. because the experimental data is diverse, statistical models need to be developed. Some methods are nonparametric, others are parametric models; some are univariate, others multivariate; some are regressions; some are survival curves,etc. Methods of analysis are also diverse, e.g. some are Bayesian, others are based on likelihoods. A review will be provided of some of the key problem issues, such as various types of bias, as well as a discussion of some of the statistical models and methods of analysis.