How to Conduct a PRISMA Systematic Review:
A Complete Step-by-Step Guide
A PRISMA systematic review is the gold standard for pulling together evidence in clinical and biomedical research. Whether you are a researcher seeking grant funding, a clinician building evidence-based guidelines, or a PhD student finishing your thesis, this skill matters. Knowing how to run a solid systematic review — and report it under the PRISMA 2020 guidelines — is one of the most useful skills you can build.
In this guide, I walk you through every stage of the process — from shaping your research question to sending in your final manuscript. I also cover the tools you need and the mistakes that most often lead to desk rejections. You will also see how professional systematic review services can help you get your review published in a top journal on the first try.
What Is a PRISMA Systematic Review?
PRISMA stands for Preferred Reporting Items for Systematic Reviews and Meta-Analyses. The PRISMA 2020 statement came out at the same time in PLOS Medicine, The BMJ, and the Journal of Clinical Epidemiology. It gives you a 27-item checklist and a standard flow diagram that every systematic review must follow to be ready for major medical journals.
A systematic review differs from a narrative review in one key way: it uses a set, repeatable method to find, pick, check, and pull together all relevant evidence on a clear clinical or research question. This cuts the risk of bias and cherry-picking that weakens narrative reviews. That is why journal editors and peer reviewers hold systematic reviews to a much higher standard.
PRISMA 2020 is not just a reporting guideline — it is a signal of quality method. A PRISMA-compliant review tells editors that your study followed a set plan, searched widely, and reported openly. Not meeting PRISMA is one of the most common reasons for desk rejection.
Why PRISMA 2020 Matters for Publication
Since the original PRISMA statement in 2009, and the big update in 2020, following these guidelines has become a must. Journals including The Lancet, JAMA, BMJ, Cochrane Library, and hundreds of specialty journals now treat it as standard. Reviewers are trained to check your flow diagram, your risk of bias tables, and your GRADE evidence summary before they even look at the science itself.
Journals now routinely use PRISMA checklists during peer review — and many require authors to send in a completed checklist with the manuscript. Not meeting PRISMA 2020 standards is not a minor formatting issue. It is treated as a serious flaw that can lead to rejection at any stage, from desk review to final revision.
PRISMA 2020 — What’s New vs PRISMA 2009
The 2020 update introduced several important changes that affect how reviews must be conducted and reported:
- New items on protocol and registration — PROSPERO disclosure now mandatory
- Expanded guidance on searching, including trial registers and grey literature
- Risk of bias must now be assessed at outcome level, not just study level
- Separate reporting requirements for pairwise and network meta-analyses
- Updated flow diagram distinguishing new searches from updated searches
- Certainty of evidence (GRADE) now explicitly required in the abstract
Step-by-Step: How to Conduct a PRISMA Systematic Review
The following eight steps reflect the complete lifecycle of a PRISMA 2020-compliant systematic review — from clinical question to accepted manuscript. These are the exact steps followed in our professional systematic review services.
Formulate Your Research Question Using PICO
Define your Population, Intervention, Comparator, and Outcome (PICO) with precision. A well-defined PICO sets your eligibility criteria, guides your search strategy, and makes sure your review answers a clinically useful question. An unclear PICO is one of the most common early-stage errors, and it causes problems at every later stage.
Register Your Protocol on PROSPERO
Before you run any searches, register your systematic review plan on PROSPERO (International Prospective Register of Systematic Reviews). Registration records your methods before data collection, stops duplicate reviews, and is now required by most Q1 journals. Your PROSPERO registration number must appear in your manuscript’s methods section.
Develop a Comprehensive Search Strategy
A PRISMA-compliant search must cover at minimum: PubMed/MEDLINE, EMBASE, Cochrane CENTRAL, and one more relevant database (Scopus, Web of Science, CINAHL). Search terms should use MeSH terms, plain-text synonyms, and Boolean operators. Grey literature sources — ClinicalTrials.gov, WHO ICTRP, and thesis repositories — must also be searched and recorded.
Screen Titles, Abstracts & Full Texts (Dual Independent)
Two independent reviewers must screen everything, using tools like Rayyan or Covidence. Screen titles and abstracts first, then pull and screen the full texts of studies that might qualify. All disagreements must be settled by discussion or a third reviewer, with every decision recorded in your PRISMA flow diagram.
Extract Data Using Standardised Forms
Build tested data extraction forms that capture study design, population details, intervention details, comparator, outcomes, and results. Two independent reviewers should extract the data, and any differences should be worked out together. Extracted data is the base for your results tables and meta-analysis input.
Assess Risk of Bias
Use the right validated tool: Cochrane RoB 2.0 for RCTs, ROBINS-I for non-randomised studies, QUADAS-2 for diagnostic accuracy studies, or Newcastle-Ottawa Scale for cohort and case-control studies. Risk of bias must be checked at outcome level for each included study and shown in a clear table — one of the areas peer reviewers check most closely.
Synthesise Evidence & Apply GRADE
Pull your findings together in words, or with numbers through meta-analysis. If you run a meta-analysis, pick the right statistical model, check heterogeneity (I², Cochran’s Q), and run pre-planned subgroup and sensitivity checks. Apply the GRADE framework to rate certainty of evidence for each outcome — now required in the abstract and summary of findings table by nearly all major journals.
Write & Submit the PRISMA 2020-Compliant Manuscript
Structure your manuscript to the target journal’s author guidelines, and complete and attach the PRISMA 2020 checklist with your submission. Your abstract must include GRADE certainty ratings. Your PRISMA flow diagram must match every stage exactly. Include the PROSPERO number, conflict of interest statements, and data availability statement as required.
Essential Tools for Systematic Reviews
A high-quality systematic review needs the right tools at each stage. Here is the toolkit used in our professional systematic review services:
- PROSPERO — Protocol registration (free, mandatory for most Q1 journals)
- Rayyan / Covidence — Dual-reviewer title/abstract and full-text screening
- PubMed, EMBASE, Cochrane CENTRAL — Core literature databases
- ClinicalTrials.gov, WHO ICTRP — Trial register and grey literature searches
- Cochrane RoB 2.0 / ROBINS-I — Risk of bias assessment tools
- RevMan 5 / R (meta, metafor) — Meta-analysis and forest plot generation
- GRADEpro GDT — Evidence certainty profiling and summary of findings tables
- Zotero / Mendeley / EndNote — Reference management and deduplication
Common Mistakes That Get Systematic Reviews Rejected
After supporting over 20 Q1 publications through professional systematic review services, these are the errors I see most often in manuscripts rejected before they reach us:
- No PROSPERO registration, or registration after data collection began
- Search limited to PubMed only — missing EMBASE, Cochrane, and grey literature
- Single-reviewer screening without inter-rater reliability documentation
- Risk of bias assessed at study level rather than outcome level (PRISMA 2020 requirement)
- GRADE certainty ratings absent from abstract and summary of findings tables
- Heterogeneity reported but not explored through subgroup or sensitivity analyses
- Eligibility criteria not pre-specified or inconsistently applied during screening
- PRISMA 2020 checklist not completed or submitted with the manuscript
- Discussion overstates certainty of evidence relative to GRADE ratings
- No data availability statement or protocol accessibility disclosure
The most common reason high-quality systematic reviews fail peer review is not flawed science — it is poor reporting. A review with excellent method, reported poorly, will be rejected. Reporting quality is within your control, and it is where professional systematic review services add the most value.
Frequently Asked Questions About Systematic Reviews
Q. What is a PRISMA systematic review?
A PRISMA systematic review is a full round-up of published research on a specific clinical question, done and reported under the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines. PRISMA gives you a 27-item checklist and standard flow diagram to keep reporting open, repeatable, and complete — now required by nearly all major medical journals.
Q. How long does a systematic review take to complete?
A typical systematic review takes 8 to 16 weeks, depending on scope, number of databases, number of studies, and how complex the synthesis is. Simple reviews with a narrow PICO and a small literature base can be done in 8 weeks. Bigger reviews — especially those with a network meta-analysis — usually need 12 to 16 weeks. Rush timelines can be arranged for grant deadlines or conference submissions.
Q. Do I need to register my systematic review on PROSPERO?
Yes — PROSPERO registration is strongly recommended and now required by most Q1 journals before peer review. Registering your protocol before data collection begins stops outcome reporting bias and duplicate reviews, and shows your methods are open. Registration is free at crd.york.ac.uk/prospero. Our systematic review services include full PROSPERO protocol preparation and submission.
Q. What is the difference between a systematic review and a meta-analysis?
A systematic review uses a set, planned method to find, pick, and closely check all relevant studies on a clear research question. A meta-analysis is a statistical method used inside a systematic review to combine numeric results from multiple studies into one summary estimate with confidence intervals. Not all systematic reviews include a meta-analysis — pooling only makes sense when the included studies are similar enough.
Q. What tools are used to conduct a systematic review?
Core tools include Rayyan or Covidence for screening, and RevMan 5 or R (meta, metafor packages) for meta-analysis and forest plots. Risk of bias is checked with Cochrane RoB 2.0 or ROBINS-I, and evidence certainty is rated with GRADEpro. Database searches are run in PubMed/MEDLINE, EMBASE, Cochrane CENTRAL, Scopus, and Web of Science, plus trial registers and grey literature.
Q. Can I get professional help with my systematic review?
Yes. I offer end-to-end systematic review services — from shaping your PICO and PROSPERO registration to multi-database searching, dual screening, and data extraction. This also covers risk of bias checks, GRADE profiling, statistical meta-analysis, PRISMA 2020-compliant manuscript writing, and peer review response. Available to researchers, clinicians, PhD students, and academic institutions worldwide. Contact me for a free consultation.