Description
This workshop will provide a practical introduction of meta analysis. We will show how meta analyses are used to set policy (as in evidence-based medicine), to publish review articles, and as part of planning a new studey. We will review the techniques used to synthesize data for various indices of effect size including means, odds ratio, risk differences, and others. We will discuss variations on these techniques such fixed effects vs. random effects models. We will also discuss some of the controversies that surround the use of meta analysis, such as the relative validity of meta analysis vs. large scale randomized trials. We will discuss methods used to assess heterogeneity among studies, and approaches suggested to deal with this heterogeneity.
This session will strike a balance between theory and practical issues. Toward this end, the theoretical issues outlined above will be addressed with reference to practical examples. In particular, we will show how the Cochrane Database of meta analyses is used to recommend treatment options for patients in any number of clinical fields. We will also show how the computer program Comprehensive Meta Analysis (developed with funding from the National Institutes of Health) may be used to run a meta analysis, and create charts and graphs that summarize the results. Attendees should be familiar with the basic concepts of meta analysis.
|
