"Meta-analysis: statistical alchemy for the 21st century": discussion. A plea for a more balanced view of meta-analysis and systematic overviews of the effect of health care interventions.

The paper discusses some of the most common criticisms to meta-analysis presented by Professor Feinstein in this Conference. As many of the points raised in his contributions are not new, a critique to them is presented in the context of the type of contribution given by systematic reviews (meta-analysis) to the analysis of the effects of health care interventions. After discussing some terminological issues, the paper challenges Feinsteins' arguments indicating that meta-analysis is inherently faulted on four grounds: (a) reproducibility, (b) precision, (c) suitable extrapolation, (d) fair comparison. Each point is discussed providing examples drawn from the published literature with a view to indicate that--despite their current limitations--systematic reviews are a necessary step to synthesize information, orient clinical research and help produce practice guidelines.