How To Do A Systematic Literature Review In Nursing | Step-By-Step Guide

What A Systematic Review In Nursing Sets Out To Do

A systematic literature review in nursing gathers and judges all relevant studies on a precise question, then reports findings with enough detail for anyone to follow the same path. The aim is to reduce bias at each step and present a balanced picture that helps with care, education, or management. Clear methods, repeatable decisions, and transparent reporting are the hallmarks of good work in this area.

Steps For Doing A Systematic Literature Review In Nursing

Use an ordered plan from idea to write-up. The outline below keeps the work on track and makes every decision traceable.

Set A Focused Question

Pick a question that matters to nursing care or service delivery, and shape it with a structure that enables searching. PICO suits intervention topics (Population, Intervention, Comparison, Outcome). PEO suits qualitative work (Population, Exposure, Outcome). Lock the scope early: patient group, setting, study types, outcomes, and time span. A firm scope prevents drift later.

Write And Register A Protocol

Describe your plan before the first search. List aims, eligibility criteria, databases, search limits, screening steps, appraisal tools, extraction fields, and synthesis plan. Registration on PROSPERO adds time stamps and helps avoid duplication. Keep the protocol short, clear, and stable; log any change during the project.

Build Reproducible Searches

Map concepts to subject headings and free-text terms. Combine with Boolean operators and proximity operators where the platform allows. Capture synonyms, spelling variants, and nursing terms used in practice. Avoid language or date limits unless the protocol justifies them. Save each strategy and export full search histories for the report.

How To Conduct A Systematic Review For Nursing Research Projects

Strong searching and careful screening are the backbone of any review. The steps below keep bias low and data tidy.

Choose Databases And Grey Sources

Search at least two major databases plus nursing-specific sources. CINAHL indexes nursing and allied health. MEDLINE and Embase widen medical reach. PsycINFO adds mental health. Add trial registries, theses, and conference items to reduce publication bias. Record every source, date, and platform.

Craft Database Strategies

Start with one database to refine logic, then adapt to others. Use controlled vocabulary like MeSH and CINAHL Headings, paired with search terms in titles and abstracts. Add truncation and phrase searching with care. Test the string against a small set of known seed papers; if the test set appears, sensitivity is likely acceptable.

Manage Records And Remove Duplicates

Export results to a reference manager or screening tool. Keep raw exports and the merged library. Run de-duplication in a clear sequence: exact match on record IDs, then on title, year, and author. Record counts at each step.

Apply Eligibility Criteria Consistently

Screen titles and abstracts in pairs with blinding where the software allows. Move to full-text screening with the same paired approach. Predefine inclusion and exclusion rules so reviewers make the same call on similar cases. Use a pilot round to align decisions, then proceed to the full set. Track and report reasons for exclusion at full-text stage.

Appraise Study Quality

Pick tools that match study designs. Randomized trials need risk-of-bias domains such as sequence generation, allocation concealment, blinding, incomplete data, and selective reporting. Observational studies need items on confounding, selection, measurement, and follow-up. Qualitative research needs checks on reflexivity, sampling, data collection, and analysis rigor. Rate independently and compare; settle differences through consensus or a third reviewer.

Extract Data With Standard Forms

Design a form that captures study ID, country, setting, design, sample size, participant traits, intervention details, comparators, outcomes, time points, and results. Add fields for funding and conflicts of interest. Pilot the form on a few papers, refine the wording, and then lock it. Extract in duplicate and reconcile line by line.

Plan Synthesis And Decide On Meta-Analysis

Before running numbers, judge if studies are similar enough in design, participants, interventions, and outcomes. If alignment looks sound, a meta-analysis may suit; if not, a structured narrative approach may be better. Either way, state the plan and stick to it unless the protocol allows a justified shift.

Meta-Analysis Choices

Select effect measures that match the outcome type. Risk ratio or odds ratio for dichotomous data. Mean difference or standardized mean difference for continuous data. When heterogeneity is low, a fixed-effect model may fit; when variation exists across studies, a random-effects model may fit better. Quantify heterogeneity with statistics such as I² and τ², and check forest plots for spread and overlap.

Assess Small-Study Effects And Bias

Use contour-enhanced funnel plots and tests like Egger’s where you have enough studies. Check for selective reporting by comparing protocols, trial registries, or methods sections against reported outcomes. Sensitivity checks—such as removing high risk-of-bias studies or switching between fixed and random models—help test the strength of the signal.

When A Narrative Synthesis Fits Best

For mixed methods, complex interventions, or diverse outcomes, group studies by design, population, setting, or intervention features. Summarize patterns in direction and size of effects, and link them to study quality. Tables that line up main characteristics next to results help readers see patterns without guesswork.

