Guidelines on systematic review
Purpose
Findings from this piece of work lays the background and foundation for, or be related to, the primary research.
It is not possible to establish concrete rules for the review (e.g. the number of hits one must have from literature search), as every review is different and the literature in each area varies. This is especially the case since there are now no less than 40 types of reviews following systematic principles.
Here are some general guidelines on the structure for write-up and the review methods for a standard systematic review.
Structure of the review
Follow the PRISMA Statement and the MOOSE proposal on how to report systematic reviews and meta-analyses.
Relevant headings and areas to cover are listed below. They are not exclusive and some may not be needed or that others need to be added.
Abstract (250 words or less)
A summary of the review (i.e. what has been done and what was found) written in a structured format: have separate sections with sub-headings and structure that correspond to that in the main text. Please include five to eight keywords which will enable the paper to be classified in scholarly indices.
Background
The need for your review is explained and justified here, including evidence of a scoping search for existing reviews and meta-analyses. Please point out how this review is going to differ from previous reviews and how your review may add to the current knowledge or the previous review(s) (e.g. Your review is more up to date; conducted with more methodological rigour; question not previously addressed).
Aims & objectives
Clearly state the research questions in terms of; population, interventions / experience of interest, outcomes and study designs.
Inclusion criteria
Establish the criteria for the literature search and selection in terms of: Population
Intervention Comparator Outcome Study type(s)
Language (only restrict language in exceptional circumstances — discuss with supervisor)
A copy of the inclusion/exclusion form you used for each paper goes into an appendix.
Methods
The methods section contains details of the search process and strategies, assessments of study quality/risk of bias, data extraction and data synthesis. Sufficient details should be provided in this section to allow for an approximate replication of the review.
Sources of literature
Include details of data sources, the time span of the literature covered for each source and the date/week(s) when the searches were carried out. The sources are usually presented in the following order:
Electronic bibliographic databases (e.g. PsyclNFO, MEDLINE, EMBASE, ASSIA, ERIC)
most relevant to your research topic.
Grey literature sources (e.g. Google search, theses portals/databases or government/relevant major NGO websites)
reference lists from identified reviews / meta-analyses
contact with experts (including supervisors in the university and at placement who have experience in your research area)
hand-search of key journals
Search strategy
List all search terms in this section and put the actual syntax in the appendix. This is crucial and syntax must be included. Make sure terms defined in the criteria as well as their synonyms or historical terms (e.g. learning difficulty/ disability -> mental retardation) are all included.
Study selection
State how studies were screened and selected using the pre-defined inclusion / exclusion criteria.
Quality assessment (or Assessment of the risk of bias)
Describe the study quality assessment criteria, the application of those criteria to studies and the interrater reliability between the primary and second reviewers. The quality assessment criteria used in your review goes into an appendix.
Data extraction
Briefly describe the data extraction form (developed from your PICO and piloted before actual use) and how data extraction was performed. The data extraction form goes into an appendix.
Reso/Is
Description of studies
Explicitly report the study selection process, stating:
how many hits you had;
how many were duplicates or not relevant;
how many references were included;
how many references were excluded and reasons for exclusions;
how many references were left in the final analysis and how many studies they contain.
Please use the PRISMA flowchart to present this process. This is essential.
Report the findings of the included studies in a clear and concise manner.
present the characteristics and main findings in a table and summarise them narratively (descriptive data synthesis). A quick scan through these tables should allow readers to judge if the studies are similar in terms of population, intervention and outcomes.
Characteristics of included studies
Provide an overall picture and present important characteristics of all the studies included, such as; the overall number of participants included in the review (by adding the number of participants within each study together), the types of participants studied, and the intervention / exposure, outcomes / findings and study designs (if including more than one type of study design).
Ouality of included studies
Include another table summarising the results of your quality assessment according to types of bias.
Descriptive data synthesis
Summarise the results of the included studies and do a ‘vote counting’ exercise.
Comment on discernible patterns rather than describing results study by study.
Interpret statistical significance and confidence intervals.
Use quality assessment results in your interpretation of findings, e.g. any association between the level of study quality/risk of bias or any particular type of methodological strengths/flaws and a particular direction/kind of findings.
Consider quantitative data synthesis (meta-analysis) when the data are suitable.
Meta-analysis should be done if possible.
Discussion
Include:
the main findings of the review
interpretation of the findings
strengths and weaknesses of the evidence included in the review (remember clinical versus statistical significance).
applicability of your findings to your population of interest (young people).
if possible, relate your review to other reviews, especially noting the differences in quality and results.
Salient limitations of your review methods (show awareness of major problems but don’t contradict your own rationale for the review questions and methods).
implications of the findings on practice and research
recommendations on future research based on your findings and their limitations
Appendices (not included in word limit) include search syntax from your electronic search, inclusion / exclusion form, quality assessment form, data extraction form or any other materials. These must be added.
Useful resources:
Systematic Reviews: CRD’s Guidance for those Carrying Out or Commissioning Reviews.
January 2009 (Think before you print. There are nearly 500 pages!) Cochrane Handbook for Systematic Reviews of Interventions
A good calculator for working out standardised effect sizes is available at: https://www.campbeIIcoIIaboration.org/escaIc/html/EffectSizeCaIcuIator-SMD5.php
To convert different results into a standardised metric is available at: https://www.psychometrica.de/effect_size.html
Review Manager (RevMan) http://community.cochrane.org/tools/review-production- tools/revman-5
Meta Analysis in SPSS: Dr. Wilson MACRO — YouTube https://www.youtube.com/pIayIist?Iist=PL5ObXcjKAfyLSRZbNZIxZQgStQXYdmjqX
Dr Wilson’s meta-analysis guidance (D. B. Wilson) – George Mason University http://mason.gmu.edu/ dwiIsonb/ma.html based on a book by Mark Lipsey and Dr Wilson titled, Practical Meta-analysis, Sage (2001).
