Statistics from Altmetric.com
Systematic reviews are increasingly being seen as the optimal source of knowledge for evidence-based practice. A good systematic review will provide an unbiased summary of existing evidence and, provided it is applicable to local patients, should guide clinical practice. Being able to appraise systematic reviews is therefore a crucial skill for emergency physicians.
The use of complex statistical techniques in meta-analysis often distracts the clinician attempting to appraise a systematic review. As previously suggested in this series, complex statistical issues are best left to a statistician. Instead, we should focus upon the many important insights that clinical experience can bring to appraisal.
WHAT IS A SYSTEMATIC REVIEW?
A systematic review is a scientific study. It follows the introduction, methods, results and discussion approach. The conclusion should represent an unbiased synthesis of available data relating to a specific question. It may not be very entertaining to read but, if undertaken properly, will provide an objective answer based upon the best scientific evidence.
A narrative review is not a scientific study. The authors present their opinions of a particular topic with reference to primary studies they have selected. A good narrative review should be interesting, entertaining or provocative, but it should not be considered to provide scientific evidence. The differences between a systematic and a narrative review are summarised in table 1.
STAGES OF A SYSTEMATIC REVIEW
The process of identifying, selecting and assessing studies for inclusion in a systematic review should be open, explicit and objective. Data collection for a systematic review typically involves three stages: (1) literature searching and retrieval; (2) the selection of appropriate papers; (3) quality assessment of selected papers.
These three steps should be based upon explicit criteria and should ideally be carried out by two independent assessors who are blind to each other’s decisions. The review should report the total number of articles identified by …
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.