Skip to Main Content

Systematic review guide

A step by step guide to doing a systematic review

Writing and Reporting

Your primary consideration in writing your review, reporting methodology and findings should be your audience - the future readers of your review. These may be other researchers, clinicians, policy and decision makers, but also your co-reviewers, peer reviewers and journal editors. They need to be able to properly judge the quality of your review and it's relevance to them - full and transparent reporting of methods and a recognition of your review's strengths and limitations will enable this. 
Writing the methods section will be guided by an appropriate reporting standard and you may need to consider author guidelines and word counts and style guides from your desired journal of publication at the outset of writing.

Writing a discussion

The key parts of a discussion include:  

Statement of principal findings: a summary of your review's findings. Include all findings even if there were findings of no effect or negative effects. 

Interpretation of the review's findings: discuss the quality of evidence and nature, strength and size of the effects (internal validity), and the applicability of the findings of the review (external validity).
Evidence in context - place your research in the context of wider literature. Are the findings in line or different from previous evidence?
Strength and weaknesses of the review - no review can be completely free of bias. Readers will appraise your review for quality so be upfront and transparent about any limitations and biases as they will be found anyway and it is better to show you are aware of them and any impact they may have on the review.  

Implications for practice: remember implications are not the same as recommendations. Systematic reviews are not a suitable platform for recommending changes in practice unless they are being produced by institutions such as Cochrane or as part of Guidelines. Be clear about the strength of the evidence supporting your statements for implications for practice and which populations, settings or circumstances the evidence does (or does not) apply to.   

Implications for research: recommendations for research is appropriate in a systematic review but avoid the "further research is needed" cliché by being clear on what you think should happen next as a result of your review. Is any further research actually needed - perhaps there are now enough quality studies pointing to the same result. If more research is needed then what type of research? Where are the gaps in the literature? What is the quality of the current evidence? How can research quality in this area be improved? 

Reporting

There are multiple guidelines available for reporting in systematic review and meta-analyses, including specific ones for different types of reviews. PRISMA is the main guidelline used and although it was designed for effectiveness reviews it can be used for many review types. You will need to complete the PRISMA 2020 27 point checklist of reporting items, as well as the flow diagram (tracking the identification and selection of relevant studies).   
There are also specific PRISMA extentsions to use alongside PRIMSA 2020 such as:

  • PRISMA-Search (which should be used routinely to increase depth and accuracy in reporting searches).
  • PRISMA-ScR (scoping reviews)
  • PRISMA-NMA (network meta-analyses) are examples of extensions relating to different types of reviews or methods of synthesis.

Other, non-PRISMA guidelines may also be relevant to your review, for instance SWiM (Synthesis Without Meta-analysis) for reviews with a narrative synthesis.

The sooner you familiarise yourself with the appropriate reporting guidelines for your review the better, making certain all mandatory reporting items are being documented reliably as you go will prevent later issues during the writing stages of the review.
As well as ensuring your review is accurately and transparently reported, the checklists help guide you through the writing process by providing structure and essential points and features to include. Keep the checklists close while writing - you are required to note the page on which each reporting items is present. Peer reviewers will check against this to see if all items have been reported properly to help determine review quality for publication.

Journals publishing reviews will have word, table and figure limits which means methods may need to be succinctly described, but supporting material such as full search strategies, table of characteristics of excluded studies and the reporting checklists themselves can instead be published with the article as appendices or supplementary files. However the PRISMA Flow Diagram should be accessible in the main text of the article.