Systematic reviews are widely used to identify, evaluate, and synthesise existing research in a transparent and reproducible manner. However, the quality of a systematic review depends not only on how studies are searched and analysed, but also on how clearly the entire review process is reported.
PRISMA 2020, which stands for Preferred Reporting Items for Systematic Reviews and Meta-Analyses, provides a structured framework for reporting systematic reviews. It helps authors explain how studies were identified, screened, assessed, included, and synthesised.
For PhD scholars, researchers, faculty members, and authors preparing systematic reviews for journal publication, understanding PRISMA 2020 is essential. The framework improves transparency, reproducibility, and confidence in review findings.
This step-by-step guide explains how researchers can apply PRISMA 2020 throughout the systematic review process.
What Is PRISMA 2020?
PRISMA 2020 is an updated reporting guideline for systematic reviews and meta-analyses.
It provides:
a 27-item checklist,
an expanded checklist,
a revised flow diagram,
and guidance for transparent reporting.
PRISMA does not tell researchers exactly how to conduct every methodological decision. Instead, it focuses on ensuring that authors clearly report what they did.
This distinction is important.
A review can follow PRISMA reporting requirements while still having methodological weaknesses if the search strategy, eligibility criteria, appraisal process, or synthesis methods are poorly designed.
PRISMA should therefore be used alongside sound systematic review methodology.
Step 1: Define a Clear Review Question
Every systematic review begins with a focused research question.
The question should define what the review intends to investigate.
Depending on the field, researchers may use frameworks such as:
PICO,
PICOS,
SPIDER,
PECO,
or other structured question formats.
For example, in transport research, a question might be:
“What factors influence public transport use among urban commuters in developing countries?”
A clear question helps determine:
databases,
search terms,
eligibility criteria,
study designs,
outcomes,
and synthesis methods.
Without a focused review question, the review can become unnecessarily broad.
Step 2: Develop a Review Protocol
Before beginning the search, researchers should prepare a protocol.
The protocol may include:
research objectives,
review question,
eligibility criteria,
databases,
search strategy,
screening procedure,
quality appraisal method,
data extraction framework,
and synthesis approach.
Where appropriate, researchers may register the protocol in a suitable registry or repository.
A protocol reduces the risk of changing methods after seeing the results.
It also improves transparency by showing that the review process was planned in advance.
Step 3: Define Inclusion and Exclusion Criteria
Eligibility criteria should be established before screening begins.
These may include:
publication years,
language,
geographical area,
study design,
participant characteristics,
intervention or exposure,
outcome measures,
document type,
and publication status.
For example, researchers may decide to include:
peer-reviewed journal articles,
published between 2015 and 2026,
written in English,
focused on urban public transport users,
and containing primary empirical data.
Exclusion criteria may remove:
editorials,
book reviews,
conference abstracts,
non-empirical papers,
duplicate studies,
or papers outside the study scope.
Eligibility criteria should be stated clearly in the final manuscript.
Step 4: Select Appropriate Databases
A systematic review should search databases relevant to the research field.
Common databases include:
Scopus,
Web of Science,
PubMed,
MEDLINE,
Embase,
IEEE Xplore,
ERIC,
PsycINFO,
and subject-specific databases.
Researchers should avoid choosing databases only because they are convenient.
The database strategy should provide adequate coverage of the literature.
In some cases, grey literature, institutional repositories, conference proceedings, or reference-list searching may also be useful.
Step 5: Develop a Reproducible Search Strategy
The search strategy is one of the most important components of a systematic review.
Researchers should identify:
main concepts,
synonyms,
related terms,
spelling variations,
controlled vocabulary,
and Boolean operators.
For example:
("public transport" OR "public transit" OR "mass transit") AND ("accessibility" OR "connectivity" OR "first mile" OR "last mile")
Search strings may need to be modified for different databases because search syntax differs.
PRISMA 2020 encourages transparent reporting of the search process.
Researchers should record:
database name,
search date,
complete search string,
filters applied,
and number of records retrieved.
Step 6: Export and Manage Search Results
After searching databases, records should be exported into reference-management or review-management software.
Common tools include:
Zotero,
EndNote,
Mendeley,
Rayyan,
Covidence,
and spreadsheet-based systems.
Researchers should retain bibliographic information such as:
title,
authors,
year,
abstract,
DOI,
journal,
and database source.
Maintaining a clean master database helps prevent errors during screening.
Step 7: Remove Duplicate Records
The same article may appear in multiple databases.
Duplicates should therefore be identified and removed before title and abstract screening.
Duplicate removal may be performed:
automatically,
manually,
or through a combination of both.
Researchers should document the number of duplicate records removed.
This number later appears in the PRISMA flow diagram.
Care is required because automatic duplicate detection can occasionally remove different papers with similar titles.
Step 8: Conduct Title and Abstract Screening
The remaining records are screened against the eligibility criteria.
At this stage, researchers review the title and abstract to determine whether each study is potentially relevant.
Records that clearly do not meet the criteria are excluded.
Where possible, two reviewers may screen records independently.
Disagreements can be resolved through discussion or by involving another reviewer.
For greater transparency, researchers should describe:
number of reviewers,
whether screening was independent,
and how disagreements were resolved.
Step 9: Retrieve Full-Text Articles
Studies that appear relevant after title and abstract screening move to full-text assessment.
Researchers should attempt to obtain the complete publication.
Sometimes full text cannot be retrieved.
PRISMA 2020 allows researchers to report records that were sought for retrieval but could not be obtained.
