Systematic Reviews

Question Formulation

A well-defined research question will help to identify the topics to search for in order to answer your research question. See the "Formulating Your Question" page for more specific frameworks and additional info. When formulating your question, clearly state the objectives of your review, and consider using the FINER components:

• Feasibility A feasible review poses a question that can be answered using the available evidence and avoids generating unmanageable amounts of information. Authors should conduct initial scoping to set clear boundaries and ensure the question isn't unanswerable due to lack of evidence.
• Interest Starting a review on a topic that is interesting to authors and others is important, as reviews require substantial effort and authors must be committed to completing the process.
• Novel A novel review should fill a real knowledge gap, so authors must check existing and ongoing reviews to avoid duplication and understand the broader research context.
• Ethical Considerations in conducting a review include opportunity costs, the prioritization and framing of research questions, and the potential political impacts, such as unintentionally widening health inequalities.
• Relevant Due to the significant effort required for a systematic review, authors should ensure its relevance by involving stakeholders in shaping its focus and by presenting findings in a way that supports decision-making. The GRADE framework supports this goal and should be applied throughout the review process.

Search Strategy

A systematic search for studies should be thorough and aim to balance recall (finding as many relevant studies as possible) with precision (limiting irrelevant results), recognizing that not all retrieved articles will be directly relevant.

Most studies will typically be found through the library's databases.

Depending on the topic’s breadth, search results may range from a few hundred to several thousand articles.

Key steps in the search process include:

• Selecting appropriate databases
• Defining relevant search terms
• Constructing and executing database searches
• Hand-searching specialized journals
• Reviewing reference lists of pertinent studies
• Contacting experts in the field
• Exploring grey literature sources like conference papers, government documents, and dissertations

Screening

At this stage of the review, a large number of studies must be screened, ideally by at least two team members to minimize bias. Creating a standardized form can help track and select studies that meet the eligibility criteria. Studies should only be included if they satisfy all inclusion criteria and none of the exclusion criteria.

Follow this process:

• Review titles and abstracts
• Retrieve full-text articles for potentially relevant studies
• Assess full texts for eligibility
• Make final inclusion decisions

Be sure to document all decisions to ensure transparency, reproducibility, and to reduce the risk of errors.

Data Extraction

Data refers to key details from a study, such as its design, participants, setting, interventions, and results. To minimize errors and reduce bias, data should ideally be extracted independently by two or more researchers; at a minimum, one person should extract the data and another should verify it. A clear method for resolving disagreements, such as involving a third reviewer, should be in place.

To ensure consistency and validity, it’s important to develop a well-thought-out data extraction process. This can be implemented using word processors, spreadsheets, citation managers, or other tools.

Care should be taken to gather enough relevant information to thoroughly assess each study. Collecting too little risks missing key data, while collecting too much can waste time.

Synthesis

Synthesis is the process of bringing together the findings from the studies included in a review to summarize results and draw trustworthy conclusions based on the evidence's quality. This can be done quantitatively using methods like meta-analysis, or qualitatively through a narrative approach.

When writing your report or article, be sure to consult the PRISMA Checklist, which is required by many journals and commissioning bodies for systematic review reporting.

Thomas J, Kneale D, McKenzie JE, Brennan SE, Bhaumik S. Chapter 2: Determining the scope of the review and the questions it will address [last updated August 2023]. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.5. Cochrane, 2024. Available from Cochrane Handbook

Adapted from University of South Florida Systematic Reviews for Social Sciences Guide https://guides.lib.usf.edu/systematicreviews/introduction

PRISMA stands for Preferred Reporting Items for Systematic Reviews and Meta-Analyses. It is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses.