Conducting a Systematic Literature Review: Protocol, PRISMA, and Methods (2026)

A systematic literature review is not just a thorough literature review — it’s a specific research methodology with defined protocols for searching, screening, and synthesising evidence. Many PhD students in Indian universities learn this distinction too late, usually when their doctoral committee or examiners ask why the review cannot be replicated. Use it when you need to answer a precise research question by identifying and appraising all relevant studies, minimising bias throughout.

Systematic vs. Narrative Literature Reviews

The key difference is reproducibility:

Use a systematic review when your research question requires complete evidence synthesis and your field expects the methodology — medicine, public health, education, and social policy most commonly.

Step 1: Formulate a Focused Research Question Using PICO

Systematic reviews begin with a structured research question — and the framing matters more than most people realise. In health and social sciences, PICO is the standard framework:

• Population: Who is the study about? (e.g., postgraduate students)
• Intervention or exposure: What is being done or examined? (e.g., plagiarism detection software)
• Comparison: What is the alternative? (e.g., manual checking)
• Outcome: What are you measuring? (e.g., detection accuracy)

A PICO question: “In postgraduate students (P), does plagiarism detection software (I) compared to manual checking (C) improve detection of paraphrased plagiarism (O)?”

In other disciplines, variations like SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type) are used for qualitative research.

Step 2: Register Your Protocol

Before searching, register your review protocol on PROSPERO (for health-related reviews) or the Open Science Framework. Registration creates a public, time-stamped record of your planned methodology — preventing post-hoc changes to inclusion criteria after seeing what comes up in the search.

Even if registration isn’t required in your field, writing a protocol document before you begin forces clarity about your methodology and protects against accusations of bias.

Step 3: Develop and Run the Search Strategy

A systematic search must meet three criteria:

• Multi-database: Run in at least three databases. For most disciplines: Scopus + Web of Science + a field-specific database (PubMed for medicine, ERIC for education, PsycINFO for psychology).
• Reproducible: Document your exact search string for each database, including all Boolean operators, truncations, and field codes.
• Wide-ranging: Search grey literature too — government reports, conference proceedings, theses — if publication bias is a concern in your field. It usually is.

Record the number of results from each database before and after deduplication — this feeds directly into your PRISMA flow diagram.

Step 4: Screen for Eligibility

Screening happens in two stages:

1. Title and abstract screening: Apply your inclusion/exclusion criteria to each result. When in doubt, keep the paper for full-text review — it’s easier to exclude at the next stage than to re-run the search.
2. Full-text screening: Apply the same criteria to the complete paper. Document the reason for excluding each paper — you’ll report these in the PRISMA diagram.

For high-stakes reviews, two independent screeners should assess each paper, with disagreements resolved by discussion or a third reviewer. Inter-rater reliability (Cohen’s kappa) should be reported.

Step 5: Assess Study Quality Using Formal Appraisal Tools

Systematic reviews require formal quality assessment of included studies. (This is the step most thesis guides skip, by the way — and the one examiners are most likely to probe.) Common tools by study type:

• RCTs: Cochrane Risk of Bias tool (RoB 2)
• Observational studies: Newcastle-Ottawa Scale (NOS)
• Qualitative studies: CASP (Critical Appraisal Skills Programme) checklist
• Mixed methods: Mixed Methods Appraisal Tool (MMAT)

Quality appraisal affects how much weight you give each study in the synthesis — low-quality studies should be noted and may be excluded from quantitative synthesis (meta-analysis) even if included in the narrative.

Step 6: Extract Data and Synthesise

Create a data extraction form before you start — applied consistently to every included study. Record: study design, population, sample size, intervention/exposure, outcome measures, main findings, and quality appraisal score.

If the included studies are sufficiently homogeneous — similar populations, interventions, outcomes — you can combine results statistically in a meta-analysis. Heterogeneous studies call for a narrative synthesis instead: describe findings thematically and note where study differences explain the variation.

Step 7: Report Using PRISMA

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines require:

• A PRISMA flow diagram showing: records identified → after deduplication → screened by title/abstract → assessed for eligibility (full text) → included
• Reporting of each PRISMA checklist item (27 items in PRISMA 2020)

Most journals publishing systematic reviews require PRISMA compliance. And even for a thesis, following PRISMA signals methodological rigour — the kind that prevents your examiners from picking apart your methods in the viva.