How to write a literature review in nursing

A nursing literature review has a job that reviews in most other fields do not: it supports decisions that affect patients. Examiners read your review as a test of whether you can find, appraise, and weigh clinical evidence, because that is exactly what evidence-based practice will demand of you after graduation.

That purpose changes everything downstream: which databases you search, what counts as evidence, how you appraise it, and how the chapter is structured. Nursing reviews are also more likely than most to be formally systematic, which brings PRISMA reporting and reproducible search methods into scope.

This guide covers the core nursing databases, the hierarchy of evidence, PICO questions, critical appraisal with CASP and JBI tools, PRISMA reporting, a worked outline, and how to keep AI assistance grounded in sources you actually hold.

What makes a nursing literature review different

Nursing reviews come in several recognised formats, and you should be able to name yours. A systematic review follows a predefined protocol and answers a narrow question. An integrative review, common in nursing, combines quantitative and qualitative evidence to answer broader practice questions. A scoping review maps a field without judging effectiveness. A literature review chapter inside a dissertation borrows rigour from all three without necessarily claiming any label.

The format determines the standard you are marked against. Claiming a systematic review without a reproducible search is one of the fastest ways to lose an examiner's confidence, while an integrative review that ignores qualitative evidence on a patient-experience question has missed its own point.

Check your programme handbook early, because many nursing courses specify the format, the minimum number of databases, and the appraisal tool you are expected to use. Aligning with those instructions from the start is far easier than retrofitting a finished chapter to them.

Where to search: CINAHL, MEDLINE, and the core nursing databases

Nursing review protocols usually require multiple named databases, and each covers ground the others miss:

CiteDash's Literature Finder searches an internal corpus plus live PubMed, OpenAlex, Semantic Scholar and arXiv in one pass, which gets your biomedical core and citation trail moving quickly. For a formal systematic review, run CINAHL and Embase through your university library as well, and record every string and date so the search can be reported in full.

• CINAHL: the core index for nursing and allied health, with nursing-specific subject headings
• MEDLINE via PubMed: biomedical coverage, essential for clinical topics and trial evidence
• Cochrane Library: systematic reviews and a register of controlled trials
• Embase: broader European and pharmacological coverage, often required by review protocols
• JBI: evidence syntheses and appraisal resources built around nursing and midwifery practice
• Grey literature: clinical guidelines, government health publications, and professional body reports

Source types and the hierarchy of evidence in nursing

Nursing teaches an explicit hierarchy of evidence, and examiners expect your review to use it. From strongest to weakest for effectiveness questions:

Treat the hierarchy as a weighing tool, not a filter. A well-conducted qualitative study answering a patient-experience question outranks a randomised trial that does not address your question at all. What examiners want to see is that you weighed each study for what it can and cannot support.

Mixed methods studies need both lenses: appraise the quantitative strand and the qualitative strand on their own terms, then judge how well the study integrates them. Nursing questions often span effectiveness and experience at once, which is exactly why the integrative review exists as a form.

• Systematic reviews and meta-analyses of randomised controlled trials
• Individual randomised controlled trials
• Cohort and case-control studies
• Cross-sectional and descriptive studies
• Qualitative studies, which sit outside the quantitative pyramid but are essential for experience and implementation questions
• Expert opinion and consensus guidelines, valuable for practice context rather than as primary evidence

Start from a PICO question

PICO turns a topic into a searchable question: Population, Intervention, Comparison, Outcome. A question like whether a structured repositioning schedule, compared with standard care, improves skin integrity outcomes in adult inpatients at risk of pressure injury gives you four concept blocks to search and clear inclusion criteria to screen against.

Use the variants when they fit better: PEO for qualitative questions, PICOT when timing matters. Each concept block becomes a set of synonyms and subject headings combined with OR, and the blocks join with AND. Write the question down before you search; it is the anchor for every later decision in the review.

Expect to iterate. First searches usually return either thousands of records or almost none, and the fix is adjusting one block at a time: broaden the population terms, tighten the intervention terms, or drop a comparison the literature does not actually make. Record each iteration so the final strategy is the documented product of decisions, not luck.

Turn the PICO into a search strategy

Each PICO block becomes a cluster of search terms: the natural-language synonyms clinicians actually use, plus the subject headings each database applies, such as MeSH in MEDLINE and CINAHL headings in CINAHL. Combine within blocks using OR, across blocks using AND, and apply limits such as date range, language, and human studies only where your protocol justifies them.

