SPEAKER: Dr. Thomas Lumley,  Assistant Professor of Biostatistics, Department of Biostatistics, University of Washington, Seattle

TITLE: Network Meta-Analysis for Indirect Treatment

ABSTRACT:

Randomized trials provide the most reliable evidence of the relative effectiveness of medical treatments, so summarizing the results of many randomized trials is important in evaluating treatments.  Traditionally this 'meta-analysis' compares just two treatments and the summary is simply a weighted average of the trial results.  When there are multiple interventions available, as for hypertension, there may be relatively few randomized trials comparing two particular treatments, but many trials comparing one of those treatments to something else.  I will describe a method for incorporating these indirect comparisons in a meta-analysis, and its origin in the graph structure of the collection of all randomized trials for a given condition.
 
 

COFFEE: 3:45 p.m., 104 Snedecor Hall