How to systematically review the medical literature.

In contrast to traditional narrative reviews, systematic reviews are true hypothesis-driven research. Meta-analysis is a form of systematic review in which studies are selected and combined by use of a predefined protocol to reduce bias and subjectivity. A sensitivity analysis shows how results vary through the use of different assumptions, tests, and criteria. The most valid synthesis of information occurs when published and unpublished materials are subjected to the same rigorous evaluation and when results are calculated with and without unpublished sources of data. A good systematic review captures the reader's attention through a skillful blend of numbers and narrative and qualifies for publication as original research in a peer-reviewed journal. Otolaryngologists have published systematic reviews of varying quality since 1990. This article should help improve the quality and validity of future efforts.