Keep a search log as you go: database, interface, string, filters, date run, and results returned. Screening then happens in two passes, titles and abstracts first, full text second, with reasons recorded for full-text exclusions. This is the paper trail that makes your methods section write itself.

Critical appraisal: CASP, JBI checklists, and evidence tables

Every included study needs a documented appraisal, not a private impression. CASP checklists cover the common designs, and JBI appraisal tools are widely used in nursing programmes. Appraise as you read, while the study is fresh, and record the outcome next to the study rather than in a separate pile you plan to reconcile later.

Summarise the results in an evidence table: author, year, country, design, sample, intervention, outcomes, key findings, limitations, and appraisal result. Many examiners turn to the evidence table before they read the narrative, because it shows in one place whether the review rests on strong or weak foundations.

Appraisal is not pass or fail. A study with limitations can stay in the review as long as the narrative weighs its findings accordingly and the evidence table records the concerns. What examiners penalise is not weak evidence honestly handled but weak evidence presented as strong.

PRISMA reporting for nursing systematic reviews

PRISMA is the reporting standard your markers will expect for a systematic review. The flow diagram tracks records identified, duplicates removed, records screened, exclusions with reasons, and studies included. Those numbers are painful to reconstruct afterwards, so keep the running counts from day one.

Report the exact search string for each database and the date you ran it, and fix your inclusion and exclusion criteria before screening starts. Our walkthrough of a PRISMA systematic review with AI covers screening, the evidence matrix, and the flow diagram step by step.

If two reviewers are expected, report how disagreements were resolved; if you screened alone, as most students do, say so and describe the checks you used instead, such as re-screening a sample of excluded records after a break. Examiners accept single-reviewer constraints when they are acknowledged and mitigated rather than hidden.

Check for retracted studies before they reach your reference list

Clinical literature has seen high-profile retractions, and a retracted trial cited in support of a practice recommendation is the kind of error that can sink an otherwise careful thesis. Screen your anchor citations with a free retraction check before the review settles around them.

Inside CiteDash the screen is automatic: retracted papers carry a visible badge in your library and are blocked from citation, so a retraction cannot slip through into the final reference list.

Make the check part of your routine rather than a one-off: run it when a paper first enters your shortlist and again when the final reference list is assembled, since retractions are issued continuously and a paper that was clean at the start of your project may not be by the end.

A worked outline for a nursing literature review

Here is an outline for a review supporting a project on pressure injury prevention, adaptable to most clinical questions:

Let the themes emerge from the included studies rather than imposing them in advance. A review whose themes mirror its evidence reads as honest synthesis; one whose themes were fixed before screening reads as confirmation.

• Introduction: clinical background, significance, and the PICO question
• Search methods: databases, strings, dates, and inclusion and exclusion criteria
• Screening results: the PRISMA flow and the final included set
• Critical appraisal: quality assessment across the included evidence
• Theme 1: effectiveness evidence for the intervention
• Theme 2: patient experience and qualitative findings
• Theme 3: implementation factors, barriers, and facilitators
• Synthesis: what the evidence supports, where it conflicts, and the gap your project addresses

How CiteDash grounds a nursing literature review

In a clinical field, fabricated or misattributed citations are not embarrassing, they are disqualifying. CiteDash is built so every citation is a database object that resolves to a real paper; free-text references do not exist in the system.

The workflow follows the chapter. Save papers to your library, upload PDFs or fetch details from a DOI, and use Ask and extract to pull populations, interventions, and outcomes with sentence-level provenance, so each extracted finding binds to the passage that states it. Build your evidence table as an evidence matrix in the Synthesis Lab, which also supports PRISMA-friendly systematic review work.

When you draft the narrative, the Fact Checker verifies each claim against the full-text PDFs you hold before the text reaches you, and flags anything it cannot verify. Reading statuses keep the screening pile honest while you work through it.

When the chapter is done, the Reference Manager formats the bibliography in your required style; its 12 styles include Vancouver, AMA, and APA, which covers the common nursing programme requirements.

Mistakes examiners flag in nursing reviews

External examiners in nursing see the same problems repeatedly:

Each of these is a distance-from-sources problem. Keep the protocol written, the appraisals documented, and every claim resolvable to a paper you hold, and your review will do what a nursing review must: justify a clinical decision to a sceptical reader.

• Search methods that cannot be reproduced from what is reported
• Included studies with no documented critical appraisal
• The hierarchy of evidence used as a filter rather than a weighing tool
• Guidelines cited as primary evidence instead of tracing the studies beneath them
• An evidence table that summarises while the narrative never synthesises
• No qualitative evidence in a review whose question is really about experience or implementation