2012 HSR&D/QUERI National Conference Abstract
2012 — An Introduction to Conducting and Interpreting Meta-Analyses
White DL, and Hysong SJ, Houston VA HSR&D COE;
Meta-analysis has become a powerful and widely used tool for systematically reviewing and quantitatively summarizing research findings. Policy makers increasingly rely on meta-analytic findings to inform clinical decision-making and health policy guidelines. The workshop’s purpose is teaching participants how to conduct and interpret a meta-analysis. First, we will briefly overview applied methods for conducting systematic literature searches and efficiently extracting data. We will concentrate on applied statistical methods for pooling data, quantification of between-study heterogeneity, tests for bias and sub-group effects, presentation of findings (statistical/graphical), and result interpretation. Our goal is to teach health services researchers how to conduct meta-analyses for optimal impact and usability to policy makers and other researchers. The instructors for this workshop are highly experienced in conducting and publishing meta-analyses in clinical epidemiology and psychology.
The workshop will be a combination of didactic presentation and interactive exercises. The following topics will be covered: 1) Overview of sources/strategies to validly and reliably identify all relevant research studies, and to efficiently extract relevant data (examples and searchable data sources provided); 2) Demonstration of software programs available for conducting meta-analysis (STATA including syntax and Comprehensive Meta-Analysis, also lists of other shareware and commercial software); 3) Applied examples of how to enter data extracted from studies in software usable format, how to interpret statistical tests of heterogeneity and decide what pooling method to use, and how to obtain summary estimates for common categorical and continuous outcome variables (e.g., odds ratios, Cohen’s d); 4) Assessing the reliability and validity of your findings (tests for publication bias and identifying sources of between-study heterogeneity); and 5) Controversial issues in meta-analysis (e.g., quality scores, non-peer reviewed articles). Specific exercises will teach attendees to: 1) Understand applied steps in obtaining pooled estimators and assessing their reliability and validity including statistical and graphical assessment; and 2) Practice interpreting meta-analytic findings. Participants will receive a list of relevant software packages/useful readings/weblinks for additional information.
Researchers interested in conducting meta-analyses or appropriately interpreting and using meta-analytic findings.
Assumed Audience Familiarity with Topic:
We will assume little/no knowledge of meta-analytic techniques.