A systematic literature review is one of the most rigorous ways to identify, evaluate, and synthesize existing research on a specific question. Unlike a traditional literature review, a systematic review follows a predefined and transparent methodology for searching, selecting, evaluating, and synthesizing evidence.

For students, researchers, academics, and professionals, learning how to conduct a systematic literature review can make the research process more structured, reproducible, and defensible.

This guide explains how to write a systematic literature review step by step, including how to formulate a research question, develop a search strategy, select studies, assess risk of bias, synthesize findings, and use the PRISMA 2020 checklist to report the review.

Important: PRISMA is primarily a reporting guideline. It helps authors transparently report what they did and found; it is not, by itself, a methodology for conducting a systematic review or a tool for judging the overall quality of a review.

For additional support, explore our research support services to strengthen your manuscript. You may also find our guide on writing a literature review and guide on finding sources for literature review helpful for your review process.

What Is a Systematic Literature Review?

A systematic literature review is a structured synthesis of research that addresses a clearly defined question using explicit, reproducible methods.

A typical systematic review involves:

  • Defining a focused research question
  • Establishing eligibility criteria
  • Developing a comprehensive search strategy
  • Searching relevant databases and other information sources
  • Removing duplicate records
  • Screening titles and abstracts
  • Assessing full-text articles for eligibility
  • Extracting relevant data
  • Assessing the risk of bias in included studies
  • Synthesizing the evidence
  • Reporting the methods and results transparently

The purpose is not simply to summarize papers. Instead, a systematic review attempts to identify the available evidence comprehensively and synthesize it using a predefined approach.

Systematic Literature Review vs. Traditional Literature Review

The terms are sometimes used interchangeably, but they describe different approaches.

Feature Traditional Literature Review Systematic Literature Review
Research question May be broad Clearly defined
Search process Often flexible Planned and documented
Search strategy May not be fully reported Explicit and reproducible
Study selection Often narrative Predefined eligibility criteria
Quality/risk assessment Variable Usually formally assessed
Synthesis Primarily narrative Structured synthesis
Reporting Flexible Guided by reporting standards
Reproducibility Often limited Greater emphasis

A systematic review therefore requires considerably more planning and documentation.

How to Write a Systematic Literature Review Step by Step

Step 1: Choose a Clear Research Question

The first step is to define exactly what your review will investigate.

A weak question may be too broad:

"What is the impact of artificial intelligence on education?"

A more focused question might specify the population, intervention or exposure, outcome, and context.

Depending on the discipline, frameworks such as PICO, PECO, SPIDER, or other question-formulation frameworks may be appropriate.

For example, a PICO question can include:

  • P — Population: Who is being studied?
  • I — Intervention: What intervention or exposure is being investigated?
  • C — Comparison: What is the intervention being compared with?
  • O — Outcome: What outcomes are being measured?

The framework you choose should match your research question and study design.

Tip: Before beginning the search, write your research question in one sentence. Then identify the main concepts and synonyms that will eventually become your search terms.

Step 2: Define the Inclusion and Exclusion Criteria

Before screening studies, establish clear criteria determining which studies will be included and excluded.

Possible inclusion criteria:

You might specify:

  • Publication date
  • Language
  • Population characteristics
  • Study design
  • Intervention or exposure
  • Outcomes
  • Geographic location
  • Publication type
  • Required methodological characteristics

Possible exclusion criteria:

You might exclude:

  • Duplicate publications
  • Studies outside the target population
  • Studies that do not address the research question
  • Editorials or opinion pieces, when inappropriate for the review
  • Studies lacking required outcome data
  • Publications outside the predefined date range

The criteria should be determined before or according to the planned review protocol, rather than being changed simply because particular results are inconvenient.

PRISMA 2020 specifically asks authors to report the eligibility criteria and how studies were grouped for synthesis.

Step 3: Develop a Search Strategy

A systematic review depends heavily on the quality and transparency of its search strategy.

Start by breaking your research question into major concepts.

For example, suppose your research question concerns the use of artificial intelligence in higher education.

Potential concepts could include:

Concept 1: Artificial intelligence

  • "artificial intelligence"
  • AI
  • "machine learning"
  • "generative AI"

Concept 2: Higher education

  • university
  • universities
  • "higher education"
  • college
  • undergraduate

You can then combine synonyms using Boolean operators.

Example search string:

("artificial intelligence" OR "machine learning" OR "generative AI") AND ("higher education" OR universit* OR college*)

The exact syntax should be adapted to each database.

For example, database-specific subject headings, field tags, truncation symbols, and proximity operators may differ.

Document Your Search

Record:

  • Database name
  • Platform, where relevant
  • Date searched
  • Complete search string
  • Filters or limits
  • Number of records retrieved

PRISMA 2020 recommends reporting the full search strategies for databases, registers, and websites, including filters and limits.

This is important because another researcher should be able to understand how your evidence base was identified.

Step 4: Select Appropriate Information Sources

Do not automatically assume that one database is sufficient.

Depending on your research topic, potentially relevant sources may include:

  • Bibliographic databases
  • Subject-specific databases
  • Citation indexes
  • Study registers
  • Government websites
  • Professional organizations
  • Reference lists of relevant studies
  • Grey literature sources

The appropriate sources depend on the discipline and research question.

PRISMA 2020 asks authors to identify the databases, registers, websites, organizations, reference lists, and other sources consulted, as well as the date each source was last searched.

Step 5: Remove Duplicate Records

When several databases are searched, the same article may appear multiple times.

Use reference-management or systematic-review software to identify duplicates.

Commonly recorded information includes:

  • Total records identified
  • Duplicate records removed
  • Records remaining for screening

Do not simply delete duplicates without documenting the process.

Your final PRISMA flow diagram should allow readers to understand how many records moved through each stage of the review.

Step 6: Screen Titles and Abstracts

The next stage is screening.

Review the titles and abstracts against your predefined eligibility criteria.

For example:

  • Include: Studies involving university students that evaluate the specified intervention.
  • Exclude: Studies involving primary-school students or unrelated interventions.

In a rigorous review, screening procedures should be clearly described, including how many reviewers participated and whether they worked independently.

PRISMA 2020 specifically asks authors to describe the study-selection process and the number of reviewers involved at relevant screening stages.

Step 7: Perform Full-Text Screening

Potentially relevant studies should then be assessed in full.

At this stage, determine whether each article actually meets every inclusion criterion.

Keep a record of why potentially eligible full-text reports were excluded.

Examples of exclusion reasons might include:

  • Wrong population
  • Wrong intervention
  • Wrong study design
  • Wrong outcome
  • Ineligible publication type
  • Insufficient relevant data

Recording these reasons is essential for transparent reporting.

Step 8: Create a Data Extraction Form

Once studies have been selected, extract the information required to answer your research question.

A data-extraction table might include:

Field Example
Author/year Smith et al., 2025
Country India
Study design Randomized trial
Sample size 250
Population University students
Intervention/exposure AI-based learning tool
Comparator Conventional instruction
Outcome Academic performance
Follow-up 12 weeks
Main findings Reported outcome
Risk of bias Assessment result

The exact fields should be determined by your research question.

Avoid collecting large amounts of information that will never be used in the analysis.

Step 9: Assess Risk of Bias

A systematic review should consider the limitations of the included evidence.

Risk-of-bias assessment is different from simply asking whether a study is "good" or "bad."

Depending on the study designs included, researchers may use appropriate validated tools or methodological frameworks.

The assessment may consider issues such as:

  • Randomization
  • Allocation procedures
  • Missing outcome data
  • Measurement of outcomes
  • Selective reporting
  • Confounding
  • Deviations from intended interventions

The appropriate tool depends on the type of evidence being reviewed.

PRISMA 2020 asks authors to report the methods used to assess risk of bias and to present risk-of-bias assessments for the included studies.

Step 10: Decide How to Synthesize the Evidence

Not every systematic review requires a meta-analysis.

Narrative synthesis:

A narrative synthesis may be appropriate when studies are too heterogeneous to combine statistically.

You may organize findings according to:

  • Population
  • Intervention
  • Outcome
  • Study design
  • Research setting
  • Major themes

Quantitative synthesis or meta-analysis:

A meta-analysis statistically combines appropriate results from multiple studies.

