It is important to distinguish writing a systematic review from conducting one. These are not interchangeable terms. While conducting a systematic review refers to the process of planning, identifying, screening, analyzing, and evaluating relevant studies, writing a systematic review is the final, yet crucial, step of that process.
If you're looking to learn how to conduct a systematic review — from planning and protocol development to search strategies and bias assessments — check out our Comprehensive Systematic Review Guide, packed with infographics, video tutorials, and practical resources to get you started.
Now, let’s walk through how to write a systematic review effectively.
Begin with the PRISMA 2020 Statement
When writing a systematic review, follow the PRISMA 2020 Statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), which provides a structured reporting framework. PRISMA recommends structuring your systematic review into the following sections:
Background
Methods
Results
Discussion
Let’s explore each section in detail.
1. Background
This is where you provide the rationale for conducting your systematic review. Your justification should be rooted in an identified research gap—this could come from a traditional literature review or a scoping review that highlights underexplored areas.
For example:
“While several individual studies have investigated the effectiveness of mindfulness-based interventions in reducing adolescent anxiety, no comprehensive synthesis has yet evaluated the quality and consistency of this evidence.”
After highlighting the gap, narrow down to your review aim or research question. Use structured formats like PICOS (Population, Intervention, Comparison, Outcome, Study Design) to formulate clear research questions.
2. Methods
This section should be comprehensive and transparent. It must detail every decision and process undertaken to ensure scientific integrity and reproducibility.
Start by stating:
“This systematic review was conducted in accordance with the PRISMA 2020 Statement…”
Key Components of the Methods Section:
a) Eligibility Criteria
Define inclusion and exclusion criteria, such as:
Population (e.g., adults over 60)
Interventions (e.g., digital CBT)
Study designs (e.g., RCTs only)
Languages included/excluded
Time frame of publication
b) Search Strategy
List and justify the databases searched (e.g., PubMed, Scopus, PsycINFO). Include search terms and Boolean operators. Attach the full search string as an appendix if possible.
“We searched PubMed, Embase, and Cochrane Library using terms like 'digital therapy', 'online CBT', 'depression', and their synonyms, covering studies published between 2000 and 2023.”
c) Study Selection
Explain how titles, abstracts, and full texts were screened. Indicate if multiple reviewers were involved and how disagreements were resolved (e.g., through consensus or third-party adjudication).
d) Data Extraction
Specify what data was extracted and how (e.g., using Excel or systematic review software). Examples include:
Study design
Sample size
Population characteristics
Intervention details
Outcomes measured
Follow-up period
e) Risk of Bias and Quality Assessment
State the tool used, such as:
Cochrane Risk of Bias Tool 2.0 for RCTs
ROBINS-I for non-randomized studies
Joanna Briggs Institute tools for various study types
Cite the tools and justify their selection.
f) Data Synthesis
Describe the synthesis approach:
Narrative synthesis (common for qualitative data) using thematic analysis
Meta-analysis (for quantitative data), including effect size measures, statistical models, and heterogeneity assessments
g) Certainty of Evidence
If your review seeks relationships or intervention effects, assess the certainty of evidence using GRADE, developed by the GRADE Working Group. GRADE assesses domains like consistency, directness, and precision.
“Certainty of evidence was assessed using GRADE to determine confidence in effect estimates.”
All decisions and steps in this section must be well-justified and clearly reported.
3. Results
This section must describe your findings in a logical and data-rich manner.
a) Study Selection Process
Report the number of:
Records identified from each source
Duplicates removed
Titles and abstracts screened
Full-texts retrieved and screened
Records excluded at each stage (with reasons)
Tip: Tools like Rayyan and Covidence can track and document each stage. For example:
Records removed due to “wrong population” or “inappropriate outcomes”
Duplicates across databases
Full-text retrieval challenges
These tools enhance the transparency and reproducibility of your selection process.
Present this data using a PRISMA flow diagram.
b) Characteristics of Included Studies
Create an evidence summary table. Common characteristics include:
Author(s), Year
Country of study
Study design
Population and sample size
Intervention characteristics (type, duration, delivery method)
Comparison group (if any)
Outcomes measured
Key findings
c) Risk of Bias/Quality Assessment
Present findings from your quality assessment using charts or summary tables, depending on the tool used.
d) Synthesis of Results
Present the synthesized findings. If meta-analysis was performed, include forest plots and heterogeneity metrics (e.g., I² values). For narrative synthesis, organize findings under themes or categories.
4. Discussion
a) Summary of Findings
Start by revisiting your review aim or research question, then summarize the main findings.
“This review found consistent evidence that school-based mindfulness programs reduce anxiety symptoms among adolescents…”
b) Interpretation and Comparison with Previous Studies
Critically compare your findings with previous reviews or related studies.
For example:
“Unlike previous reviews that found limited evidence for digital CBT, our review reveals stronger support—likely due to the inclusion of more recent, large-scale trials.”
Discuss why results may differ:
Methodological differences
Target populations
Measurement tools
Study quality
c) Limitations
Mention limitations in both your process and the included studies:
Limited to English-language studies
Exclusion of unpublished data (publication bias)
Inconsistent outcome measures across studies
d) Implications
Discuss:
Policy: “Findings support integrating digital CBT into primary care settings.”
Practice: “Training for school counselors on mindfulness-based approaches is recommended.”
Future research: “More longitudinal studies with diverse populations are needed.”
Final Thoughts
Writing a systematic review isn’t just about summarizing studies — it’s about presenting a rigorously conducted process in a structured, transparent, and reproducible manner. Follow PRISMA guidelines, document each step clearly, and use available tools to support accuracy.
If you haven't already conducted your review and are unsure where to begin, remember to visit our Comprehensive Systematic Review Guide for step-by-step instructions, templates, and visuals.