6 Conducting a systematic review
This section describes the basic steps of conducting a systematic review. There is more in-depth guidance in the literature. (Muka et al. (2020), Cherry, Dickson, and Boland (2024), Egger, Higgins, and Smith (2022), Gough, Oliver, and Thomas (2017))
6.1 Research question
Develop an answerable research question. Use the FINER criteria and frameworks such as PICO to guide you through the process.
6.2 Scoping search
Run a quick scoping search (e.g. in PubMed to achieve three goals:
Find out whether a systematic review was recently published on your topic.
Get a feeling how much primary literature is available on the topic.
Find relevant seed papers to inform your systematic search.
It is highly recommended to search relevant protocol registers (PROSPERO, OSF, INPLASY) for similar projects to avoid research waste.
6.3 Study design decision
Systematic reviews are conducted by a team in a collaborative manner and take a substantial amount of time to complete. Systematic reviews often take approximately 12-18 months, from protocol to publication.
Determine if a systematic review is an appropriate methodology for your research question.(Sutton et al. (2019), Grant and Booth (2009)) Consider the Review section for guidance and alternatives. The website Right Review provides a tool to determine the most suitable study type for your research.
6.4 Building a team
A proper systematic review cannot be conducted by one person. Gather a team of experts for your systematic review project, including:
- Subject experts with clinical and methodological expertise
- At least two screeners to review the results independently
- A third person or tiebreaker to make decisions if there is disagreement about a study meeting the inclusion criteria
- A statistician, if performing a meta-analysis
- An information specialist or a medical librarian trained to conduct a systematic literature search.
Information specialists …
- … develop the search strategies which are essential for data acquisition.
- … document the systematic search in a transparent and reproducible manner.
- … draft the Methods section detailing the relevant steps of the search.
- … review and edit the manuscript and provide approval of the Methods section.
- … share all obligations as co-authors with the rest of the team.
For more information on the role of information specialists see Spencer and Eldredge (2018), Metzendorf (2016), Foster (2015), Koffel (2015), Rethlefsen et al. (2015), Schellinger et al. (2021), Meert, Torabi, and Costella (2016).
The International Committee of Medical Journal Editors (ICMJE). provides recommendations regarding the roles of authors and contributors, which should be considered in decisions about co-authorship.
6.5 Protocol registration
A protocol is a detailed plan for your review. It includes a rationale for the research project, the research question, the inclusion/exclusion criteria, an outline of the planned literature search, methods for data abstraction, data management and the quality assessment of the studies.
The idea of a protocol is to counteract publication bias and to inform the research community of your project and thus to avoid research waste. (The PLoS Medicine Editors (2011))
Write a protocol and register it in a register suitable to your review type, such as PROSPERO, OSF or INPLASY. Make sure to follow a reporting guideline, e.g. PRISMA-P.
6.6 Literature search
Perform a systematic search of the literature, preferably supported or conducted by information specialist.
6.7 Screening of results
Screen and classify studies based on the eligibility criteria defined in your protocol. This is usually done by screening titles and abstracts first, then full texts.
It is highly recommended to use a screening tool. Do this in a team of at least two screeners to counteract selection bias. Make sure to keep track of the numbers of included and excluded records at every stage of the screening process using a PRISMA flow diagram.
6.8 Data extraction
Compile all the relevant information for data analysis from the eligible studies.
6.9 Quality assessment
Perform a risk of bias analysis and critical appraisal of the studies. Use a current risk-of-bias tool and checklists for critical appraisal.
See also Section 4.2
6.10 Synthesis of research
Analyze the results from the compiled data. Consider performing a meta-analysis if the data is suitable and a statistical examination of the data makes sense.
6.11 Interpretation
Draw conclusions from the your results and summarize your findings according to your target audience.
6.12 Write and publish
Select a suitable journal for publication. Make sure to avoid predatory publishers. Draft a manuscript together with your co-authors and publish your systematic review. Make sure to properly document your review and your systematic search according to PRISMA and PRISMA-S. In case you need a reference management program, make sure you consult Section 3.9 for guidance.