Systematic reviews are concise summaries of the best available evidence from studies that address well defined practice questions. They are referred to as secondary forms of research because they gather, analyse and collate findings from a number of primary studies. Systematic reviews differ from traditional narrative reviews because they undertake an exhaustive search for all relevant research in relation to a predefined question, according to pre-stated inclusion and exclusion criteria, including specifying which kinds of studies will be accepted (e.g. only RCTs; only studies involving individuals of a certain age group). This is done in an attempt to minimise the likelihood of reviewers including, or excluding studies that they personally favour or reject. This does not mean that value judgements can be totally isolated from the review process, but compels reviewers to be explicit about the procedures used in creating the review, giving readers enough information about studies so they can gain a good understanding of the original research

Systematic reviews can highlight whether specific interventions work best with particular populations and illustrate any gaps in a field of study or recurrent methodological flaws in existing research. Busy professionals can benefit from such reviews, which present them with a collated, appraised set of findings from a wide range of studies. Similarly, service providers can take advantage of these types of investigation when assessing how best to employ available resources. Like all reviews however, findings are applicable to large populations rather than any given individual; professional judgement is also needed to establish to what extent any given person would benefit.

A systematic review has to rely on the findings from good quality primary studies for inclusion, which can often act as an obstacle to this method of investigation. Although randomised controlled trials (RCTs) are considered to produce the most robust findings in terms of the effectiveness of interventions, a paucity of such designs may call for the use of other research methods to provide answers to a question . When other designs are included in a systematic review, the attributive confidence associated with the results from such methods needs to be established. Reviewers must assess the quality of a study by examining how far its design, content and analysis have helped or hindered the minimisation of bias. It should be also noted that methodology is not the only criteria of reliability - well conducted studies using less robust designs will usually be of more use to use than a poorly conducted RCT.

Meta-analysis

One way of examining data from a systematic review is via a meta-analysis, which involves collating all the data from separate studies and treating the results as a single figure. C ombining all available data increases the power of the review to detect important intervention effects which may be less noticeable "“or may be misleading "“ in any single study.