To begin a medical literature review, define a focused question, draft a protocol, and plan searches, screening, and data extraction.
Readers come to this task with one aim: turn scattered studies into a clear, useful picture. The fastest path starts with scope, a tight question, and a plan you can carry out. This guide walks through early moves that set you up for clean methods, reproducible notes, and a smooth write-up.
Pick The Right Review Type And Scope
Before you write a line, decide what kind of review fits your goal and time. Pin down topic boundaries, populations, interventions, comparators, and outcomes. Keep the question small enough to search in a day or two, yet broad enough to matter. If you plan a formal meta-analysis, keep study designs consistent. If you aim to map a field, lean toward a scoping approach.
| Review Approach | Goal | When It Fits |
|---|---|---|
| Systematic | Answer a narrow clinical question with structured methods and, when suitable, meta-analysis. | Clear PICO, enough trials, and a need for pooled estimates. |
| Scoping | Map concepts, types of evidence, and gaps. | Broad or emerging topics with varied designs. |
| Narrative | Synthesize themes with expert context. | Heterogeneous evidence where pooling is not suitable. |
| Rapid | Provide timely guidance with trimmed steps. | Policy or service needs on short timelines. |
Define A PICO-Style Question You Can Answer
A crisp research question makes every later step easier. Use PICO (Population, Intervention, Comparator, Outcome) or a close cousin like PECO for exposure questions or SPIDER for qualitative work. Write the question in one sentence, then list exact terms and acceptable variants for each element. That list becomes your search seed and your screening yardstick.
Set Up A Simple Protocol You Will Follow
Protocols reduce drift and keep teams aligned. Outline objectives, eligibility criteria, databases, grey-literature sources, screening steps, data items, and planned synthesis. Add a short timeline and who does what. If your review will inform practice or policy, register the protocol to promote transparency and avoid duplication. A widely used registry is PROSPERO.
What Belongs In Eligibility Criteria
State study designs you will include, language limits, date limits, and settings. Define age ranges, diagnostic criteria, and outcome windows. Pre-write your primary outcome and any secondary outcomes. Name exclusion points that can end a screen in seconds, such as wrong population, wrong intervention, or conference abstracts with no full text.
Plan Searches That Balance Sensitivity And Precision
Good searches are planned, tested, and logged. List core concepts from your PICO elements. For each concept, add synonyms, acronyms, and spelling variants. Build a draft string that combines subject headings and free-text terms. Test it in one database and scan the first 100 hits. Tweak until you see a healthy share of known key papers without drowning in noise. Save and document the final strings.
Databases And Grey Sources
Use at least two major databases that fit your topic, such as MEDLINE via PubMed, Embase, or CINAHL for nursing topics. Add trial registries and preprint servers when recency matters. Hand-search reference lists of the most relevant studies, and check citing articles using the “cited by” feature where available. Store every source and date in a log.
Subject Headings And Keywords That Work Together
Subject headings (like MeSH) capture concept families, while keywords catch fresh terms that have not yet been indexed. Pair them. Truncate where appropriate, mind phrase searching, and avoid terms that inflate unrelated hits. A balanced string usually mixes two or three concepts with OR within concepts and AND between concepts.
Close Variant H2: Starting A Medical Literature Review The Right Way
This step-by-step path sets a clean foundation: a tight question, a light protocol, tested searches, and a calm plan for screening. Keep it simple, keep it written, and save every decision. You will thank yourself during PRISMA flow and data extraction.
Build A Lean Screening Workflow
Screen in two passes. First, scan titles and abstracts against your inclusion list. Second, review full texts for studies that pass the first gate. Use two reviewers when stakes are high. Resolve disagreements by chat or a third reader. Track counts at each step so you can finish a PRISMA flow diagram without a scramble.
Quick Rules That Speed Decisions
Pre-decide deal-breakers: wrong population, off-topic outcomes, single-arm case reports for a comparative question, or non-human work when human data are required. Keep a running note of borderline calls and how you resolved them so the logic stays consistent from start to finish.
Design Data Extraction Forms Before Full Screening
Build a pilot form that captures study ID, setting, design, sample size, intervention details, comparators, outcome measures, time points, effect estimates, and risk of bias items. Test the form on three papers. Trim fields that never get used, and add boxes that you needed but missed. Decide how you will handle multi-arm trials, clustered designs, and crossover trials.
Plan For Risk Of Bias Assessment
Pick tools that fit study designs: RoB 2 for randomized trials, ROBINS-I for non-randomized studies, QUADAS-2 for diagnostic accuracy, or CASP for qualitative designs. Agree on signaling questions and judgments in advance, then pilot on two studies to align thresholds.
Write Early With A Reproducible Outline
Do not wait for the last study to start drafting. Create a skeleton with standard sections: background, question, methods, results placeholders, and a short implications box. Drop in your protocol text where it matches the journal template. As you screen and extract, paste numbers into the PRISMA flow and fill the study characteristics table.