However, studies should not be combined simply because they address a superficially similar topic.

Before conducting a meta-analysis, consider:

  • Clinical or substantive differences
  • Methodological differences
  • Outcome definitions
  • Measurement methods
  • Statistical compatibility
  • Risk of bias

The synthesis method should be planned and clearly justified.

Step 11: Analyze Heterogeneity

Heterogeneity refers to differences between studies.

These differences can occur in:

  • Participants
  • Interventions
  • Comparators
  • Outcomes
  • Study designs
  • Settings
  • Measurement methods

For quantitative reviews, statistical methods may also be used to evaluate heterogeneity.

However, statistical heterogeneity should not be interpreted independently of the substantive characteristics of the studies.

A statistically pooled result does not automatically mean that the underlying studies are sufficiently comparable.

Step 12: Write the Results

The Results section should clearly explain what the review found.

It should normally cover:

  1. Search and selection results
  2. Characteristics of included studies
  3. Risk-of-bias assessments
  4. Results of individual studies
  5. Results of the synthesis
  6. Additional analyses, where applicable

PRISMA 2020 recommends reporting the study-selection process from identified records through included studies, ideally with a flow diagram.

Tables can make the evidence easier to understand.

For example:

Study Design Sample Intervention Outcome Main finding
Study A RCT 200 Intervention X Outcome Y Result
Study B Cohort 450 Exposure X Outcome Y Result
Study C RCT 150 Intervention X Outcome Z Result

Step 13: Write the Discussion

The Discussion should interpret the findings without overstating what the evidence demonstrates.

A useful structure is:

  • Principal findings: What were the major findings?
  • Comparison with previous research: How do the findings relate to existing evidence?
  • Possible explanations: What factors might explain similarities or differences?
  • Limitations: What limitations exist in the included evidence and in the review itself?
  • Implications: What might the findings mean for research, practice, policy, or future studies?

Be especially careful not to turn association into causation or to make claims stronger than the evidence supports.

Step 14: Write the Conclusion

The conclusion should directly answer the review question.

Avoid introducing new evidence.

A strong conclusion usually includes:

  • The central finding
  • The strength or limitations of the evidence
  • Important uncertainty
  • Appropriate implications for future research or practice

What Is PRISMA 2020?

PRISMA stands for Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

PRISMA 2020 replaced the original PRISMA 2009 statement and provides updated guidance for reporting systematic reviews. It includes a 27-item checklist, an expanded checklist, an abstract checklist, and revised flow diagrams.

The purpose of PRISMA is to improve the transparency and completeness of systematic-review reporting.

PRISMA does not mean that a review is automatically methodologically rigorous simply because the checklist has been completed. Instead, it provides a framework for reporting what was done.

PRISMA 2020 Checklist: Complete Overview

The PRISMA 2020 checklist contains 27 main items and sub-items organized across the major sections of a systematic-review report.

Title and Abstract

  • 1. Title: Identify the report as a systematic review.
  • 2. Abstract: Use the PRISMA 2020 for Abstracts checklist to ensure that the abstract includes the necessary information.

Introduction

  • 3. Rationale: Explain why the review is needed in the context of existing knowledge.
  • 4. Objectives: State the review's objectives or research questions explicitly.

Methods

  • 5. Eligibility Criteria: Specify inclusion and exclusion criteria and explain how studies were grouped for synthesis.
  • 6. Information Sources: Identify every information source searched or consulted and provide the last search date.
  • 7. Search Strategy: Report the complete search strategies, including applicable filters and limits.
  • 8. Selection Process: Describe how studies were screened and selected, including reviewer involvement and relevant automation tools.
  • 9. Data Collection Process: Explain how data were collected from reports and how many reviewers were involved.
  • 10. Data Items: Define the outcomes and other variables that were sought.
  • 11. Study Risk of Bias Assessment: Identify the methods used to assess risk of bias.
  • 12. Effect Measures: Specify the effect measures used for each outcome, where applicable.
  • 13. Synthesis Methods: Explain how studies were prepared and synthesized.
  • 14. Reporting Bias Assessment: Describe how potential reporting biases were assessed.
  • 15. Certainty Assessment: Describe how certainty or confidence in the evidence was assessed, where applicable.

