Methods of meta-analysis: an analysis

To understand the principles of systematic reviewing and meta-analysis, using recent examples from the medical literature to highlight some of these points.

The word 'meta-analysis' is an intimidating one, and its associated jargon makes it seem incomprehensible. Actually, it is only a mathematical maneuver to add up data in a systematic review; it might be better called 'meta-addition'. The systematic review is a process of using the best available evidence to answer a particular clinical question. Data combination (usually done with meta-analysis) increases the power to see small differences and makes a more precise estimate of a treatment effect. Its major drawback is heterogeneity (the proverbial problem of adding apples and oranges). Systematic reviews have been used by medical societies to create position statements. Such statements have suggested that parenteral nutrition is far less efficacious than previously believed. Systematic reviews in some areas of nutritional support have clarified type II errors. Problems exist, however, in a number of the published meta-analyses of aspects of this therapy.

Especially in an era of resource restraint, we need to become more skilled at interpreting evidence from clinical research.