How Reporting Standards Shape Your Draft
Reporting guidance saves time and reduces revision cycles. Two resources anchor most health reviews: the PRISMA 2020 statement and the Cochrane Handbook. Use the checklist to plan which items you will report, and lean on the Handbook for methods like effect measures, heterogeneity, and bias assessment.
See PRISMA 2020 for item-by-item reporting guidance, and consult the Cochrane Handbook for stepwise methods detail including search planning, screening, and synthesis. Both links open to the specific pages you need.
Create Clear Inclusion And Exclusion Rules
Write rules that a new teammate could apply without guesswork. Keep them short and measurable. When you meet a tough case, add a one-line note to your rule set showing how you handled it. That becomes a record that guards against bias and keeps the team steady during late-stage screening.
| Criterion | Include If | Exclude If |
|---|---|---|
| Population | Adults with the target condition as defined in the question. | Pediatric-only samples when adults are required, or mixed samples with no separable adult data. |
| Intervention | Named drug, device, or program that matches the scope. | Wrong dose, combined bundle where the target cannot be isolated. |
| Comparator | Placebo, standard care, or the specified alternative. | Single-arm designs for comparative questions. |
| Outcomes | Predefined primary outcome present with a valid measure. | Surrogate-only outcomes when clinical outcomes are required. |
| Study Design | Designs listed in protocol (e.g., RCTs). | Non-listed designs or conference abstracts without full texts. |
| Language/Date | Within set language and date limits. | Outside language or date window without translation or access. |
Document Everything As You Go
Strong notes save hours later. Keep a running log with database names, platforms, exact search strings, dates, filters applied, and the number of records returned. Save PDFs with a consistent file name pattern, and store full-text access notes. Capture reasons for exclusion at the full-text stage in a simple sheet.
Draft The Methods Section First
Methods rarely change once you start screening, so write them early. Describe databases, dates, search strategy basics, screening process, risk of bias tools, data items, effect measures, and plans for synthesis. Add a short line on software used for screening and meta-analysis if you plan to pool results.
Plan Figures And Tables Up Front
Sketch a PRISMA flow diagram with placeholders for counts. List the core tables you will need: study characteristics, risk of bias summary, and main effects by outcome. Decide on any subgroup or sensitivity analyses early so the extraction form captures what you need.
Keep Ethics And Transparency In View
Disclose funding, roles, and conflicts. If you involve patients or the public in shaping the question or outcomes, say how. Share your data extraction sheets and analytic code when possible through a repository so readers can follow or reuse your work.
Pick A Journal And House Style Before You Write Too Far
Each journal sets length, table limits, and style. Check aims and scope, typical article formats, and reference style early. Many medical titles follow ICMJE guidance and the NLM style for references, which affects punctuation and order. Sorting this now avoids painful edits at submission time.
Choose Tools That Keep You Organized
Use a reference manager you trust for de-duplication and PDF storage. Set a folder scheme that mirrors your workflow: one for search logs, one for screening, one for extraction, and one for figures. Name files with year, first author, journal, and an internal ID. Keep a short README in each folder that states what lives there and how you use it.
Coordinate The Team Without Friction
Assign roles before you begin: one person drafts search strings, two people screen, one person resolves ties, one person leads extraction, and one person checks risk of bias calls. A short kickoff chat helps align thresholds, wording, and file naming. Keep a shared tracker for counts and dates so anyone can step in mid-stream.
Plan Time Like A Mini Project
Work in short, well-defined blocks. A helpful cadence is: design the question and protocol in week one, run pilot searches in week two, screen titles and abstracts in weeks three and four, and run full-text screening in weeks five and six. Adjust based on scope and team size. The key is to set check-ins and lock methods early.
Common Early Mistakes You Can Avoid
Scope That Is Too Broad
Questions that bundle many populations or outcomes lead to unmanageable searches and thin synthesis. Split big ideas into a series, or fix one axis while you vary another. Try this: fix one population and compare two interventions, or fix one outcome and compare two populations.
Vague Eligibility Rules
Soft phrases like “relevant outcomes” or “adequate follow-up” slow screening and spark disputes. Swap them for crisp thresholds such as “mortality within 30 days” or “follow-up ≥ 6 months.”
No Pilot Testing
Unpiloted search strings and forms produce avoidable rework. Test on a small batch and tune before you scale.
One-Page Starter Checklist
Before You Search
- Choose review type and write a one-line question.
- Draft a mini-protocol with scope, sources, roles, and timeline.
- Pre-register when suited to policy or practice decisions.
While You Search
- Pair subject headings with strong keywords.
- Record every string, date, and platform.
- Export to a reference manager and de-duplicate.
While You Screen
- Apply inclusion rules in two passes and log reasons for exclusion.
- Pilot data extraction and risk of bias tools on three papers.
- Update counts for your PRISMA flow after each stage.
While You Draft
- Write methods first and paste protocol text where it fits.
- Build tables and figures as you extract.
- Align with journal style and reference format before you polish.