Meta-analysis: Conceptual bases, statistical analysis and interpretation

Abstract

The term meta-analysis was first used by G.V. Glass in 1976 in an article called "Primary, secondary and meta-analysis of investigations" ("Primaria, secundaria y meta-análisis de la investigación"). He used this term to refer to the statistical analysis of all the results obtained in different clinical studies regarding the same subject and that were to be analyzed together. At the beginning, this type of analysis was mainly used for the examination of social studies and psychology investigations, but later, during the 1980s, it became a popular method used in medicine; particularly in the cardiovascular, cancer and perinatal specialties. Nowadays it is not rare to find several medical articles using this method. For this study we decided to perform a meta-analysis and also combine the results of the studies because when the sample size increases, the statistical potential increases as well. Furthermore, when including studies and researches performed in different Health Centers, the results obtained can be easily generalized. Nevertheless, the meta-analysis method has its controversies, many of them due to an excessive use of the method together with a lack of methodological rigor; there are many limitations to be considered when evaluating the results of a meta-analysis. In this article we will focus mainly on the statistical aspects of the subject, from the point of view of the description of the methods, indications and interpretations, without specifying other details such as the protocols for carrying out a systematic review or statistical formulas. The focus will be on the analytical methods used in meta-analyses of controlled clinical trials evaluating therapeutic efficacy or adverse reactions.