This stage is important because a paper that appears eligible from its abstract may not meet the criteria after full-text examination.
Step 10: Conduct Full-Text Eligibility Screening
During full-text review, researchers assess each article against the complete inclusion and exclusion criteria.
Studies may be excluded for reasons such as:
wrong population,
wrong study design,
inappropriate outcome,
wrong geographic context,
insufficient data,
or duplicate reporting.
The specific reasons for exclusion should be documented.
This information is required for transparent PRISMA reporting.
Avoid using vague statements such as “not relevant.”
Instead, provide meaningful exclusion categories.
Step 11: Complete the PRISMA Flow Diagram
The PRISMA flow diagram visually summarises how studies moved through the review.
It generally reports:
records identified,
duplicates removed,
records screened,
records excluded,
full texts sought,
full texts not retrieved,
full texts assessed,
full texts excluded with reasons,
and final studies included.
The flow diagram allows readers to understand exactly how the final evidence base was created.
The numbers must be internally consistent.
Researchers should verify that all totals add up correctly before submission.
Step 12: Conduct Quality or Risk-of-Bias Assessment
A systematic review should not treat all included studies as equally reliable.
Researchers should assess methodological quality or risk of bias using an appropriate tool.
The tool depends on the study design.
Examples may include instruments for:
randomised controlled trials,
cohort studies,
cross-sectional studies,
qualitative studies,
diagnostic studies,
and mixed-method research.
The chosen tool should be justified.
Researchers should report:
who performed the assessment,
how disagreements were handled,
and how quality findings influenced the synthesis.
Step 13: Extract Data Systematically
Once studies are included, relevant information should be extracted using a structured form.
Typical variables may include:
author,
publication year,
country,
study objective,
sample size,
methodology,
participant characteristics,
exposure or intervention,
outcome measures,
key findings,
and limitations.
A standardised extraction template helps maintain consistency.
If multiple reviewers extract data, procedures should be established for resolving inconsistencies.
Step 14: Synthesise the Evidence
The synthesis approach depends on the type of studies and available data.
Researchers may conduct:
narrative synthesis,
thematic synthesis,
meta-analysis,
subgroup analysis,
sensitivity analysis,
or mixed-method synthesis.
A meta-analysis is not required for every systematic review.
If studies are too heterogeneous in population, methodology, measures, or outcomes, a narrative or thematic synthesis may be more appropriate.
Researchers should clearly justify the chosen approach.
Step 15: Report Study Characteristics
Readers should be able to understand the evidence base.
A study-characteristics table commonly includes:
citation,
location,
sample,
study design,
variables,
methodology,
and principal findings.
This table helps readers compare included studies.
However, the text should interpret the evidence rather than merely repeat the table.
Step 16: Report Risk of Bias and Limitations
PRISMA 2020 expects transparent discussion of limitations.
These may include:
language restrictions,
limited database coverage,
publication bias,
missing full texts,
heterogeneity,
small study numbers,
or methodological weaknesses in included research.
Authors should also distinguish between:
limitations of the evidence,
and limitations of the review process.
This improves credibility.
Step 17: Report Search and Selection Transparently
One of the most common weaknesses in systematic reviews is inadequate reporting of the search process.
A reader should ideally be able to reproduce the search using the information provided.
Therefore, authors should report:
all databases,
search dates,
complete search strings,
filters,
screening methods,
and selection criteria.
Where journal word limits are restrictive, detailed search strategies may be placed in supplementary material.
Step 18: Use the PRISMA 2020 Checklist Before Submission
Before submitting the manuscript, researchers should complete the PRISMA 2020 checklist.
Each item should be matched to the page or section where it is addressed.
The checklist covers areas such as:
title,
abstract,
rationale,
objectives,
eligibility criteria,
information sources,
search strategy,
selection process,
data collection,
risk of bias,
synthesis,
results,
discussion,
registration,
funding,
and conflicts of interest.
Many journals request the completed checklist as a supplementary file.
Common Mistakes to Avoid
Researchers frequently make avoidable errors when applying PRISMA.
Common problems include:
calling a review “systematic” without a reproducible search,
using only one database without justification,
changing inclusion criteria during screening,
failing to report excluded full-text studies,
inconsistent flow-diagram numbers,
missing search dates,
incomplete search strings,
no quality appraisal,
and treating PRISMA as a substitute for methodology.
Another common mistake is stating that a review is “PRISMA compliant” simply because a flow diagram is included.
PRISMA involves much more than the diagram.
PRISMA 2020 and AI-Assisted Reviews
AI tools are increasingly being used to support:
search-term development,
screening,
data extraction,
and literature organisation.
However, researchers should still report the review process transparently.
If AI significantly influenced screening or extraction, authors should describe:
the tool used,
how it was used,
whether human verification occurred,
and how final decisions were made.
AI should improve efficiency without making the review process less reproducible.
Conclusion
PRISMA 2020 provides a comprehensive framework for transparently reporting systematic reviews and meta-analyses.
A strong PRISMA-based review begins with a clear research question, predefined protocol, transparent eligibility criteria, reproducible search strategy, systematic screening, documented exclusions, structured data extraction, quality appraisal, and appropriate evidence synthesis.
The PRISMA flow diagram is an important component, but it is only one part of the complete reporting framework.
For researchers and academic support organisations such as EduPub, using PRISMA 2020 correctly can improve the transparency, reproducibility, and publication quality of systematic reviews.
The central principle is simple: every important decision in the review process should be documented clearly enough for readers to understand how the final body of evidence was created.