Results

  • 16. Study Selection: Report the search and selection process, ideally with a PRISMA flow diagram.
  • 17. Study Characteristics: Present the characteristics of each included study.
  • 18. Risk of Bias: Present risk-of-bias assessments for the included studies.
  • 19. Results of Individual Studies: Report the results of individual studies for relevant outcomes.
  • 20. Results of Syntheses: Present the results of each synthesis.
  • 21. Reporting Biases: Present assessments of possible reporting biases.
  • 22. Certainty of Evidence: Present assessments of certainty or confidence in the evidence, where applicable.

Discussion

  • 23. Discussion: Interpret the results in the context of existing evidence. Discuss limitations of the evidence, limitations of the review process, and implications for practice, policy, and future research.

Other Information

  • 24. Registration and Protocol: Report registration details and where the review protocol can be accessed, when applicable. Also report important amendments to the protocol or registration.
  • 25. Support: Report sources of financial and other support for the review.
  • 26. Competing Interests: Declare relevant competing interests.
  • 27. Availability of Data, Code, and Other Materials: State which review materials are publicly available and where they can be accessed, when applicable.

This structure follows the PRISMA 2020 checklist and should be checked against the official checklist when preparing a manuscript.

How to Create a PRISMA Flow Diagram

The PRISMA flow diagram shows how records move through the review process.

A typical flow includes:

  • Records identified
  • Duplicate records removed
  • Records screened
  • Records excluded
  • Reports sought for retrieval
  • Reports not retrieved
  • Reports assessed for eligibility
  • Reports excluded
  • Studies included in review
  • Studies included in synthesis

The exact diagram depends on the type of review and the sources searched.

The official PRISMA website provides separate templates for new and updated reviews and for reviews using databases/registers alone versus databases/registers plus other sources.

Example:

Suppose your search produces 2,500 records:

  • 2,500 records identified
  • 400 duplicates removed
  • 2,100 records screened
  • 1,850 records excluded
  • 250 full-text reports assessed
  • 180 reports excluded after full-text assessment
  • 70 studies included

The numbers should be internally consistent throughout the flow diagram and Results section.

Do not invent numbers for a PRISMA diagram. Every number should come from your documented review process.

PRISMA Checklist vs. PRISMA Flow Diagram

These are related but different.

  • PRISMA checklist: The checklist helps you determine whether the manuscript reports the required information.
  • PRISMA flow diagram: The flow diagram illustrates the movement of records and studies through identification, screening, eligibility assessment, and inclusion.

You generally use both when applicable.

Official PRISMA resources provide downloadable checklist and flow-diagram templates.

Before vs. After: Weak Systematic Review vs. Strong Systematic Review

Aspect Weak Review Strong Review
Research question Broad or unclear Focused and answerable
Search strategy Incomplete or unreported Comprehensive and reproducible
Eligibility criteria Undefined or changed mid-review Predefined and consistently applied
Screening Not documented Transparent with reviewer details
Risk of bias Ignored Formally assessed with appropriate tools
Synthesis Simple summary of studies Structured synthesis appropriate to evidence
Reporting Incomplete Follows PRISMA 2020 guidelines
Flow diagram Missing or inconsistent Complete and internally consistent

Common Mistakes When Writing a Systematic Review

Mistake Why It Hurts Your Review
Starting the search before defining the question A broad or unclear question can produce an unmanageable evidence base.
Using only one database without justification A single database may not capture all relevant literature.
Reporting an incomplete search strategy "PubMed was searched" is not enough for reproducibility.
Changing inclusion criteria during screening Changing criteria because results are inconvenient can introduce bias.
Treating PRISMA as the review methodology PRISMA is a reporting guideline, not a complete research methodology.
Assuming meta-analysis is mandatory A systematic review can use narrative or other appropriate synthesis methods.
Ignoring risk of bias Simply counting studies does not tell readers how trustworthy their findings may be.
Overstating conclusions A review should not make stronger claims than its evidence supports.

