Biochemia Medica, Vol. 19 No. 2, 2009.
Other
Describing and interpreting the methodological and statistical techniques in meta-analyses
Alfred A. Bartolucci
; University of Alabama at Birmingham, Birmingham, Alabama, USA
Abstract
The term, meta-analysis, has become a regular description for interpreting results from the scientific literature and hopefully reaching a definitive conclusion concerning the reasonableness of results of studies investigating similar interventions, techniques or treatments. It has become a necessity for understanding the sometimes huge repository of conflicting, inconsistent or inconclusive literature or raw data on a subject and understanding the true relevancy of the statistical results in addressing a scientific question such as efficacy. One has to understand, not only the motivation for a meta-analysis, but the techniques employed to help synthesize results from different studies giving them appropriate statistical scrutiny and interpreting the conclusions. A particular challenge is the differing designs of studies, differing eligibility criteria for entry into the study, especially in the medical clinical trial and possibly non similar approaches in interpreting the results, although this latter point is not too prevalent. For the most part studies investigating the role of a particular intervention such as lowering cholesterol and the impact on cardiovascular health address a similar question and should studies not all agree on the results, the role of meta-analysis is to hopefully put all the results together and reach a definitive conclusion about the positive, negative or inconclusive impact on lowering lipid levels and cardiovascular health. One should appreciate an overview of the approach to meta-analysis assuring mostly the understanding of the statistical approaches, their meaning, purpose and ultimate impact on the interpretation of the meta-analysis. Thus having been exposed to the elements of the meta-analysis one should understand the advantages and limitations of this approach to interpreting results from the literature.
Keywords
meta-analysis; statistics; review; bias
Hrčak ID:
37595
URI
Publication date:
3.6.2009.
Visits: 10.658 *