How To Do A Literature Review For A Medical Dissertation | Step-By-Step Plan

What Success Looks Like

A strong medical literature review does three things. It shows what is known, what is uncertain, and where your project fits. Readers can retrace your steps. Your searches are reproducible. Your judgements about study quality are transparent.

You can reach that standard with a simple plan. Break the work into stages, assign outputs for each, and hold the line on scope. The table below gives a compact roadmap you can paste into your protocol or methods appendix.

Doing A Literature Review For A Medical Dissertation: Start Strong

Set a specific question. PICO works well for intervention topics. PEO suits qualitative or exposure topics. Write one crisp objective in a single sentence.

Fix your scope early. Pick study designs you’ll include. Set date and language limits only when they make sense. Avoid arbitrary date cutoffs unless methods or standards shifted.

Draft a protocol. One page is enough for a student project. List databases, planned strings, inclusion rules, and appraisal tools. Share it with your supervisor so expectations match your time and word limits.

Sources And Databases That Matter

Start with PubMed for breadth and reliable indexing. Learn filters and field tags from the PubMed Help pages. Add a second database to reduce bias from indexing gaps; Embase is common in medicine. Include a focused source if your field needs one, e.g., CINAHL for nursing or PsycINFO for mental health. If you’re reviewing trials, check the Cochrane Library and trial registries.

Use subject headings. MeSH terms in PubMed raise recall. The MeSH browser shows preferred headings and scope notes. Pair MeSH with free-text synonyms so you catch ahead-of-print items and new terms.

Grey sources help fill gaps. Look for guidelines, registries, and theses databases in your niche. Keep the set small to stay within a student timeline, and document each source and date.

Build A Reproducible Search Strategy

Write the concepts first. Then list synonyms for each concept, including spelling variants and acronyms. Decide which concepts must appear in every record and which are optional. Combine with OR within a concept and AND across concepts.

Use proximity where available. In Embase, NEAR/3 finds words within three words of each other. In PubMed, use phrase quotes for exact phrases and field tags like [tiab] for title/abstract. Add truncation only when it won’t pick up junk terms.

Example PubMed string pattern for an ICU delirium intervention review:

("Delirium"[Mesh] OR delirium[tiab] OR "acute confusion"[tiab])
AND
("Intensive Care Units"[Mesh] OR ICU[tiab] OR "critical care"[tiab])
AND
(prevent*[tiab] OR reduc*[tiab] OR prophylaxis[tiab])

Test and refine. Scan the first few pages of results for false hits and missed relevant papers. Add or drop terms to fix those issues. Save the final strings and export results with all citation fields so you can deduplicate later.

Screening: From Hits To Included Studies

Screen in two passes. First pass: titles and abstracts. Second pass: full texts. Apply the same criteria at both stages. Record reasons for exclusion at the full-text stage so your PRISMA flow is honest and traceable.

Use a simple spreadsheet or a screening tool. Two reviewers are ideal; if you’re solo, add a quick calibration step with your supervisor on a small sample. Borderline calls should be conservative at title/abstract stage and stricter at full text.

Keep a living log. Note the search dates, the exact counts per source, and the date you ran each screen. Small admin habits save hours during write-up.

Appraisal: Judge Study Quality

Match the tool to the design. RoB 2 suits randomized trials. ROBINS-I suits non-randomized comparisons. QUADAS-2 fits diagnostic accuracy work. Newcastle-Ottawa or JBI tools work for cohorts and case-control studies. For qualitative studies, use CASP or JBI critical appraisal tools.

Rate domains, not just a single score. Comment on sequence generation, allocation concealment, blinding, missing data, selective reporting, and conflicts. Explain how concerns affect confidence in the findings.

Capture the data you’ll need later. Prebuild fields for setting, participants, interventions or exposures, comparators, outcomes, time points, and notes on analysis. Add a field for funding and trial registration where relevant.

Synthesis: Turn Evidence Into Insight

Group studies in a way that matches your question. Themes, outcomes, or populations can anchor the structure. Start each section with a one-line takeaway, then back it with numbers and citations. Contrast study designs and settings when that shapes the findings.

If the data line up, run a meta-analysis. Check that populations, interventions, comparators, and outcomes are close enough. Extract effect sizes with confidence intervals. Choose a model that fits the variation you see. Report heterogeneity with I² and explain plausible sources.

When pooling isn’t sensible, write a well-signposted narrative. Use consistent subheadings, tight topic sentences, and neutral verbs. Tables help readers compare arms, doses, and outcomes across studies without scrolling.

Writing The Literature Review In A Medical Thesis: Structure That Works

Readers land on this chapter to see the map. Give them a clean path from question to gap. The outline below keeps the chapter readable and passes common exam rules.