Tips by Academic Discipline for Systematic Reviews

  • Health and Medical Sciences (Medicine, Nursing, Public Health): Use PRISMA as the standard reporting guideline. Register your protocol with PROSPERO. Use validated risk-of-bias tools appropriate to study designs.
    Example emphasis: "For clinical reviews, use Cochrane risk-of-bias tools and consider GRADE for certainty assessment."
  • Social Sciences (Psychology, Sociology, Education): Use appropriate frameworks like SPIDER for qualitative or mixed-methods questions. Consider both quantitative and qualitative synthesis approaches.
    Example emphasis: "For educational research, consider using thematic synthesis alongside meta-analysis where appropriate."
  • Environmental and Natural Sciences (Biology, Ecology, Environmental Science): Consider using systematic review methodology for environmental evidence. Use appropriate databases and grey literature sources.
    Example emphasis: "For ecological reviews, consider using the CEE (Collaboration for Environmental Evidence) guidelines alongside PRISMA."
  • Business and Management Research: Adapt systematic review methods for management research. Use appropriate databases and consider both academic and practitioner literature.
    Example emphasis: "For management reviews, consider using frameworks like SPIDER and clearly justify your inclusion of non-traditional sources."

A Practical Systematic Review Workflow

You can use the following workflow when planning your review:

Phase 1: Planning

  • Define the research question
  • Identify the review framework
  • Establish inclusion and exclusion criteria
  • Develop the protocol
  • Consider registration where appropriate

Phase 2: Searching

  • Select databases and other sources
  • Develop search strings
  • Run the searches
  • Export records
  • Document search dates and results

Phase 3: Screening

  • Remove duplicates
  • Screen titles and abstracts
  • Retrieve full texts
  • Apply eligibility criteria
  • Record exclusion reasons

Phase 4: Data Collection

  • Develop a data-extraction form
  • Extract study characteristics
  • Extract relevant outcome data
  • Check extracted information

Phase 5: Quality and Bias Assessment

  • Select an appropriate risk-of-bias tool
  • Assess included studies
  • Record judgments and supporting information

Phase 6: Synthesis

  • Determine an appropriate synthesis method
  • Group comparable studies
  • Conduct quantitative synthesis when appropriate
  • Explore relevant heterogeneity
  • Assess reporting biases where appropriate

Phase 7: Reporting

  • Write the Introduction
  • Describe Methods
  • Present Results
  • Write the Discussion
  • Prepare the Conclusion
  • Complete the PRISMA checklist
  • Prepare the PRISMA flow diagram

Systematic Literature Review Checklist

Before submitting your manuscript, ask:

  • Is the research question clearly defined?
  • Are inclusion and exclusion criteria explicit?
  • Are all relevant information sources identified?
  • Are search dates documented?
  • Are complete search strategies available?
  • Have duplicate records been handled?
  • Is the screening process documented?
  • Are full-text exclusion reasons recorded?
  • Is data extraction clearly described?
  • Has risk of bias been assessed appropriately?
  • Is the synthesis method justified?
  • Are individual study findings reported?
  • Are limitations discussed?
  • Are conclusions supported by the evidence?
  • Is the review protocol or registration information reported where applicable?
  • Are funding and competing interests disclosed?
  • Is data/code/material availability addressed where applicable?
  • Has the PRISMA 2020 checklist been completed?
  • Has the appropriate PRISMA flow diagram been prepared?

The official PRISMA checklist should be used for the final manuscript because it contains the detailed reporting requirements and sub-items.

Writing a systematic literature review requires much more than collecting and summarizing published papers. It requires a transparent process for defining the question, identifying evidence, selecting studies, extracting data, assessing risk of bias, synthesizing findings, and reporting the results.

The PRISMA 2020 checklist provides a structured reporting framework covering the title, abstract, introduction, methods, results, discussion, and other important information. The accompanying PRISMA flow diagram helps readers understand how records moved from the initial search to the final set of included studies.

For researchers preparing a systematic review, the best practice is to consult the official PRISMA 2020 materials throughout the writing process rather than treating the checklist as a final formatting exercise. This helps make the review clearer, more reproducible, and easier for readers to critically evaluate. For personalized guidance, you can book a consultation with our experts or explore our pricing plans.

Keep Exploring: Recommended Reads from AD Publications Blog

You may also find these articles valuable for your research and academic writing journey: