When beginning work on a systematic review, the best starting point is to ensure one of two things:
- The question has not been answered by a previous systematic review
- If a systematic review is already in existence on your chosen topic, enough new research has been undertaken since its publication that a further review is warranted.
Consult a librarian to discuss the question, to hone it into an answerable question using the literature. Eligibility criteria for inclusion in the review should be clearly defined from the outset. There should be at least two people reviewing the articles retrieved for eligibility so you may need to recruit a co-author.
Do a scoping search to see if there’s likely to be enough literature for a full review – too little will mean a meta analysis is difficult; too much may mean you should narrow the focus of your question.
If you’re looking at a review on the effectiveness of an intervention, then do register your protocol at PROSPERO – the International Prospective Register of Systematic Reviews. You can also search this database to see what other systematic reviews are underway.
The search needs to be well documented. It is recommeded to use the Preferred Reported Items for Systematic Reviews and Meta-Analysis (PRISMA) standards for reporting the search. You may also wish to look at the Methodological Expectations of Cochrane Intervention Reviews (MECIR) for further guidance.
Choose your databases. As a baseline, you should search at least MEDLINE, EMBASE and Cochrane CENTRAL for an effectiveness review. You should also hand search relevant journals in the clinical field if these aren’t well indexed in the databases, and look for grey literature such as reports, conference abstracts and dissertations.
If you can, involve a librarian at this stage, as they are experts in searching.
Do be aware that a literature will not retrieve all relevant papers if they are not indexed consistently, and finding relevant papers outside of the search strategy does not invalidate it. Just make sure what you do and find is reported and recorded clearly.
Once your search strategy is completed, set it up as a regular alert, so that new papers potentially fitting your review criteria are found as they are published. You will also need to re-run the search a final time just before submitting the review for publication to be sure nothing has been missed.
Selecting papers for inclusion
Abstracts need to be screened indepedently by two reviewers. Use your eligibility criteria to select these into yes/no piles (keep your “no”s as you will need these for the “characteristics of excluded studies”. The number of articles selected should be recorded on your PRISMA flowchart. At this point, it’s a really good idea to use some kind of reference management software to keep track of your papers (UHL have access to RefWorks via the Athens password system).
To retrieve the full text of the articles you choose, contact UHL NHS Libraries and information services.
The papers you select to include in the study will need appraising for methodological quality. There are many different tools available for this, such as CASP UK. You will need a data extraction tool for your paper too, so you’ll to decide what the important points are to compare across the studies you choose to include.
Writing the review
Last of all, the review needs to be written! Proper documentation along the way, and following MECIR and PRISMA guidelines will help. If you’ve had a librarian design the strategy, contribute to meetings, the methods section and have approval of the final draft before submission, then they should be listed as an author of the review. Contributing to the search only should get an acknowledgement in the final paper.
UHL NHS Libraries and information services and the Clinical Librarian Service can help at many stages of the review process.