Opening Paragraph

State the question in plain language. One sentence on why the topic matters clinically. One sentence on the setting or population. One sentence on the aim of the review.

Methods Snapshot

Short paragraph on databases, date range, and core limits. One line on screening, one line on appraisal tools, and one line on synthesis plan. Link to full details in an appendix if your word count is tight.

Map Of Existing Evidence

Organize by theme or outcome. Use subheadings that mirror your synthesis plan. Close each subsection with a one-line message that ties back to the question.

Quality And Confidence

Summarize where bias threatens the findings. Name common flaws you saw across studies. Flag any areas with stronger methods and more consistent effects.

Gaps That Justify Your Project

Point to unanswered angles, weak designs, missing outcomes, or under-studied groups. Link those gaps to your dissertation aims so the pivot into your methods feels natural.

Reporting Tools That Keep You Honest

Use PRISMA items as a self-check when you draft the chapter. The PRISMA 2020 statement and its checklist keep methods and reporting tight. Keep the checklist beside your draft and tick each item as you write. Add a PRISMA flow figure if your program asks for one.

When your review includes trials, the Cochrane guidance remains the gold standard for searching, screening, bias assessment, and synthesis choices. The online Cochrane Handbook is free to read and easy to cite.

Data Extraction That Saves Time Later

Build your extraction sheet before full-text screening. That way, you capture the same fields across papers and avoid repeat reading. Keep field names short. Use controlled lists where possible so entries stay consistent.

Include identifiers you’ll need in the methods chapter of your dissertation. Trial registry IDs, protocol links, and funding notes often explain odd results or selective outcomes. Add a quick tick box for conflicts and for whether results match the registered outcomes.

Writing Style That Scores With Examiners

Short sentences beat long ones. Active voice keeps the pace. Pick neutral verbs. Avoid claims that overreach the data. Use figures and tables to carry detail so the body text stays tight.

Use signposts. Open sections with a one-line message. Close them with the takeaway you want the reader to remember. Repeat the question at key junctions to keep the thread visible.

Second-Order Checks Before You Submit

Run through a short list before you hand the chapter to your supervisor or examiner. Catch slip-ups now and the rest of the dissertation goes smoother.

Keep Records Like A Pro

Save every search string and export file. Keep a dated log with counts per database. Store PDFs with a consistent file name that starts with year and first author. Use a reference manager to deduplicate and to insert citations while you write.

Backups matter. Keep a cloud copy of your log, your extraction sheet, and your figures. Keep versions, too. Always. A lost library folder the week before submission is avoidable pain.

Citations And Reference Managers

Pick one manager and stick with it. Zotero is free and flexible; Mendeley and EndNote are common in faculties. Create one library for the dissertation, then one collection per chapter. Add a tag for screening status. Attach PDFs and your notes so core details live next to the record.

Use the plug-in to insert citations while you draft. Set the style your school requires. Lock the style early. Keep authors’ names and titles in sentence case unless the style demands another format. Clean metadata as you import so your reference list doesn’t need a weekend of fixes.

Search Log Fields

• Database and platform
• Exact string used
• Limits and filters
• Hits retrieved and kept
• File names for exports

PRISMA Flow In Plain Terms

Report how many records you found, how many remained after deduplication, how many you screened, how many full texts you read, how many you excluded with reasons, and how many you kept for synthesis. Include counts from registries or hand searching. Label each box so the flow stands on its own.

Ethics, Fairness, And Balance

Cite studies that go against your preferred answer. Name funding sources and conflicts where they might sway results. Flag language or publication limits and say how they could tilt your findings. Balance builds trust with examiners and with future readers of your work.

Quick Templates You Can Reuse

Methods Paragraph Template

We searched PubMed, Embase, and [specialty database] from inception to [date]. Strings combined MeSH and free-text terms for [population], [intervention or exposure], and [outcome]. Two rounds of screening used preset criteria. Study quality was judged with [tool]. Findings were grouped by [theme/outcome/design].

Evidence Table Headings

Study, setting, sample size, population details, design, intervention/exposure, comparator, primary outcome, secondary outcomes, follow-up, risk of bias notes, funding, registry ID.

One-Line Takeaway Starters

• Across trials, effect sizes were small and inconsistent.
• Observational cohorts point to a dose-response link.
• Data for older adults and low-resource settings remain thin.

Common Speed Bumps And Simple Wins

Scope creep drags timelines. Stick to the protocol you agreed at the start. Strange heterogeneity? Split the synthesis by setting or dose rather than forcing a single pool. Too many abstracts to screen? Tighten one concept in the search string and rerun.

Keep your reader in mind at every page turn. State what a section will cover, deliver it cleanly, and move on. Your examiners will thank you with fewer margin notes and faster sign-off.