Reporting That Meets Established Standards

Use a reporting checklist so readers can follow each decision. The PRISMA 2020 statement lays out items to include in the title, abstract, methods, results, and interpretation sections. A flow diagram shows counts through identification, screening, eligibility, and inclusion. Append full search strings, and share extraction forms and appraisal results when possible.

Write With Clarity

Keep sentences short and active. Define acronyms on first use. State what you did, not only what you found. Use plain labels for tables and figures. Report numbers with enough detail for re-calculation, such as event counts and sample sizes for each arm at each time point.

Present Core Tables And Figures

Include a table of study characteristics, a risk-of-bias summary, and either forest plots or a structured summary of findings. Keep figure captions self-contained so a reader can understand the display without searching the text.

Practical Database Choices And Filters

The list below helps you pick sources and set filters without missing nursing-relevant evidence.

Quality Appraisal Tools That Suit Nursing Topics

Match the tool to the design. Use a single tool across similar designs to keep ratings comparable. Keep the rating process blind to study results where feasible, and store notes that explain each judgement.

Choices By Design Type

Randomized trials: a domain-based risk-of-bias tool. Cohort and case-control: a checklist that handles confounding and selection. Cross-sectional: items on sampling and measurement. Qualitative: a tool that checks reflexivity and data depth. Mixed-methods: a combined tool or parallel ratings by component.

From Appraisal To Weighting

Do not turn ratings into raw scores without a clear rule. It is safer to use ratings to guide sensitivity checks, subgroup plans, or narrative weight. Report how quality influenced the synthesis.

Data Management, Reproducibility, And Ethics

Store files in named folders: protocol, search logs, screening decisions, full texts, extraction sheets, and analysis files. Use version control or time-stamped exports. Keep a decisions log, including any deviation from the protocol and why it occurred. Credit all contributors and declare any conflicts that relate to funding or affiliations.

Software That Helps Without Taking Over

Reference managers help with de-duplication and storage. Screening platforms add blinding and conflict resolution. Spreadsheets or data tools track extraction and cleaning. Statistical software handles meta-analysis. Pick tools that your team can use well rather than chasing features that add complexity.

Common Pitfalls And Simple Fixes

Vague questions lead to messy searches. Fix by sharpening the scope and refining PICO or PEO terms. Narrow searches miss major studies. Fix by testing strings against a set of known papers and adding synonyms. Inconsistent screening inflates bias. Fix by piloting rules and using two reviewers. Weak reporting blocks reuse. Fix by following a checklist and sharing appendices.

Write-Up Outline That Saves Time

Draft while you work. Keep a living methods section that you update after each step, then move it into the final paper. Drop in search dates, database names, and exact strings as soon as you have them. Paste the flow counts after each stage. By the end, you will already have a solid first draft.

Suggested Section Order

Title with the main question. Abstract with the aim, data sources, eligibility, and main findings. Introduction with the gap and why the topic matters to nursing. Methods with protocol, eligibility, sources, strategies, screening, appraisal, extraction, and synthesis. Results with study flow, characteristics, quality, and the main findings. Implications for practice, limits, and research needs. References and appendices with search strings and forms.

Final Checks Before Submission

Run the PRISMA checklist one more time, and verify that tables, figures, and appendices match the text. Confirm that outcomes in the protocol appear in the results. Rerun searches just before submission to capture recent papers and note the update. Share data and code when allowed. If the topic spans clinical risk, cross-check major steps against the Cochrane Handbook for good practice details.

Handling Non-English And Older Studies

Set language plans in the protocol, not after screening begins. If you include English only, state why and note the risk. When a study in another language looks relevant, use an abstract, translation software, or help from a colleague to judge eligibility.

Grey Literature That Saves You From Missing Data

Many nursing audits and service evaluations sit in local repositories. Use search-term strings that mirror your main concepts, then scan pages quickly. Keep a short file name rule so you can trace each item back. When a report lacks full detail, reach out to the named contact.

Summarizing Strength Of Evidence

Readers need to know how secure the findings are. State the main reasons that increase or decrease certainty: study limits, inconsistency, imprecision, indirectness, and bias across studies. A short “summary of findings” table with outcomes, effect sizes, and certainty ratings gives a clear snapshot. Explain any judgement that alters certainty and link it to your notes and sensitivity checks.

Working With An Information Specialist

A trained librarian can translate a PICO into strong search strings and adapt them across platforms. Involve this person during the question and protocol stages, and name them as a contributor in the report. Ask for peer review of search strategies, and store signed-off versions to avoid drift while the project is running.